Purpose and Multiple Uses of ICD

The International Classification of Diseases (ICD) is a tool for reporting and grouping conditions and factors that influence health. It contains categories for diseases, health related conditions, and external causes of illness or death. The purpose of the ICD is to allow the systematic recording, analysis, interpretation and comparison of mortality and morbidity data collected in different countries or areas and at different times. The ICD is used to translate diagnoses of diseases and other health problems into an alphanumeric code, which allows storage, retrieval and analysis of the data. The ICD has become the international standard diagnostic classification for all general epidemiological and many health management purposes. These include analysis of general health situations of population groups, monitoring of incidence and prevalence of diseases, and other health problems in relation to other variables, such as the characteristics and circumstances of the affected individuals. ICD is also suitable for studies of financial aspects, such as billing or resource allocation. The ICD has evolved over the past 150 years from an International List of Causes of Death to a comprehensive classification system for use in mortality, morbidity, casemix, quality and patient safety. It can be used in primary care, secondary care and research. The ICD is used to allocate the majority of the global health resources. Users of the ICD include physicians, nurses, other health care providers, researchers, health information management professionals, coders, health information technology workers, analysts, policy-makers, insurers, patient organizations and many more.

The ICD is used in various settings with different levels of resolution ranging from a set of 100 codes to more than 10,000 codes. It therefore includes an information framework that contains a fully specified set of health concepts and their characteristics and relationships. The ICD–11 ensures consistency with traditional use cases of earlier ICD versions, because it has been built with the past revisions in mind. Past data analyses based on older versions of ICD can be linked to analyses of data based on ICD–11.

All World Health Organization (WHO) Member States are expected to use the most current version of the ICD for reporting death and illness (according to an international treaty, the ‘WHO Nomenclature Regulations’, adopted by the World Health Assembly in 1967). By 2015, many Member States of the WHO had started using the ICD. ICD–10 has been translated into 43 languages, and ICD–11 has been available in all 6 official languages since its publication. Most countries (115 in 2017) use the system to report mortality data, a primary indicator of health status.

The ICD is primarily designed for the classification of diseases and injuries. However, not every problem or reason for coming into contact with health services can be categorized in this way. Consequently, the ICD includes a wide variety of signs, symptoms, abnormal findings, complaints and social factors that represent the content from health-related records (see section on morbidity). The ICD can therefore be used to classify data recorded under headings such as ‘Diagnosis’, ‘Reason for admission’, ‘Conditions treated’ and ‘Reason for consultation’, which appear on a wide variety of health records from which statistics are derived, for treatment, prevention or patient safety.

Intended Use

The ICD has been designed to address the needs of a broad range of use cases: - Mortality - Morbidity - Epidemiology - Casemix - Quality and safety - Primary care Detailed information on the intended use is available in the sections 5 (mortality) and 6 (morbidity). A situation may arise, which anticipates using the ICD-11 for a purpose for which it has not been designed. In this situation, the categorization used within the ICD-11 and its additional features may not be able to address such a new use case. In such cases, it is recommended to consult with the WHO in order to ensure that the information collected is appropriate to the intended new use.

Classification

A classification is ‘an exhaustive set of mutually exclusive categories to aggregate data at a pre-prescribed level of specialization for a specific purpose’ (ISO 17115). Classification involves the categorization of relevant concepts for the purposes of systematic recording or analysis. The categorization is based on one or more logical rules. The purpose of a health classification varies: for example, it may be used in the analysis of cause of death (mortality), morbidity, activity limitation, or participation restriction. Low frequency concepts tend to be grouped but rare concepts may also be separately classified if necessary. Coding rules must be incorporated in the classification to achieve consistency of coding and comparability of coded data over time and space. Classifications are complementary to terminologies, since they are designed to be used for standardized coding of information for statistical purposes.

ICD in the Context of WHO Family of International Classifications (WHO-FIC)

The WHO Family of International Classifications (WHO-FIC) comprises classifications that have been endorsed by the WHO to describe various aspects of health and the health system in a consistent manner. The WHO-FIC provides standardized building blocks for health information systems and consists of 3 broad groups: Reference classifications, Derived classifications and Related classifications. The reference classifications are international reference standards, from which the derived ones have been developed to accommodate a particular detail in specific areas of health. Related classifications cover health domains beyond mere mortality and morbidity (e.g. medicaments classification). Figure 1 illustrates the types of classifications in the WHO-FIC.

Figure 1: WHO Family of Classifications

The purpose of the WHO-FIC is to assist the development of reliable statistical systems at local, national, and international levels, with the aim of improving health status and health care. The classifications are the property of the WHO or other groups. Health related information might sometimes require additional detail to that contained in the ICD. A group or ‘family’ of health relevant classifications covers these needs both by classification of domains different from those of the ICD and provision of more detail for specific uses, e.g. cancer registration. The WHO-FIC designates a suite of integrated classification products that share similar features and can be used singularly or jointly to provide information on different aspects of health and on health care systems. For example, the ICD as a reference classification is mainly used to capture mortality and morbidity. Functioning is classified in the International Classification of Functioning, Disability and Health (ICF) and health interventions in the International Classification of Health Interventions (ICHI).

In general, the WHO-FIC aims to provide a conceptual framework of information dimensions which are related to health and health management. In this way, it provides a common language that improves communication and permits comparisons of data within countries, across countries, health care disciplines, services, and time. The WHO and the WHO-FIC Network strive to build the family of classifications based on sound scientific and taxonomic principles, ensuring it is culturally appropriate and internationally applicable, and focusing on the multi-dimensional aspects of health in order to meet the needs of its different users.

WHO-FIC: Reference Classifications

Reference classifications cover the main parameters of the health system, such as death and disease (ICD), disability, functioning, and health (ICF) and health interventions (ICHI). WHO-FIC reference classifications are a product of international agreements. They have achieved broad acceptance and official agreement for use and are approved and recommended as guidelines for international reporting on health. They may be used as models for the development or revision of other classifications. The three Reference classifications are:

1. International Classification of Diseases (ICD)
2. International Classification of Functioning, Disability & Health (ICF)
3. International Classification Health Intervention (ICHI)

Disability and Functioning – ICF

The ICF is the WHO's framework for measuring health and functioning/disability at both the individual and population levels. While the ICD classifies diseases and causes of death, the ICF classifies health domains. ICD and ICF together provide tools to capture the full picture of health. The ICF classifies health and health-related states in two parts. Part one addresses functioning and disability, described from the perspectives of the body, the individual, and society, and is composed of two components:

1. Body Functions and Structures
2. Activities and Participation life areas

Part two covers contextual factors and has two components: Environmental Factors and Personal Factors (currently not classified in ICF), since an individual's functioning occurs in a context. Functioning is a generic term for body functions (e.g. memory), body structures (e.g. occipital lobe), and activities and participation life areas (e.g. walking, engaging in paid work). It denotes the neutral aspects of the interaction between an individual (related to the individual’s health) and that individual's contextual factors (environmental and personal factors). Disability is an umbrella term for impairments, activity limitations and participation restrictions. It denotes the negative aspects of the interaction between an individual (with a health condition) and that individual's contextual factors (environmental and personal factors). Disabilities are envisioned as a continuum and therefore the ICF and the codes within it do not confer an international binary status of disabled/not disabled. Levels of disability can be used descriptively in clinical settings when formulating a case. Program and policy decision-makers can apply the ICF and specify their own standards for the level of disability as eligibility criteria that are relevant for specific purposes.  ICF includes the following other definitions:

• Body functions are the physiological functions of body systems (including psychological functions).
• Body structures are anatomical parts of the body such as organs, limbs and their components.
• Impairments are problems in body function or structure such as a significant deviation or loss.
• Activity is the execution of a task or action by an individual.
• Activity limitations are difficulties an individual may have in executing activities.
• Participation is involvement in a life situation.
• Participation restrictions are problems an individual may experience in involvement in life situations.
• Environmental factors make up the physical, social and attitudinal environment in which people live and conduct their lives.

ICF includes codes for Body Functions (b), Body Structures (s), Activities and Participation (d), and Environmental Factors (e). ICF codes are only complete with the presence of a qualifier, which denotes the level of health (i.e. severity of the problem from ‘no problem’ to ‘complete problem’). Without qualifiers, codes have no inherent meaning. The ICF acknowledges that every human being can experience a decrement in health and thereby experience some disability. Disabilities can be temporary and may be brief (such as staying home from work for a few days with the flu); they can also be chronic or permanent and may fluctuate in severity over time.

Interventions – ICHI

Intervention classifications are designed to include all kinds of health interventions for treatment, diagnosis or prevention. ICHI includes interventions across all functional sectors of the health system, covering acute care, primary care, rehabilitation, assistance with functioning, prevention, public health, and ancillary services. Interventions provided by all types of providers have been included. The importance of describing and classifying health interventions has long been understood. An International Classification of Procedures in Medicine (ICPM) was published by WHO in 1978, but was not maintained. ICHI is much broader than the former ICPM because it includes the full range of health interventions. Development of ICHI began in 2007, as a joint effort of the WHO-FIC Network and WHO. Its structure has been completed, an alpha version published in 2012 and a beta version in 2015. Finalisation is planned for 2017.

First axis: ‘Target’	Second axis: ‘Action’	Third axis: ‘Means’
The Target axis contains the entities on which the action is carried out. Targets include:	The Action is defined as a deed which is done by an actor to a target during a health care intervention. Actions include:	The Means axis contains the entities describing the processes and methods by which the action is carried out. Means include:
Anatomy	Investigation	Approach: the process by which the target of the action is accessed, e.g. open, endoscopic
Human function	Treating	Technique used as part of the action, e.g. radiation, magnetic resonance
Person or client	Managing	Method: describing how the action is undertaken, e.g. law enforcement, method of transport.
Group or population	Informing
	Assisting
	Preventing

Other attributes of interventions are included as means in the ICHI Content Model. The content of the axes has been restricted to attributes that are common to a wide range of interventions. In particular:

• Devices have not been included as an axis because the majority of interventions do not involve a device and devices change rapidly
• Drugs or other substances administered through an intervention are classified elsewhere (ICD, ATC/DDD, INN).

The coding system comprises a 7-character category structure for the three axes:

• Three letters for the Target
• Two letters for the Action
• Two letters for the Means

ICHI is a flat file comprising valid 7 letter combinations of the three axes. For each intervention included in ICHI, the appropriate 7 letter combination is identified. Not every possible combination of the three axes represents a valid ICHI domain.

WHO-FIC: Derived Classifications

Derived classifications are often tailored for use at the national or international level or for use in a particular specialty. Derived classifications are based upon reference classifications. Derived classifications may be prepared by: - adopting the reference classification structure and classes - providing additional detail beyond that provided by the reference classification, or through rearrangement or by aggregation of items from one or more reference classifications.

Related Classifications

Related classifications are included in WHO-FIC to describe important aspects of health or the health system not covered by reference or derived classifications. Related classifications are:

1. International Classification of Primary Care (ICPC)
2. International Classification of External Causes of Injury (ICECI)
3. Technical aids for persons with disabilities (ISO9999)
4. The Anatomical Therapeutic Chemical Classification with Defined Daily Doses (ATC/DDD)
5. The International Classification for Nursing Practice (ICNP)

Use in Health Information Systems

Health information systems include a range of different components for collection, analysis and use of the data. Information sources could for example be population-based, health facility-based or focus on particular diseases. The main population-based sources of health information are census data, household surveys, and (sample) vital registration systems.

Health facility-related data sources include public health surveillance, health services data (that may be referred to as health management information systems or routine health information systems), and health system monitoring data (e.g. human resources, health infrastructure, financing).

National health accounts are designed to provide a comprehensive picture of health financing.

Coding enables the recording of health information in a language independent way.

Standardization of coding enables both intra- and international data comparison. For example, ICD coded data can be compared across different sectors of the health system – if the same coding rules are applied.

The health information systems are increasingly based on digital (electronic) reporting and coding. ICD–11 is designed to be used in such environments. In many places information collection is based on paper reporting in a traditional analog way. ICD–11 can be produced in a printed version for use in paper based systems.

Use of ICD–11 in a Digital Setting

The ICD-11 is used in an electronic version for coding of electronically reported diagnoses, in electronic health records or electronic death certificates, or in other places. Specific tools facilitate the coding, allowing access to the specific ICD-11 code using any of the several dimensions that define an ICD-11 entity or category. Additional detail can be added using multiple codes for one condition. Relevant rules and instructions to guide the coder are shown in the context of individual categories. Data analysis draws on the framework of the ICD-11 with its multiple dimensions to create previously agreed upon groupings to answer epidemiological or clinical questions. Retaining the unique identifier of the coded ICD-11 entity, allows the same information to be reused across different versions of the ICD, between different translations, and to report conditions at a finer granular level than what is possible with individual ICD-11 codes.

Use of ICD–11 in an Analog Paper Based Setting

The ICD-11 is used in analog printed versions in many countries. Information is reported on paper and then coded with the ICD-11. In order to start coding with ICD-11 and to do initial reporting of mortality and morbidity statistics this may be one way to become acquainted with ICD-11 and health information. Paper based recording however requires manual transcription of the information into electronic systems and should be substituted by electronic reporting as early as possible in the information chain. Further problems with paper based recording include readability and timeliness. ICD-11 supports many ways of computer assisted coding including sanctioning of code combinations and other possible plausibility checks. Therefore, the long term goal for all users should be coding of ICD-11 in an electronic environment, even though the preceding steps and the following steps in the processing of health information may still be carried out on paper. The electronic coding support facilitates coding and improves consistency and reliability of the coding.

Links with other Classifications and Terminologies

ICD coded entities or categories can be used in conjunction with other health relevant classifications or terminologies to fully document an episode of care, or a case for research.

Integrated use with Terminologies

Classification involves grouping information according to logical rules. The grouping is driven by a specific purpose. Terminology allows the reporting of information at any desired level of detail: for example, body parts, findings, or other elements that constitute a disease. Terminologies have no mechanism to report new information that has not previously been added to the terminology. In contrast, a classification has residual classes ('other specified') that ensures that all cases can always be classified. In a terminology, as much as in a modern disease classification, a disease can be defined, for example establishing linkages between its elements, such as anatomy or findings. Terminologies are able to retain the information without emphasizing any aspect of the recorded information. In contrast, classification allows identification of ‘relevant parts’ of the content, for example for public health. International agreement about these relevant parts makes sure that the aggregated information is internationally comparable. The standardized use of the aggregation logic of a classification and the standardized use of the detailed information of a terminology aim at the same result: comparability. International agreement processes are necessary in both cases – and must be the same as soon as the same question has to be answered by the aggregation/classification. Terminologies and classifications should be considered complementary. As an example for a linked terminology within ICD-11 the Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT) can be referenced to, and information coded with the aid of SNOMED CT could be categorized with ICD. Links to additional terminologies, and nomenclatures include for example ICD-O, INN or ICECI. The two standards used together allow better support of data collection, more efficient reporting, and meaningful exchange of data in health information systems.

Joint Use With ICF

Historically, the ICD has used certain disability concepts as common disease or disorder entities, such as: Blindness, Deafness, Mental Retardation, Learning Disability, or Paraplegia, as well as certain disability concepts for other purposes, such as ‘disability as a sequela of injury’, and ‘limitation of activities due to disability’. The ICF was developed after the publication of ICD–10. The ICD–11 has been created both to share concepts and be used jointly with, the ICF. This partnership may assist with the following tasks:

• evaluation for general medical practice (e.g. fitness for work)
• evaluation for social benefits (e.g. disability, pension)
• payment or reimbursement purposes
• needs assessment (e.g. for rehabilitation, occupational assistance, long term care.)
• outcome evaluation of interventions

Signs and symptoms in the ICD were aligned with body functions in the ICF, and ‘factors influencing health status’ in the ICD with contextual factors in the ICF.

Additional selected ICF categories (see sections 1.2 and 1.4.1) are drawn from the component activities and participation (A&P) and help to describe the functional limitations commonly associated with the specific health condition in a functioning pattern. The impact of the disease or disorder in daily activities of a person may vary depending on the severity of the condition as well as the contextual factors (e.g. environmental factors) and possible co-morbidities. The ICD takes an approach that identifies severity as a property of the disease/disorder and describes the impact of the health condition on the daily life of a person as functioning pattern (FP). Details about implementing FPs are described in Section 3.2.5. Instructions on how to code FPs are described in Section 6.4.

Structure and Taxonomy of the ICD Classification System

The chapter and block structure of the ICD has evolved in 11 iterations over 100 years. The authoring of ICD follows a set of rules that ensure the functional and structural integrity of the classification. The evolution of ICD carefully balances the need for categories that match current knowledge while allowing statistical comparability over space and time.

The chapter structure of ICD reflects major aspects of diseases. Chapters are not intended to delimit areas of medical expertise or domains of specialties. The ICD has categories for diseases, disorders, syndromes, signs, symptoms, findings, injuries, external causes of morbidity and mortality, factors influencing health status, reasons for encounter of the health system and traditional medicine. ICD-11 complements these categories with additional detail such as anatomy, substances, infectious agents, or place of injury. ICD-11 also comes with a set of rules and explanations for its use, required reporting formats, and necessary metadata.

The most widespread use of ICD over time and geographically, is for cause of death statistics. The second important use is classification of clinical documentation to provide standardized, language independent information for morbidity use, as resource allocation, case mix, patient safety and quality of care as well as primary care and research. ICD and its definitions are also used as a framework in legislation.

A statistical classification of diseases must be confined to a limited number of mutually exclusive categories able to encompass the complete range of diseases or morbid conditions. The categories are chosen to facilitate the statistical study of disease phenomena. A specific disease entity that is of particular public health importance, or that occurs frequently, should have its own category. Otherwise, categories are assigned to groups of separate but related conditions. Every morbid condition must have a well-defined place in the list of categories. Consequently, throughout the classification, there will be residual categories for other and miscellaneous conditions that cannot be allocated to the more specific categories. The following measures are used to determine whether an entity qualifies to become a unique category:

1. Epidemiological evidence: frequency analyses of coded mortality and morbidity data
2. Clinical evidence: disease evidence provided by the medical specialties
3. Granularity: minimum detail reported and useful in mortality (mortality data) or primary care
4. Continuity: preserve the level of detail pre-existing in ICD
5. Parsimony: the need to limit the number of categories for international mandatory reporting.

The concepts of classification, nomenclature and terminology are closely related. It is the element of grouping that distinguishes a statistical classification from a nomenclature or terminology, which must have separate titles for each known morbid condition. However, nomenclatures or terminologies are also often arranged systematically. A statistical classification can make allowances for different levels of detail if it has a hierarchical structure and subdivisions.

A statistical classification of diseases should retain the ability to identify specific disease entities while allowing statistical presentation of data of broader groups to enable the generation of useful and understandable information. The same general principles apply to the classification of other health problems, and reasons for contact with health-care services, which are also incorporated in the ICD. The ICD has developed as a practical, rather than a purely theoretical classification, in which there are a number of compromises between classification based on aetiology, anatomical site, circumstances of onset, or other criteria.

ICD-11 draws extensively on the method of combining several codes to describe a morbid entity to the desired level of detail. Its electronic architecture allows assignment of unique identifiers to any condition listed - independently whether the condition is grouped in a statistical class or whether it represents a class of its own. The two approaches together allow the option of keeping coding simple where required diagnostic detail is limited; and the alternative to add detail where diagnostic reporting requires a high level of sophistication.

Taxonomy

The authoring of ICD follows a set of guiding principles that ensure the functional and structural integrity of the classification. The international standardized, and most widespread use of ICD over time and geographically, is cause of death statistics. The second important use is classification of clinical documentation to provide pertinent information for resource allocation, case mix, patient safety and quality of care as well as primary care and other kinds of statistics. A statistical classification of diseases must be confined to a limited number of mutually exclusive categories able to encompass the complete range of morbid conditions. The categories are chosen to facilitate the statistical study of disease phenomena. A specific disease entity that is of particular public health importance, or that occurs frequently, should have its own category. Otherwise, categories are assigned to groups of separate but related conditions. Every disease or morbid condition must have a well-defined place in the list of categories. Consequently, throughout the classification, there will be residual categories for other and miscellaneous conditions that cannot be allocated to the more specific categories. The following measures apply in determining whether an entity qualifies to become a unique category:

1. Epidemiological evidence: frequency analyses of coded mortality and morbidity data
2. Clinical evidence: disease evidence provided by the medical specialties
3. Granularity: minimum detail reported and useful in mortality (mortality data) or primary care
4. Continuity: preserve the level of detail pre-existing in ICD
5. Parsimony: the need to limit the number of categories for international mandatory reporting

The concepts of classification, nomenclature and terminology are closely related. It is the element of grouping that distinguishes a statistical classification from a nomenclature or terminology, which must have separate titles for each known morbid condition. However, nomenclatures or terminologies are also often arranged systematically. A statistical classification can allow for different levels of detail if it has a hierarchical structure and subdivisions. A statistical classification of diseases should retain the ability both to identify specific disease entities and to allow statistical presentation of data for broader groups, to enable the attainment of useful and understandable information. The same general principles apply to the classification of other health problems and reasons for contact with health-care services, which are also incorporated in the ICD. The ICD has developed as a practical, rather than a purely theoretical classification, in which there are a number of compromises between classification based on aetiology, anatomical site, circumstances of onset, or other criteria. ICD-11 draws extensively on the method of combining several codes to describe a clinical condition to the desired level of detail. Its electronic architecture allows assignment of unique identifiers to any condition listed - independently whether the condition is grouped in a statistical class or whether it represents a class of its own. The two approaches together allow the option of keeping coding simple where diagnostic detail is limited; and the alternative to add detail where diagnostic reporting requires a high level of sophistication.

Chapter Structure

The ICD is a variable-axis classification. The structure has developed out of that proposed by William Farr in the early days of international discussions on classification structure: - epidemic diseases - constitutional or general diseases - local diseases arranged by site - developmental diseases - injuries

These groups remain in the chapters of ICD–11. The structure has stood the test of time and, though in some ways arbitrary, is still regarded as more useful for general epidemiological purposes than any of the alternatives tested. The conservation of the structure acknowledges the need for stability while allowing incorporation of additional sections. The special groups bring together conditions that would be inconveniently arranged for epidemiological study were they to be scattered, for instance in a classification arranged primarily by anatomical site. These conditions formulate the “special groups” chapters:

Chapter	Title
Chapter  1	Certain infectious or parasitic diseases
Chapter 2	Neoplasms
Chapter  3	Diseases of the blood or blood-forming organs
Chapter  4	Diseases of the immune system
Chapter 18	Pregnancy, childbirth or the puerperium
Chapter 19	Certain conditions originating in the perinatal or neonatal period
Chapter 20	Developmental anomalies
Chapter 22	Injury, poisoning or certain other consequences of external cause

The distinction between the ‘special groups’ chapters and the ‘body systems’ chapters has practical implications for understanding the structure of the classification, for coding to it, and for interpreting statistics based on it. It has to be remembered that, in general, conditions are primarily classified to one of the ‘special groups’ chapters.

Where there is any doubt as to where a condition should be positioned, the ‘special groups’ chapters should take priority. This principle is enforced in the excludes notes at the beginning of each chapter in the ICD. For example, cervical dysplasia grade 1 is coded to the chapter 2 ‘Neoplasms’ because distinction between dysplasia and neoplasia and clinical management are subject to a set of recommended criteria that may change over time.

Revision major steps

The revision of ICD-11 has taken place in several phases.

First, a list of issues that were known from the use of ICD-10 and that could not be solved in its classification structure was compiled and possible solutions were formulated.

Second, input was received from many scientific groups in the key subject areas with a focus on the clinical perspective.

Finally, centralised editing occurred, aimed to adjust imbalances in content generated by multiple independently operating expert groups in the previous phase of the revision, and to ensure the overall structure is consistent and practicable for users in mortality and morbidity statistics. The 'guiding principles' were an essential tool particularly in the last phase. The content, terminology and suggestions for specific groupings by the scientific groups has been preserved, though the proposed structure and location of the entities in the classification has undergone changes necessary to the main uses of ICD. The multiple parenting preserved the visibility of the conditions in the preferred location of the scientific groups. The final version also received input from field testing, Member State comments, and ongoing submission and processing of proposals.

Guiding Principles

Allocation of entities in the classification follows a set of rules that serve to maintain the structural and functional integrity of the classification. The core set of rules listed here is complemented by additional rules that address special cases or serve to ensure consistent user guidance (see annex). They are listed in order of priority.

1. No changes to the classification, including movement of categories or groups between chapters, without rationale and documented change in aetiology or prevention method. (e.g. Diseases of the immune system was added as a new chapter as there was sufficient scientific evidence to support this move. Alternatively, while it was suggested to move 'wounds of skin' to Disease of the skin, the wound of the skin, being an injury, remains grouped with injuries. Prevention will focus on the cause of the wound.

2. Conditions are classified predominantly by their aetiology.

   a. Local manifestations of important ‘aetiologies’ are located in the aetiology chapter (e.g. Viral hepatitis is in Certain infectious or parasitic disease).

   b. Where one condition can be due to multiple different aetiologies, and it is more relevant to retain the affected body system, it is usually classified with the body system, (e.g. some gastric ulcers are caused by bacteria, but they remain in the Digestive system chapter).

   c. Where the aetiology of the condition is unknown, it is allocated to the most relevant organ system (e.g. Costen syndrome is in the Digestive chapter).

   d. Systemic ‘aetiologies’ are primarily in their relevant aetiology chapter (e.g. Idiopathic inflammatory myopathy is in Diseases of the immune system).

3. Conditions that could arguably be in two or more places of the classification remain in their legacy location.

   a. For example, injuries of the eye are equally important for the eye and their prevention. Despite the suggestion of including them in the eye chapter, they remained where they were, in the injury chapter.

   b. Where aetiology and body system are equally important, the legacy location remains unchanged (ocular motor nerve palsies).

4. Keeping a group of subtypes together in one location may override anatomical or aetiological considerations (e.g. human prion diseases, some have a genetic component, others a transmissible component).

Guiding principles for special concepts

1. Clinical findings are located in the chapter ‘Symptoms, signs or clinical findings, not elsewhere classified’. (e.g. Abnormal serum enzyme levels or Results of function studies of the circulatory system)
2. Manifestations of diseases and a relevant point for a health intervention are ‘clinical manifestations’ and are located in the body system chapter where they manifest. The underlying condition has to be coded as well. (e.g. myocarditis)
3. Syndromes, where the aetiology is unknown, are allocated with the most relevant organ system. (e.g. Costen syndrome is in the Digestive chapter)
4. The number of categories with ‘due to’ in the title are restricted to certain exceptions. (e.g. Sepsis due to certain bacteria)
5. Chronic, specific postprocedural conditions are grouped at the end of the organ system chapter where they manifest. (e.g. lymphoedema due to surgery or radiotherapy). Residual categories do not exist for these groups.
6. Acute postprocedural complications are identified by combinations of codes from body system, injury and external causes chapters. (e.g. an accidental puncture during an intervention is classified with a code for the injured organ, a code identifying the accidental puncture as the mode of injury and a code describing what surgery occurred as the mechanism of injury)
7. Categories with mention of ‘multiple’ are restricted to exceptions and require coding of the different multiple conditions individually. (e.g. multiple injuries are to be coded individually)
8. Categories with mention of ‘sequelae’ are, if there are any at all, restricted to exceptions. The specific condition resulting as a sequela needs to be coded along with the underlying cause. In some instances, they will continue to exist with the label ‘late effects of…’ (e.g. late effects of cerebrovascular disease or late syphilis)
9. Categories with mention of ‘history of’ are limited to exceptions. (e.g. personal history of malignant neoplasms lists only the more frequent anatomical sites)
10. High level groupings need to be meaningful.
11. Residual categories exist only where they are meaningful. (e.g. where conditions are either congenital or acquired, there is no ‘other’ residual, but there will be an ‘unspecified’ option)

Improving user guidance

The following rules serve to provide user guidance. Users may expect to find conditions in certain places when browsing the tree structure. User groups may need to group data or create subsets for other reasons. The multiple parenting in the foundation serves to address that issue.

1. Where a condition could be in two or more places, identify these other places and add them as secondary parents, e.g. malignant neoplasm of the colon is coded to the neoplasm chapter, but is also shown in the chapter of diseases of the digestive system. In case a set of conditions needs to be shown in more than one place and there is no grouping matching that set, create a window (no primary children, no terms, no residual categories) in the appropriate place.
2. Where a condition could be confused with another condition bearing a similar name, add an exclusion note. (e.g. ‘Influenza due to seasonal influenza virus’ has a note ’Exclusion: Haemophilus influenzae [H. influenzae] meningitis’).
3. Where alternative ways of tabulating data are required, create a special tabulation list as a second parent. (e.g. infectious diseases by agent.). The coding scheme of the individual entries will remain the one used for the full international classification.
4. Where diseases of certain body systems are spread across different chapters, allow for a specialty tabulation of the pertinent diseases. The coding scheme of the individual entries will remain the one used for the full international classification. Currently there are specialty tabulations for primary care, dermatology, neurology, ophthalmology and the special cases ICD-O and ICECI.

### General features of ICD–11 ###

The main structural innovation of ICD–11 is that it is built on a foundation component from which the tabular list can be derived. The international reference Tabular list is the statistical classification for morbidity and mortality. Due to the addition of a Foundation component, and the electronic design of ICD-11, some new terminology had to be introduced that had not been used in prior versions of ICD. The table below provides examples of this new terminology. You will find more detail about individual aspects in other parts of this guide.

ICD–11 Term	Explanation
Foundation	Underlying data base content that holds all necessary information to generate print versions of the tabular list and the alphabetical index as well as additional information that is needed to generate specialty versions of ICD-11 and country specific modifications.
Stem code	Stem codes are codes that can be used alone. They are found in the tabular list of ICD-11 for Mortality and Morbidity Statistics. Stem codes may be entities or groupings of high relevance, or clinical conditions that should always be described as one single category. The design of stem codes makes sure that in use cases that require onely one code per case, a meaningful minimum of information is collected.
Extension code	Some users and settings are interested in reporting more detail than is included in a stem code. This additional detail can be coded using an Extension code. Extension codes can never be used without a stem code, but are used together with the stem codes to define the full detail of a reported disease.
Cluster coding	Stem codes can be used together with extension codes or with other stem codes to fully describe one diagnosis. In this case codes will be clustered together. This mechanism is called “cluster coding”.
Primary and secondary parents	The hierarchy of ICD-11 is defined the same as it was in previous versions of ICD. The possibility to connect specific diseases and concepts within the classification to another parent code was introduced to enable specific extracts of the Tabular list for medical specialities or for specific use cases. For example, a code for a malignant neoplasm of the skin is located in the chapter for malignant neoplasms. The primary parent for this code is a code or a block from this chapter. However, a medical doctor treating only skin diseases might want to see only codes from the classification that are relevant for his or her specific clinical purpose. Therefore, a secondary parent was defined in the skin chapter which will only show the code in this chapter if the specific extract of code for his or her use case is selected.

Details of the differences between ICD–10 and ICD–11 in individual chapters are explained in the annex.

Coding scheme:

• The chapter numbering is in Arabic numbers and not Roman numerals.
• The coding scheme for categories now has 4 characters, and there are 2 levels of subcategories.
• The coding scheme always has a letter in the second position to differentiate from the codes of ICD–10.
• In ICD–11, the first character of the code always relates to the chapter number. It may be a number or a letter. The code range of a single chapter always has the same character in the first position.
• In order to describe a causal relationship between conditions in a code title the preferred term is ‘due to’.
• In order to indicate the concurrence of two conditions in a code title the preferred term is ‘associated with’.

Extension codes and additional subclassifications

ICD–11 allows adding specific detail to coded entities by the following mechanisms:

1. The extension codes comprised of groups of codes e.g. anatomy, agent, histopathology and other aspects that may be used to add detail to a code. Extension codes are not to be used alone but have to be added to a stem code. Not all extension codes are allowed to be used with every stem code. As well there might be places where only a subgroup of an extension code can be used together with a specific stem code. This will be indicated in the tabular list by showing the valid extension codes for each stem code. Sometimes an extension code might be valid for a group of stem codes. This will be indicated in the tabular list by displaying such extension code at the highest level within the hierarchy.
2. ‘Code also’ instructions inform about additional aetiological information which needs to be coded in conjunction with certain categories, because that additional information is relevant for primary tabulation. The ‘code also’ statement marks the categories that should be used only in conjunction with the indicated second code(s). In some instances, they may be a reason for treatment in their own right, where aetiology is unknown.
3. ICD–11 has an explicit way of marking a cluster of codes that are jointly used to code one condition, called cluster coding. This is a notable new feature in ICD-11 that creates an ability to link core diagnostic concepts (i.e. stem code concepts) when desired, and/or to add clinical concepts captured in extension codes to primary stem code concepts. Either way, it should be emphasized that the clsutering ability inherent to ICD-11 is one of the significant changes relative to ICD-10.

Other general features:

ICD–11 categories have a short description and a long definition labelled ‘additional information'. The description is a short characterisation (maximum of 100 words) of the entity that states things that are always true about a disease or condition and necessary to understand the scope of the rubric. It appears in the tabular list of the classification. The long ‘additional information’ is the full definition, without length restriction, including detailed information that appears in the foundation component only.

• Special tabulation lists continue to exist in ICD-11, but there are three additional ones- the Startup Mortality List (SMoL), the list for verbal autopsy and infectious diseases by agent. Additional special tabulations can be derived from the new multiple parenting technique, e.g. all WHO notifiable diseases, listing all conditions that are assigned to the relevant section of the infectious diseases chapter. Specialty tabulations allow the representation of content from the angle of a specialty,such as dermatology or neurology, creating subsets, and allowing the precoordinatation of more detail, if desired.
• For morbidity, the definition of main diagnosis has changed to be the reason for admission after assessment at the end of the stay. This definition is less prone to interpretation, and countries that had switched from the 'biggest resources' definition to the 'reason for admission at the end of the stay'using ICD-10, noticed only small changes in their activity statistics.

Foundation Component and Tabular Lists of ICD–11

The Foundation Component is a multidimensional collection of all ICD entities. Entities can be diseases, disorders, injuries, external causes, signs and symptoms. Some entities may be very broad, for example ‘injury of the arm’, while others are more detailed, for example ‘laceration of the skin of the thumb'. The Foundation Component also has the necessary information to use the entities to build a tabular list (a mono hierarchy in the style of a traditional statistical classification). The foundation includes information on where and how a certain entity is represented in a tabular list, whether it becomes a grouping, a category with a stem code, or whether it is mentioned as an inclusion term in a particular category.

Several different tabular lists can be built from the foundation component. Drawing on the same foundation, a set of tabular lists that builds on the same hierarchical tree can be created – a set of so called congruent tabular lists. Data that is collected with any tabular list of such a congruent set can always be aggregated to the smallest common denominator (provided the same rules for reporting, coding and selection have been applied). The foundation component includes instructions on how to combine certain codes in a tabular list to achieve more detail in coding. These rules help coders and computer systems to visualize the permitted code combinations when they are using a tabular list.

In a tabular list, entities of the foundation become categories. The categories are mutually exclusive and jointly exhaustive, and linked to a mono hierarchical tree (they have only one parent). The information related to an entity that has become a category and has multiple parents is still available from the foundation. This information can be used to visualize that category in more than one place in the tabular list, e.g. showing them in black in its place for reference tabulation and in grey in any other place for browsing or alternative tabulations. Multiple parenting is explained in a separate section (see Basic coding guidelines). ICD–11 has multiple congruent tabular lists with varying levels of detail.

Core tabular lists for international use are:

• Mortality and Morbidity Statistics (MMS)
• Primary care low resources settings (PCL)
• Primary care intermediate resources setting (PCM)
• Verbal Autopsy (VA)
• Simple Mortality List (SMoL)

The full name of such a tabular list will always include ‘ICD–11’, e.g. ICD–11–MMS.

Stem Codes and Extension Codes

For instructions for use of stem codes and extension codes see later sections ‘Extension codes’ and ‘Basic Coding Guidelines’

Pre- and Post-Coordination, Cluster

A health condition may be described to any level of detail, by applying more than one code, or by 'clustering (i.e.combining)

• two or more stem codes, (i.e., code1/code2)
• stem codes with one or more extension codes. (i.e.stem code & extension code1 & extension code2)

In this manner, the classification can address a big number of items with a limited range of categories.
Stem codes contain all pertinent information in a pre-combined fashion. This is referred to as ‘pre-coordination’. When additional detail is achieved by combining multiple codes, this is referred to as ‘post-coordination’ and is handled by cluster coding in ICD-11. Below, two examples are presented to show first a precoordinated diagnostic concept, then one that benefits from cluster coding.

Example:	Single code (precoordination)
Single code:	‘squamous cell carcinoma of lung’

In precoordination, both site and pathology are combined in a single precoordinated diagnostic code

Example:	Cluster coding (postcoordination)
Stem code:	other specified neoplasm of lung
Extension code:	large cell carcinoma

In cluster coding, the condition 'large cell carcinoma of bronchus and lung' is expressed through a combination of two linked codes.

Multiple Parenting

An entity may be correctly classified in two different places, e.g. by site or by aetiology. For a disease like oesophageal cancer this would mean that it could be classified to cancers (malignant neoplasms) or to conditions of the digestive system. In the same way, cerebral ischaemic conditions could be classified to the vascular system – where the problem arises - or to the nervous system – where the ischaemia impacts and manifests with symptoms.

In the foundation component 'includes' notes for these examples will have both mentioned possible parents (multiple parents). However, for the tabulation of statistical outputs from any tabular list, there can be only one parent for primary tabulation. When there are such multiple parents, in the foundation view both parents will be displayed the same way. However, in a tabular list, the primary parent place will show the entity and its parents in black, and possibly the secondary parent place (e.g. for oesophageal cancer primary parent malignant neoplasm will appear in black and the digestive system for the oesophageal cancer) in grey. Similar to the online version, the print version (i.e. the hard copy) in grey, if needed for the specific use the tabular list was generated for.

Every time an entity is parented elsewhere, it will continue to show the code from the primary parent. The primary parent is sometimes referred to as the ‘Tabular list parent’.

The Content Model

ICD–11 holds all its content in the foundation component. Here, every entity is specified by a definition, machine readable properties that have values, and one or more parent-child relationship. Additional links provide information for post-coordination. All this multi-dimensional information is then projected on one line with mutually exclusive categories, as the tabular lists. The foundation includes information on where and how a certain entity is represented in a tabular list. An entity might become a grouping, a category, or just a term that is, for example, listed in the index. The Content Model is a structured framework that defines each entity found in the ICD in a standard way. The purpose of the Content Model is to present the background knowledge that provides the basis for the definition of each ICD entity in a systematic way to allow for computerization. Each ICD entity can be seen from different dimensions. The Content Model represents each one of these dimensions as a ‘property’. For example, there are currently 13 defined main properties in the content model to describe an entity in the ICD.

A disease is usually defined using addressing a set of relevant aspects drawn from the pattern below. A disease is a set of dysfunctions in any body system defined by:

Property	Description
1. Symptomatology or manifestations:	known pattern of signs, symptoms, and related findings
2. Aetiology:	an underlying explanatory mechanism
3. Course and outcome:	a distinct pattern of development over time
4. Treatment response:	a known pattern of response to interventions
5. Linkage to genetic factors:	e.g., genotypes, patterns of gene expression, etc.
6. Linkage to environmental factors

The key components of the definition of disease are included as different properties within the Content Model. The thirteen main properties of the Content Model are:

1. ICD Entity Title
2. Classification Properties
3. Textual Definitions
4. Terms
5. Body System/Body Part
6. Temporal Properties
7. Severity of Subtypes Properties
8. Manifestation Properties (Signs, Symptoms or Investigation Findings)
9. Causal Properties
10. Functioning Properties
11. Specific Condition Properties
12. Treatment Properties
13. Diagnostic Criteria

For each ICD entity, various properties can be given if necessary to reach the correct coding result. At the time of initial release of ICD-11 only absolutely necessary properties will be defined in order to avoid the necessity of frequent updates and to reduce the resources needed in implementing countries to update the classification within a short timeframe. Additions of property values on international level can be managed through the regular update cycle whenever coding problems indicate the necessity to do so. For example:

ICD entity: invasive ductal carcinoma of breast

Properties	Value
Anatomy	breast
Morphology	invasive ductal carcinoma

The full range of different values for a given property is predefined using standard terminologies and ontologies. This range of values is called a "Value set".

Definitions

Descriptions of ICD–11 concepts inform analysts and translators about the meaning of an entity and its descriptive characteristics. There are two different types of descriptions: a short description (maximum of 100 words) that is available in both the content model and the tabular list, and a detailed description with comprehensive detail at the level required for each entity. The detailed description appears only in the content model and possibly in electronic versions online.

Coders are cautioned not to use the descriptions to establish a diagnosis. Coders must assign codes based on the diagnosis(es) documented by the clinician.

The descriptor information that is included for the individual entities of the ICD-11 provides users of the ICD clear insight regarding the intended meaning and scope of rubrics or terms in the tabular list and the Foundation component. The descriptors guide translators, coders and users of coded data. The goal is to enhance the comparability, consistency, and interpretation of coded information for everyone, everywhere. There are four levels of descriptor information in the ICD–11 content model.

Fully Specified Term

This is an unambiguous title that does not assume context. For example, “systemic illness with predominant gastrointestinal diarrheal symptoms attributable to vibrio cholera” as opposed to “cholera” or “other” (where the meaning of other would have been clear from the hierarchical context.)

Description

The description is a short characterization (maximum of 100 words) of the entity that states things that are always true about a disease or condition and necessary to understand the scope of the rubric. Descriptions do not contain elements intended for in level 3 (common epidemiology) or things that may be true for level 4 (clinical criteria). Descriptions were formerly called ‘short definitions’.

Additional Information

This is a text field that is not mandatory, but that may contain any additional information about, or characteristics of, the diseases or conditions included in the entity. This text field provides more context for the entity. For example, the most common epidemiologic circumstances, putative or highly suspected etiologic agents, or other information that may not always be true but may be common, typical, or expected. Additional information was formally called ‘long definition’.

Clinical or Diagnostic Criteria

This element is presently unpopulated within the Foundation content model. It is intended to contain one or more scenarios of clinical criteria and characteristics that would be sufficient to diagnose the condition(s) or syndrome(s) of the given class or concept. Such scenarios should contain multiple variations, or embedded logic to accommodate “x out of n” variations, that are necessary or useful to make the diagnosis. For example, a myocardial infarction in high-resource diagnostic settings would typically include a longitudinal pattern of cardiac enzymes, specific EKG changes, and stereotypical symptoms. However, only two out of these three need be present as there are such things as “silent MIs (without symptoms) and similar variations. It is expected that these scenarios will be divided over technology capabilities. For example, diagnosing a myocardial infarction in the high-resource diagnostic settings would likely involve different technology and criteria than in low-resource settings. ICD diagnostic criteria draw on various WHO guidelines that have identified diagnostic rules (e.g. guidelines, criteria). Extensions to the ICD, as specialty tabular lists, may use such diagnostic rules. The ICD-11 architecture, and the future evolution of this component of information, could eventually serve decision algorithms based on these criteria. Assignment of diagnoses and conditions could automatically be proposed from data arising in electronic medical records. Populating the clinical criteria is a future project that requires further planning. If necessary, diagnostic criteria describe diagnostic methodology that determines how health providers diagnose cases that are classified to an entity. It contains the core diagnostic information necessary and sufficient to describe a category, and enables the digital representation of the diagnostic algorithms using standardized terminology and other elements as appropriate. There may be different sets of diagnostic criteria for different settings. Diagnostic Criteria draw on content of other attributes. ICD diagnostic criteria draw on various WHO guidelines that have identified diagnostic rules (e.g. guidelines, criteria). Extensions to the ICD, as specialty tabular lists, may use such diagnostic rules.

Functioning properties

In order to describe the impact of morbidity on the life of a person it is not only necessary to mention the diseases a patient suffers but the impact of these diseases on functioning, activity and participation of a person needs to be described, too. This is possible with the embedded functioning properties. A detailed description on how to use the functioning properties is given in a separate section below. Note: Functioning embedded in ICD allows a first documentation of functioning of an individual. Using selected subsets for the assessment of functioning has proven a useful method that is also used in the WHO Disability Assessment Scale (WHODAS). Where possible, the full ICF should be used for a complete reporting of functioning.

Language independent ICD Concepts

ICD-11 concepts are language independent. All concepts have unique identifier (URI), and have a specific place in a hierarchy of groups, categories and narrower terms. The maintenance of the ICD-11 on an international level is handled in the English language but the content model of the ICD–11 is language independent and allows binding of any desired language to the elements of its foundation. In this way, an international translation base facilitates translations or multilingual browsing. The unique identifier (implemented as an URI) remains valid independently whether an ICD concept is still valid or has been retired. The hierarchical structure of groups, categories, subcategories, and inclusions (parents, children and narrower terms) serves also as a language independent backbone for translations of ICD. This structure that is one component of the foundation is essential when building an index in a local language. It helps (in conjunction with the ICD translation platform) to identify the things that need to be translated in order to be able to address ICD categories with terms reported in the local language.

Organization of a Congruent System

Many countries use a single coding system (tabular list) for all use cases. Congruent, telescopically expandable and collapsible purpose-independent subsets for morbidity coding in different settings (comparable to Verbal Autopsy, or initial implementation lists for mortality) allow gathering of information at different levels of detail and still allow for comparison of the collected information at the level of the common description.

ICD–11 conventions

ICD–11 has standard ways of presenting its content. Conventions describe textual content and also apply to the coding structure.

Code Structure

The codes of the ICD–11 are alphanumeric and cover the range from 1A00.00 to ZZ9Z.ZZ. Codes starting with ‘X’ indicate an extension code (see Extension codes). The inclusion of a forced number at the 3rd character position prevents spelling ‘undesirable words’. A letter in the 2nd character position allows for clear distinction between a code from ICD–11 and one from ICD–10. The letters ‘O’ and ‘I’ are omitted to prevent confusion with the numbers ‘0’ and ‘1’. Technically, the coding scheme would be described as below:

EDEE.EE

• E corresponds to a 'base 34 number' (0-9 and A-Z; excluding O, I);
• D corresponds to 'base 24 number' (A-Z; excluding O, I); and
• 1 corresponds to the 'base 10 integers' (0-9)
• The first E starts with ‘1’ and is allocated for the chapter. (i.e. 1 is for the first chapter, 2: chapter 2, … A chapter 10, etc.)

The terminal letter Y is reserved for the residual category ‘other specified’ and the terminal letter ‘Z’ is reserved for the residual category ‘unspecified’. For the chapters that have more than 240 blocks, ‘F’ (‘other specified’) and ‘G’ (‘unspecified’) are also used to indicate residual categories (due to problems with the coding space).

Chapters are indicated by the first character. For example, 1A00 is a code in chapter 1, and BA00 is a code in chapter 11.

Groups are not coded within this code structure. However, hierarchical relations are retained in the 4-digit codes. There is unused coding space allocated in all blocks to allow for later updates and to keep the codes stable.

Inclusions

Within the coded categories there are typically other optional diagnostic terms. These are known as ‘inclusion terms’ and are given, in addition to the title, as examples of the diagnostic statements to be classified to that category. They may refer to different conditions or be synonyms. They are not a sub-classification of the category. Inclusion terms are listed primarily as a guide to the content of the category, in addition to the definition. Many of the items listed relate to important or common terms belonging to the category. Others are borderline conditions or sites listed to distinguish the boundary between one subcategory and another. The lists of inclusion terms are by no means exhaustive. Alternative names of diagnostic entities (synonyms) are included and shown in the electronic coding tool and the Alphabetical Index, which should be referred to first when coding a given diagnostic statement. It is sometimes necessary to read inclusion terms in conjunction with titles. This usually occurs when the inclusion terms describe lists of sites or pharmaceutical products, where appropriate words from the title (e.g. ‘malignant neoplasm of ...’, ‘injury to ...’, ‘toxic effects of ...’) need to be understood. General diagnostic descriptions common to a range of categories, or to all the subcategories in a four-character category, are to be found in the notes heading ‘Inclusions’, immediately following a chapter, group, or category title.

Exclusions

Certain categories contain lists of conditions preceded by the word ‘Exclusions’. These are terms which are classified elsewhere. An example of this is 6A40 Hyperfunction of pituitary gland which excludes Cushing syndrome (6A50) Exclusions serve as a cross reference in ICD and help to delimitate the boundaries of a category. General exclusions for a range of categories or for all subcategories are found in the notes heading ‘Excludes’, immediately following a chapter, group or category title. Some exclusions may be language dependent. The meaning of ICD entities is designed to be the same in all languages but different languages have different sets of synonyms. As a result, a language specific term can have a language specific homonym in a different part of the classification, while such homonym does not exist in English. In such instances, a language specific exclusion note would inform the user and will have a cross-reference. Multiple parenting in ICD-11 shows categories in the context of siblings that are placed elsewhere in the classification. This is also an indication of an exclusion and means ‘some sibling is coded elsewhere’. In the print and the coder version this information is displayed as an exclusion as well.

Code also - Use additional code, if desired

Code also instructions inform the user about mandatory additional information which has to be coded in conjunction with certain categories because that additional information is relevant for primary tabulation. The ‘code also’ statement marks the categories that should be used in conjunction with the indicated second code(s). However, in some instances they may be a reason for treatment in their own right, where aetiology is unknown, and the code is reported alone. Use additional code, if desired instructions inform the user about optional additional detail that can be added for a particular diagnosis.

‘NEC’ and ‘NOS’

‘NEC’

The words ‘not elsewhere classified’, when used in a category title, serve as a warning that certain specified variants of the listed conditions may appear in other parts of the classification.

‘NOS’

The letters NOS are an abbreviation for ‘not otherwise specified’, implying that the documentation that is used for classifying does not provide more detail than just that term (implying ‘unspecified’, ‘incompletely specified’ or ‘unqualified’). Sometimes an unqualified term is nevertheless classified to a rubric for a more specific type of the condition. This is because, in medical terminology, the most common form of a condition is often known by the name of the condition itself and only the less common types are qualified. For example, ‘pharyngitis’ is commonly used to mean ‘acute pharyngitis’. These inbuilt assumptions have to be taken into account in order to avoid incorrect classification. Careful inspection of inclusion terms will reveal where an assumption of cause has to be accounted for; coders should be careful not to code a term as unqualified unless it is quite clear that no information is available that would permit a more specific assignment elsewhere. Similarly, in interpreting statistics based on the ICD, some conditions assigned to an apparently specified category will not have been so specified on the record that was coded. When comparing trends over time and interpreting statistics, it is important to be aware that assumptions may change from one revision of the ICD to another. For example, before the Eighth Revision, an unqualified aortic aneurysm was assumed to be due to syphilis (this is no longer the case since ICD–10). In ICD-11 in most instances the ‘NOS’ point to unspecified categories, so that the later data analysis can take care of assumptions or not regarding the linguistic meaning.

‘Certain’

The term ‘certain’ informs that some entities that could be grouped here are grouped somewhere else outside the current chapter or block.

Residual categories – ‘Other’ and ‘Unspecified’

ICD-11 coding should always be completed to include the highest level of detail possible with the use of one code or multiple codes as described above. There are, however, circumstances when that is not possible and for that reason the ICD-11 includes categories titled ‘other’ and ‘unspecified’. In some instances, necessary information to select a specific category may not be available in the source documentation. When this is the case, the residual category ‘unspecified’ is selected. Conversely, there are instances where the information in the source documentation is very specific, but the tabular list does not include a specific category. In this case, users identify the closest category match, and code to the residual category titled ‘other’.

Sometimes an unqualified term is classified to a category for a more specific type of the condition. This is because, in medical terminology, the most common form of a condition is often known by the name of the condition itself and only the less common types are qualified. For example, ‘mitral stenosis’ historically meant ‘rheumatic mitral stenosis’. These inbuilt assumptions have to be taken into account in order to avoid incorrect classification. Careful inspection of inclusion terms will reveal where an assumption of cause has been made; coders should be careful not to code a term as unqualified unless it is quite clear that no information is available that would permit a more specific assignment elsewhere. Similarly, in interpreting statistics based on the ICD, some conditions assigned to an apparently specified category will not have been so specified on the record that was coded. When comparing trends over time and interpreting statistics, it is important to be aware that assumptions may change from one revision of the ICD to another. For example, before the Eighth Revision, an unqualified aortic aneurysm was assumed to be due to syphilis.

Use of ‘And’ and ‘Or’

The words ‘and’ and ‘or’ in ICD–11 are used in their meaning in formal logic. A term that includes a statement of the kind 'A and B' means that both, A and B, have to be present in order to use that category. A term that includes a statement of the kind ‘A or B’ means that either A or B, or both, have to be present in order to use the category. Because A or B can mean either A or B or both, ‘or’ now meaning "and/or". (The term ‘and’ meaning ‘and/or’ found in ICD–10 has not been carried over into ICD–11.)

‘Due to’ and ‘With’

‘Due to’ is the preferred term for categories where two conditions are mentioned and a causal sequence exists. Other terms, such as ‘caused by’ or ‘attributed to’ may be allowed synonyms. The phrase ‘secondary to’ is equivalent and may also be included as a synonym. ‘Associated with’ is the preferred term for categories where two conditions are mentioned and there is no causal sequence implied.

Spelling, parentheses, grammar and other conventions

Spelling and grammar of ICD-11 follow the British rules with exceptions and amendments conforming with WHO spelling rules. The detailed conventions are listed below. The alphabetical index uses the following conventions:

• Terms will be listed in their singular form
• Apostrophes have been removed For example: ‘Hodgkin lymphoma’ (instead of ‘Hodgkin’s lymphoma’)
• Entities are described using natural language For example: ‘myocardial infarction’ (instead of ‘infarction, myocardial’)
• Abbreviations will be printed using upper case letters, and followed by the complete title in full. For example: ‘MI – myocardial infarction’

Parentheses are used in the tabular list to enclose the code to which an exclusion term refers. For example:

1M7L Infectious blepharitis Exclusions: blepharoconjunctivitis (AD56)

Stem codes

ICD–11 stem codes are codes in a particular tabular list that can be used alone. Stem codes may be entities or groupings of high relevance, or clinical entities that should always be described as one entity. The design of stem codes makes sure that in use cases that require only one code per case a meaningful minimum of information is collected. The stem codes of the ICD-11 are organized in 24 chapters that follow the traditional pattern of ICD, relating to aetiology, relevant organ system, maternal status, perinatal status, external causes, and factors influencing the health status.

Extension codes

The Extension codes have been designed to standardize the way additional information is added to stem codes. Also, the adoption of multi-dimensional coding results in a substantially reduced amount of stem codes. The Extension codes should never be used alone and must always be linked to a stem code. One or more extension codes can be linked when coding a specific condition. Extension codes are provided for use as supplementary or additional codes when it is desired to identify more detail in entities classified elsewhere.

There are two main types of Extension codes. Type 1 extension codes allow the user to add detail to a stem code. The category refers to the same diagnosis with or without the Type I extension code. These extension codes provide important additional information, such as whether a condition is acute or old - and where it is located.

Type 2 extension codes describe the kind of diagnosis; they identify different types of use of the same ICD code in health and other records or for administrative detail. The same ICD code may be used as a main condition, or to specify whether that condition was present on admission, or alternatively, if it arose after admission (see also section ‘Morbidity rules’ for the international definition of main condition). The meaning of the code refers to the same condition, but the extension code alters its interpretation. For example: In both examples above, ‘Other specified benign neoplasms’ refers to the same ICD code, and the Type 1 extension codes refer to the same diagnostic situation for the patient, but the use of the Type 2 extension code provides additional detail informing that the diagnosis of the lipoma was the reason for admission.

Table: Overview of the Extension codes

Type 1	Type 2
Severity	Main Condition Reason for encounter Reason for admission Main Resource Condition
Temporality (course of the condition)	Present on Admission Developed after admission Uncertain timing relative to admission
Temporality (Time in life)	Provisional diagnosis
Aetiology	Diagnosis confirmed by… - lab - serology - histology - genetics - imaging - unspecified means
Anatomic detail Topology Anatomic location	Differential Diagnosis
Histopathology
Biological Indicators
Consciousness
External Causes detail
Injury Specific detail

Using extension codes and cluster coding

Extension codes and codes from other parts of ICD can be linked together to describe a diagnosis in detail. They have to be grouped together in order to not lose the information conveyed by the joint group of codes for one condition in data transmission and evaluation. Such a group of codes is called a cluster. For coding, the following conventions for clustering should be followed:

1. If only one stem code is coded no clustering mechanisms need to be observed.

   Example: stem code

2. If one stem code is clustered together with extension codes such a cluster has to be identified and grouped together using an unambiguous markup that will be chosen based on a specific use of the ICD-11 in a specific setting. In data submission for international reporting, this specific markup has to comply with the following rules: The stem code is to be reported first followed by a ‘&’ followed by the extension codes separated as well by ‘&’.

   Example: stem code&extension code&extension code

3. If two stem codes are clustered, it is important to follow the order according to the use case. For international reporting the first stem code will be separated from the second stem code by a slash (in data analysis for public health prevention, priority will be given to the code that describes the aetiology of a condition, if only one code can be retained).

   Example: stem code/stem code

4. If a stem code is clustered with extension codes and another stem code with some more extension codes, the specific markup should be designed to make a clear distinction of which extension codes in the cluster belong to which stem codes. For international reporting, the following markup has to be followed: The first stem code is reported, followed by a ‘&’ followed by one or more extension codes, each of them separated by ‘&’. Then a slash separates this first section of the cluster from the next stem code which is followed by a ‘&’ and the extension codes for this specific stem code, each again separated by ‘&’.

Example: stem_code & extension_code / stem code & extension_code & extension_code

Clustering is only to be used to combine codes to describe a single diagnosis. If a patient suffers from two independent diagnoses they are not to be reported as clusters. Some type 3 extension codes are used as prefixes (see 3.5.2.2 ). Permissible combinations of stem codes and extension codes are described by sanctioning rules that are embedded in the foundation of ICD–11. They will prevent impossible combinations, and the creation of combinations that already exist in pre-coordination in the tabular list in question.

Special extension codes

The inclusion of the new Extension codes in ICD–11 provides capacity for coding qualifying information of linked stem codes.

“History of” and “Family history of”

Chapter 24 of the classification includes a number of codes that describe both a personal history of various conditions, and a family history of various conditions. If a coder is in a situation of wanting to use ICD-11 to capture such coded information, they have the option of either using the available codes in chapter 24 in isolation, or of using the ‘history of’ (and ‘family history of’) codes, clustered to more specific diagnoses existing elsewhere in the classification.

Example 1: A patient has a personal history of colon cancer that was curatively resected, and a coder wishes to capture this concept.

• Option 1: ‘code for personal history of neoplasm’ (coded alone)
• Option 2: ‘code for personal history of neoplasm’/ ‘code for colon cancer’
• Option 1 simply captures the notion that the patient has a personal history of cancer.

In the coding depicted in option 2, the clustering mechanism has allowed the coder to increase the specificity of information for the history of neoplasm, specifying the type of neoplasm.

Example 2: A patient with a family history of macular degeneration, and a coder wishes to capture this concept.

• Option 1: ‘code for family history of eye disorder’ (coded alone)
• Option 2: ‘code for family history of eye disorder’/ ‘code for macular degeneration’

Note that in both examples 1 and 2, a backslash is used in the clustered coding approach, because the clustered codes are both stem codes in the two examples.

“Present on Admission”

The inclusion of the new Extension codes in ICD–11 provides capacity for coding qualifying information of linked stem codes. Among the new qualifying features is the particularly important status display feature that allows for distinction of diagnoses present at admission from diagnoses arising after hospital stay began. The latter distinction is particularly important, because it allows for the targeted identification of a number of in-hospital diagnoses that may represent adverse events associated with health care. A majority of coded concepts in a hospital record are present at admission. Recognizing this, the most common operational desire in ICD–11 will be to flag a diagnosis that developed after admission.

Example 1:
A patient with long-standing type 1 diabetes, admitted to hospital because of a myocardial infarction.
	Main condition:	Myocardial infarction
	Other condition:	Diabetes mellitus, type 1

In this instance, both conditions are present at admission, but one of them (myocardial infarction) does not need to be coded as being ‘present on admission’ because it is the main condition, designated in this example as being the “reason for admission after assessment at the end of the stay”. The appropriate coding of this scenario therefore includes a combination of two clustered coding entities, each of which involves a stem code linked to an accompanying extension code i.e.: ‘Code for myocardial infarction’/ ‘code for reason for admission assessed at end of stay’ ‘Code for diabetes mellitus type 1’/ ‘code for present on admission’

Note that for both of these coded entities in the above example, a single backslash is used. In the first cluster, the stem code for myocardial infarction is linked to an extension code for main condition diagnosis type. In the second cluster, the stem code for diabetes mellitus type 1 is linked to an extension code for present on admission.

Example 2:
A patient with long-standing type 1 diabetes, admitted to hospital because of myocardial infarction. The patient develops deep vein thrombosis as an in-hospital complication of care.
	Main condition:	Myocardial infarction
	Other condition:	Diabetes mellitus, type 1
		Deep vein thrombosis (arising after hospital stay began)

In this example, an extension code for ‘developed after admission’ is linked by cluster coding to a stem code for ‘deep vein thrombosis’. The first two diagnostic concepts, meanwhile, are coded exactly as per the preceding example. i.e., ‘Code for myocardial infarction’/ ‘code for reason for admission assessed at end of stay’ ‘Code for diabetes mellitus type 1’/ ‘code for present on admission’ ‘Code for deep vein thrombosis’/ ‘code for developed after admission’ Again, each of the three cluster entities uses a ‘&’ because the second code in the cluster is an extension code.

ICD Print and Electronic version

The ICD provides a standard for reporting, coding, selecting, and tabulating conditions for different use cases. It provides guidance on finding the right code from a reported condition. In the electronic version of the ICD, most information is interlinked and visible in the relevant context. Only the content of the Reference Guide should be consulted additionally when coding with ICD-11. In the print version, the information is divided into 3 volumes, the tabular list, the reference guide, and the index. All three are needed to use the ICD correctly.

Volume 1: Tabular List, Special Tabulation Lists, Qualifiers, and Modifiers

Volume 1 contains the Tabular list, which is an alphanumeric listing of diseases and disease groups, inclusion and exclusion notes, and some coding rules. The ICD has 25 chapters, and approximately 15 000 entities at the 4, 5 or 6-character level. In addition, there is a section on extension codes and one on traditional medicine. At the end of Volume 1 the special tabulation lists are presented. These are not designed for coding, but are for tabulation only.

Volume 2: Reference Guide

The Reference Guide contains an introduction to the context, components, and intended use of the ICD. It describes the diverse components of ICD–11, provides guidelines for certification, recording, rules for mortality coding (i.e. causes of death) and morbidity coding (e.g. hospital statistics) and lists for tabulation of statistical data.

Volume 3: Index

The Alphabetical Index is a list of approximately 120 000 clinical terms (including synonyms or phrases). The Index is used to find the relevant ICD codes or code combinations for terms.

Electronic core: The Foundation Component

The Foundation Component is a data source for production and maintenance of Volume 1 and 3, in parts, for Volume 2 as well. It also includes additional content (see ‘content model’) that goes beyond the traditional paper based use of a classification. Depending on the setting within a country it can be decided to use the full Foundation component or to focus on the parts that are essential to production and maintenance of the Index and the Tabular list.

The foundation serves to align the different tabular lists in content and to define the categories. As such it allows standardised use of the ICD-11, independent of the setting in which it is used. The foundation component includes for example links to other classifications or terminologies that can be expanded in the future. Only if relevant for a country this information, or subsets of it, can be used in the application of ICD-11.

Online Tools

The WHO provides the ICD–11 browser for browsing the ICD in multiple languages. This tool allows the user to retrieve concepts by searching terms, anatomy or any other element of the content model. With this browser, users can also contribute to the updating and continuous improvement of ICD with comments and solutions. Such input is reviewed for consideration for inclusion on an annual basis. ICD–11 can also be accessed using web services with user specific software. The IT guide to the ICD provides more details on compatibility requirements. Both, the web services and the online browser allow access to all Tabular lists of the ICD, for mortality and morbidity statistics, primary care, or for a specialty adaptation for certain specialized domains.

Basic Coding and Reporting Guidelines

Coding is the assignment of one or more codes in order to represent the meaning of a condition in as much detail as required. Before attempting to code, the coder should be acquainted with the principles of classification and coding. In some instances, using one code will provide sufficient detail. In other instances, it may be necessary to use several codes together in order to express the level of detail required by the use case, setting, or laws. For coding, users may use a print version of ICD, an online version, or local software.

Finding a stem code

An ICD entity may be: - A category - A group - A chapter

A category (which is the most common reference to an ICD class within a tabular list) may be a disease, disorder or syndrome, a sign or symptom, other health problems such as injuries, or a combination of the above. In addition, the ICD also classifies ‘external causes’ or ‘other reasons for encounter’. A group refers to a set of categories with at least one common property, while a chapter is the highest level of aggregation within the ICD.

Coding step by step – clinical term

The table below compares the coding steps in a paper and an electronic environment. The essential component of coding is finding a match to the reported clinical term – having a good dictionary in the relevant language, and verifying the resulting code against additional rules. In an electronic environment a sanctioning mechanism can verify compliance with the coding rules.

Electronic	Paper
Enter the statement or term in the search window of your user interface	Look up the lead term in the Alphabetical Index and applicable secondary terms.
Select the matching term, or one closest to what you are looking for, among the displayed options	Select the appropriate term, or one closest to what you are looking for, among the listed options
Verify the result in the tabular list browser view for exclusions, inclusions and notes given at the level of that category, its grouping levels and at the chapter level.	Verify the result in the tabular list (Volume 1) for exclusions, inclusions and notes given at the level of that category, its grouping levels and at the chapter level.

WHO has made available a coding tool online for the ICD-11: click here. The WHO online browser is available here. You can also access browser and coding tool through the WHO website: www.who.int/classifications/ICD

Adding Detail – cluster coding with multiple stem codes and extension codes

All cases should be coded in a way to inform about aetiology and the manifestation of the condition of interest. In some instances, the ICD category refers to both, in other instances more than one code needs to be used in order to express the relevant detail. Cluster coding as described in 3.5.1 will be used in these cases. It is necessary to search for the relevant stem codes individually in order to find the correct code. If using only one stem code, search using the full name of the condition in the search field (or the paper index). This name may point to an entity in the foundation that already contains the information regarding what stem code and what Extension code(s) are to be combined in order to code that condition. In the case when there is no such entry, it is necessary to look up the stem code and the Extension code separately. Clusters must never be used to replicate the meaning of a condition with an existing stem code. For example: Because there is a code for fracture of ulna, it is not allowed to build a cluster with the two codes

1. Other specified fracture of forearm and
2. ulna.

There may be less obvious cases across the ICD. Sanctioning rules will help to avoid this kind of mistake, in an electronic environment. For reporting purposes, the correlated codes are linked, using a separator between codes (see ‘3.7.1 ’). In both cases the first code will be the one for the more specific condition prompting the health care contact (i.e. neither the external cause, nor a chronic underlying condition).

If a case has several conditions, it can be coded with one or more codes. Every condition would be a cluster (or linked) in its own right and reported in a separate data field or section.

Special cases ‘multiple’ Some categories refer to multiple parts of the body being affected by one condition, e.g. ‘multiple fractures of…’, or ‘multiple valve disease’. In such cases report first the code with ‘multiple’ and then list in the same cluster all the specific conditions. For example:

Multiple fractures of pelvis/fracture of os pubis/fracture of os sacrum/fracture of os ilium

Special case ‘prosthetic’ Some categories refer to ‘disease of prosthetic…’. In such cases a local coding hint will inform what conditions need to be also coded to be specific about the disease occurring in the context of the prosthesis. For example:

Prosthetic valve disease /Nonrheumatic valve stenosis

Electronic reporting

Electronic documentation will follow the principle of lossless collection of information at the source. Best practice includes: 1. A text field that captures the clinical term or cause of death with the exact wording reported by the health provider 2. A data field that retains the identifier (URI) of the most exact matching chosen term of ICD-11 (index, code title or other element). 3. A data field for the relevant ICD-11 code In this way, the quality of the coding can be verified at any point in time. Also, specific conditions can be identified and analyzed, independently of them being linked to an individual ICD code or lumped together in a code with other conditions.

Main Uses of the ICD: Mortality

This section concerns the rules and guidelines adopted by the World Health Assembly regarding the selection of a single cause or condition for routine tabulation from death certificates. Guidelines are also provided for the application of the rules and for coding of the condition selected for tabulation. Implementation of the ICD for mortality requires setting up an infrastructure for reporting and storing information, designing information flows, quality assurance and feedback, and training for classification users working with the input or output of data.

Mortality statistics

Mortality statistics are widely used for medical research, monitoring of public health, evaluating health interventions and planning and follow-up of health care. Analysis of mortality data typically involves comparisons of data sets, for example those representing different regions or different points in time. Unless the data have been produced by the same methods and according to the same standards, such comparisons will yield misleading results. To standardize production of mortality data, WHO issues international instructions on data collection, coding and classification, and statistical presentation of causes of death. It is of utmost importance that production of mortality data follows the procedures detailed next, since any deviation from the international instructions will impair international comparability. The definition of a single underlying cause of death, and selected approaches to multiple causes of death enables the identification of trends in health for a given population. Analysis of mortality data typically involves comparisons of data sets, for example those representing different regions or different points in time. Unless the data has been produced by the same methods according to the same standards, such comparisons will yield misleading results. The following section contains information on coding causes of death for mortality statistics. It explains the basic concepts, how to code multiple causes, and how to select the underlying cause of death.

The aim of these instructions is to optimize the mortality statistics from a public health point of view. Some of the instructions may appear wrong or questionable from a purely medical perspective. They should still not be set aside, since they may be motivated by well-founded epidemiological and public health principles. If an apparent error is found, it should be reported to WHO. WHO will either explain the rationale or take steps to correct the error at the international level. Individual countries should not correct what is assumed to be an error, since changes at the national level will lead to data that are less comparable to data from other countries, and thus less useful for analysis.

Coding instructions for mortality: underlying cause of death

It was agreed by the Sixth Decennial International Revision Conference that the cause of death for primary tabulation should be designated the underlying cause of death. From the standpoint of prevention of death, it is necessary to break the chain of events or to effect a cure at some point. The most effective public health objective is to prevent the precipitating cause from operating. For this purpose, the underlying cause has been defined as ‘(a) the disease or injury which initiated the train of morbid events leading directly to death, or (b) the circumstances of the accident or violence which produced the fatal injury’. However, for some diseases or injuries special rules apply.

Sections 6.1–6.3 contain instructions on coding causes of death for mortality statistics. The first section, 6.1, explains the basic concepts, section 6.2 explains how to identify the underlying cause of death, and section 6.3 gives further details on how to code multiple causes of death.

The international death certificate

The international mortality coding instructions presuppose that data have been collected with a death certificate conforming to the International form of medical certificate of cause of death (see Annex 14.1). Otherwise, the causes of death cannot be coded according to the international standard and the data will not be internationally comparable. For example, some coding instructions apply to conditions reported as caused by certain other conditions, and in such cases it is important to have a clear distinction between causes reported in Part 1 and in Part 2 of the certificate. Further, information reported elsewhere on the certificate, such as manner of death or whether pregnancy contributed to the death, is essential when assigning multiple cause codes to the conditions stated on the certificate.

It is the responsibility of the medical practitioner or other qualified certifier signing the death certificate to indicate which morbid conditions led directly to death and to state any antecedent conditions giving rise to this cause. The certifier should use his or her clinical judgement in completing the medical certificate of cause of death. Automated systems must not include lists or other prompts to guide the certifier, as these necessarily limit the range of diagnoses and therefore have an adverse effect on the accuracy and usefulness of the report.

The medical part of the form is split into two parts: Part 1 is for diseases related to the train of events leading directly to death, and Part 2 is for unrelated but contributory conditions. On the certificate, all additional data that are necessary to code the correct underlying cause should be recorded, and the form (see Annex 14.1) indicates which other information should be collected. In order to align the way this information is collected internationally, the form should be followed as closely as possible. The information can then be used for manual or electronic coding of the underlying and multiple causes of death.

Basic concepts

Mortality coders must be familiar with the basic concepts introduced in this section.

Sequence

The term ‘sequence’ refers to a chain or series of medical events in which each step is a complication of, or is caused by, the previous step.

|Example 1|||| |--|--|--| ||1|(a)| Myocardial infarction| ||||due to| |||(b) |Coronary thrombosis| ||||due to| |||(c)| Coronary atherosclerosis|

The myocardial infarction is caused by the coronary thrombosis, which, in its turn, is a complication of coronary atherosclerosis. Consequently, the sequence is myocardial infarction caused by coronary thrombosis caused by coronary atherosclerosis.

|Example 2:|||| |--|--|--| ||1|(a)| Extensive haemorrhage| ||||due to| |||(b) |Traumatic amputation of right leg| ||||due to| |||(c)| Run over by street car|

The haemorrhage is a complication of the traumatic amputation, which, in its turn, is caused by the street car accident. Consequently, the sequence is extensive haemorrhage caused by traumatic amputation of the right leg caused by being run over by a street car.

Causal relationship

A causal relationship exists if a condition mentioned on the certificate can be caused by another condition also mentioned on the certificate. However, whether a causal relationship is considered acceptable or not for mortality coding is founded not only on a medical assessment but also on epidemiological and public health considerations. Therefore, a medically acceptable relationship might be listed as unacceptable in the coding instructions, because a later step in the sequence is more important from a public health point of view.

Therefore, to decide whether a stated causal relationship is acceptable or not, first check the instructions in Section 6.2.3, Special instructions on accepted and rejected sequences. Stated relationships that are not listed in Section 6.2.3 should be accepted as far as possible, because the certifier’s opinion about the causes leading to death should not be disregarded lightly. If a stated relationship seems highly improbable, refer to internationally recognized decision tables for mortality coding. A reported sequence that appears improbable should be accepted if one or more intervening steps would explain the causal relationship. For example, if haematemesis is stated as due to cirrhosis of the liver, assume that the haematemesis was caused by ruptured oesophageal varices, the varices were caused by portal hypertension, and the portal hypertension by liver cirrhosis. Such assumed intervening causes must not be used to modify the coding. Note that a condition A can never be caused by a condition B if condition A has a longer duration or earlier onset than condition B.

Duration

On death certificates, each reported condition should also include information about duration. The duration refers to the time period between the onset of the disease or condition and the time of death. Note that it is not always the same as the time of diagnosis of the condition, which may be at the same time as, or after, the onset of symptoms.

Terminal cause of death

The terminal cause of death is the condition entered first on the first line of Part 1 of the death certificate.

Example 3:
	(a)	Myocardial infarction and pulmonary oedema
		due to
	(b)	Coronary atherosclerosis
> Myocardial infarction is the terminal cause of death, since it is entered first on the first line of the certificate.

Starting point

The starting point is the condition or event that started the sequence of acceptable causal relationships ending with the terminal cause of death. In a correctly completed certificate, the condition reported on the lowest used line in Part 1 is the starting point of the sequence.

Example 4:
	(a)	Myocardial infarction and pulmonary oedema
		due to
	(b)	Coronary atherosclerosis

Example 5:
	(a)	Pneumonia
		due to
	(b)	Hip fracture
		due to
	(c)	Tripped on carpet

Tripped on carpet is the starting point, since it started the sequence of events leading to death.

Tentative starting point

In a correctly completed certificate, the condition reported on the lowest line in Part 1 is the starting point, but if the certificate is not correctly filled out, the starting point may be reported somewhere else. The instructions on how to identify the starting point in such cases are complex. Sometimes, several instructions apply to the same death certificate, and it is important to apply the instructions step by step as described in Section 6.2.1, ‘Find the starting point’. In each step where a tentative starting point is identified, a condition that is provisionally considered as the starting point but that, in later steps, might turn out to be caused by something else. The tentative starting point may change several times as the instructions are applied to the certificate.

Also, take additional information on causal relationships that the certifier has provided into account. This applies also if the information appears in the ‘wrong’ place of the certificate. For example, if the sequence in Part 1 starts with a disease A, and information elsewhere on the certificate states that this disease A was due to a disease B, then consider B as the tentative starting point.

Obvious cause

Several coding instructions will instruct you to check whether the tentative starting point is itself obviously caused by another condition mentioned on the same line or below on the certificate. The word ‘obviously’ is important, and there must be no doubt about the relationship between the conditions. Further instructions are given in Section 6.2.1, Step SP6 – Obvious cause, and in Section 6.2.4, Special instructions on obvious cause (Step SP6).

Example 6:
	1(a)	Sepsis
due to
	(b)	Peritonitis
	2	Appendicitis with rupture

Peritonitis started the sequence of events reported in Part 1, so it is the tentative starting point. However, appendicitis with rupture is an obvious cause of peritonitis. Therefore, the sequence of events starts with appendicitis, which consequently is the starting point of the sequence of events ending with sepsis, the terminal cause of death.

First-mentioned sequence

A death certificate may contain several sequences, and the coding instructions will tell you to find the starting point of the first-mentioned sequence. To identify the starting point of the first-mentioned sequence, begin with the terminal cause of death (the first-mentioned condition on the uppermost line in Part 1). Establish whether the first condition listed on the next line in Part 1 can cause the terminal cause of death. If it cannot, and if there are more conditions on the line, establish whether the second condition listed on this line can cause the terminal cause of death. Continue until you have found a condition that could cause the terminal cause of death. This is the tentative starting point of the sequence.

If no condition on the next line can cause the terminal cause of death, there is no sequence ending with the terminal cause of death.

If you found a tentative starting point but there are conditions reported on lower lines in Part 1, repeat the procedure for the next line. Start with the tentative starting point you identified in the previous step. Establish whether the first condition listed on the next lower line in Part 1 can cause the tentative starting point. If it cannot, and if there are more conditions on the line, check whether the second condition listed on that line can cause the tentative starting point. Continue until you have found a condition that could cause the tentative starting point. This is the new tentative starting point.

If there are still conditions reported on lower lines in Part 1, repeat the procedure for as long as a new tentative starting point can be identified. When no condition can be found that could cause the tentative starting point, the last identified tentative starting point is also the starting point of the first-mentioned sequence. The figure illustrates examples of certificates with several sequences. The starting point of the first-mentioned sequence is in grey, with a bold black circle.

Figure 2: Sequences

Example 7:
	1(a)	Pneumonia
due to
	(b)	Hip fracture and heart failure
due to
	(c)	Tripped on carpet, coronary atherosclerosis

Pneumonia can be due to hip fracture, and therefore hip fracture is the tentative starting point. Hip fracture can be due to tripping, which is the new tentative starting point. Since there are no causes reported below line 1(c), tripping on carpet is the starting point of the first-mentioned sequence.

Example 8:
	(a)	Pneumonia
		due to
	(b)	Heart failure and hip fracture
		due to
	(c)	Coronary atherosclerosis and tripped on carpet

Pneumonia can be due to heart failure, and therefore heart failure is the tentative starting point. Heart failure can be due to coronary atherosclerosis, which is the new tentative starting point. Since there are no causes reported below line 1(c), coronary atherosclerosis is the starting point of the first-mentioned sequence.

Example 9:
	(a)	Pneumonia
		due to
	(b)	Hip fracture and heart failure
		due to
	(c)	Coronary atherosclerosis and tripped on carpet

Pneumonia can be due to hip fracture, and therefore hip fracture is the tentative starting point. However, hip fracture cannot be due to coronary atherosclerosis but hip fracture can be due to tripping, which is the new tentative starting point. Since there are no causes reported below line 1(c), tripped on carpet is the starting point of the first-mentioned sequence.

First-mentioned condition

Some coding instructions refer to the ‘first-mentioned’ condition. When identifying the first-mentioned condition, start from the top line of Part 1 downwards, and from left to right. [Note for translators: If the local language is not written from left to right and from top to bottom, adapt the instruction so that it agrees with the direction of writing.]

Underlying cause of death

Most, but not all, mortality statistics show a single cause of death for each individual, regardless of how many conditions are reported on the certificate. The underlying cause of death is the condition selected for such single-cause tabulation. In most cases, the underlying cause of death is the same as the starting point. However, sometimes a condition other than the starting point is selected as underlying cause of death for use in the statistics. See also ‘Modification’, next.

Example 10:
	(a)	Bronchopneumonia
		due to
	(b)	Hemiplegia
		due to
	(c)	Cerebral infarction

Cerebral infarction started the sequence of events leading to death, so it is the starting point. In this case, it is also the underlying cause of death.

Modification

Special coding instructions on specific sequences and ICD categories may have the effect that a condition other than the starting point is selected as the underlying cause of death for use in the statistics. In such cases, the code for underlying cause often expresses a combination of the starting point with another reported condition, or a complication or consequence of the starting point that is of particular importance to public health. The procedure by which the ICD code for the starting point is replaced by another code is called modification.

Example 11:
	(a)	Heart disease
		due to
	(b)	Generalized atherosclerosis

Generalized atherosclerosis started the sequence of events leading to death, so it is the starting point. However, according to a special instruction on generalized atherosclerosis, generalized or unspecified atherosclerosis leading to heart disease is assigned to atherosclerotic heart disease in mortality statistics. Because of this modification, atherosclerotic heart disease is the underlying cause of death.

Tentative underlying cause of death

Several special instructions on modification may apply to the same death certificate. If so, apply the instructions step by step. The code selected as the outcome of each step in the process is called the tentative underlying cause of death.

Example 12:
	(a)	Myocardial infarction
		due to
	(b)	Coronary atherosclerosis
		due to
	(c)	Generalized atherosclerosis

Generalized atherosclerosis started the sequence of events leading to death, so it is the starting point. There are special modification instructions relating to atherosclerosis and coronary heart disease in the ICD, and, in the next step, coronary atherosclerosis is selected as the tentative underlying cause of death. But there are further instructions on coronary atherosclerosis and myocardial infarction, and in the final step, myocardial infarction is selected as the underlying cause.

Coding instructions for mortality: selecting the underlying cause of death

When coding and classifying causes of death, you must first assign ICD codes to all the conditions mentioned on the death certificate. Many coding instructions are based on specific ICD codes and, to determine whether any of the instructions apply, you need to know the ICD codes for all conditions on the certificate. This is called multiple-cause coding (see section 6.3, Coding instructions for mortality: multiple causes). Next, you select an underlying cause of death to be used in the mortality statistics. This is called classification of the underlying cause of death.

For most death certificates, selecting the underlying cause of death is a fairly uncomplicated procedure. There are, however, many cases where the underlying cause is not immediately obvious. To ensure that both straightforward and complex cases are coded according to international regulations, it is important to follow the coding instructions carefully, step by step. Otherwise, the resulting mortality statistics will not be internationally comparable, which seriously reduces the value of the data for public health purposes.

Selecting the underlying cause of death involves two separate steps. First, you identify the starting point – the disease or event that started the chain of events leading to death. Next, you check whether any special instructions apply to the starting point you identified. If so, the next step is to modify the starting point you identified in the first step.

Note that the purpose of the selection procedure is to produce the most useful mortality statistics possible. Thus, the following instructions may reflect importance for public health rather than what is correct from a purely medical point of view. The following instructions always apply, whether they might be considered medically correct or not.

In the coding examples that follow, the ‘due to’ statement between the lines in Part 1 is no longer included. Still, it is important to bear in mind that anything reported on an upper line in Part 1 is meant to be due to what is reported on the line below.

Find the starting point (Steps SP1 to SP8)

To identify the starting point, follow the eight steps specified in this section. The steps are named SP1 to SP8 (Starting point rule 1 to Starting point rule 8). Each step contains one selection rule. At each step, there is a description of the selection rule itself and an instruction on what to do next. For some of the rules, there are also bullet points with more detailed instructions.

Step SP1 – Single cause on certificate

If there is only one condition reported on the certificate, in either Part 1 or Part 2, this is the starting point and it is also the underlying cause. Next, go to Step M4. If there are two or more conditions on the certificate, go to Step SP2.

Step SP2 – Only one line used in Part 1

If the certifier has used only one line in Part 1 but entered two or more conditions on this line, then the first-mentioned condition is the tentative starting point. Next, go to Step SP6.

Also, if there is only one condition reported in Part 1 but one or more conditions in Part 2, then the single condition in Part 1 is the tentative starting point. Next, go to Step SP6.

If the certifier has used more than one line in Part 1, go to Step SP3.

Example 1:
	1(a)	Myocardial infarction and diabetes mellitus
	(b)
	(c)
	(d)
	2

Myocardial infarction is mentioned first on the certificate and is the tentative starting point. Next, go to Step SP6, to check whether further selection and modification rules apply.

Example 2:
	1(a)	Myocardial infarction
	(b)
	(c)
	(d)
	2	Diabetes mellitus

Myocardial infarction is mentioned first on the certificate and is the tentative starting point. Next, go to Step SP6, to check whether further selection and modification rules apply.

Step SP3 – More than one line used in Part 1, first cause on lowest line explains all entries above

If there are conditions reported on more than one line in Part 1, check whether all of the conditions reported on the line(s) above the lowest used line in Part 1 can be caused by the first condition on the lowest used line. If all conditions on the line(s) above the lowest used line in Part 1 can be caused by the first condition on the lowest used line, then this condition is – tentatively – the starting point. Next, go to Step SP6. If all conditions on the line(s) above the lowest used line in Part 1 cannot be caused by the first condition on the lowest used line, try to get clarification from the certifier. If no further information is available, go to Step SP4. At Step SP3, it is not necessary to assess the causal relationships between conditions reported on the lines above the lowest used line. It is sufficient that each one of the conditions on the lines above the lowest used line can be due to the condition reported first on the lowest used line. At Step SP3, there is no requirement that the conditions entered above the lowest used line have successively longer durations from the top line downwards. The condition mentioned first on the lowest used line may still have caused all conditions reported on the lines above, as long as none of them has a duration that is longer than that of the condition mentioned first on the lowest used line. - Note that whether a causal relationship is listed as correct or not may reflect importance for public health rather than what is acceptable from a purely medical point of view. Therefore, check the instructions in Section 6.2.3, Special instructions on accepted and rejected sequences, first. Always follow the instructions in Section 6.2.3, whether they appear to be medically correct or not. - Stated relationships that are not listed as rejected in Section 6.2.3 should be accepted, as far as possible. They reflect the certifier’s opinion about the causes leading to death and should not be disregarded lightly. - If a stated relationship appears highly improbable, refer to internationally recognized decision tables for mortality coding.

Example 3:
	1(a)	Bronchopneumonia
	(b)	Hemiplegia
	(c)	Cerebral infarction
	(d)
	2

Both bronchopneumonia and hemiplegia can be caused by cerebral infarction. This means that cerebral infarction is the tentative starting point.

Example 4:
	1(a)	Kaposi sarcoma 1 year
	(b)	HIV 3 years
	(c)	Blood transfusion 5 years
	(d)	Haemophilia since birth
	2

Kaposi sarcoma, HIV and blood transfusion can all be caused by haemophilia, which is the first (and also only) condition mentioned on the lowest used line in Part 1. This means that haemophilia is the tentative starting point.

Example 5:
1(a)	Pneumocystosis 6 months
	(b)	HIV 5 years
	(c)	Ruptured spleen 7 years
	(d)	Assault – fist fight 7 years
	2

Assault by fist fight is the only condition mentioned on the lowest used line in Part 1. It can cause everything on the lines above, assuming a blood transfusion as treatment for the ruptured spleen. See also Section 6.1.3, Basic concepts, where assumption of intervening cause is described in the section on causal relationship.

Example 6:
	1(a)	Liver metastases 2 months
	(b)	Bronchopneumonia 4 days
	(c)	Stomach cancer 6 months
	(d)
	2

Both liver metastases and bronchopneumonia can be caused by stomach cancer. This means that stomach cancer is the tentative starting point, even though bronchopneumonia cannot cause liver metastases and the bronchopneumonia has a shorter duration than the liver metastases.

Example 7:
	1(a)	Liver metastases and pulmonary oedema
	(b)	Bronchopneumonia
	(c)	Stomach cancer
	(d)
	2

Liver metastases, pulmonary oedema and bronchopneumonia can all be caused by stomach cancer. This means that stomach cancer is the tentative starting point, even though bronchopneumonia cannot cause liver metastases.

|Example 8:|||| |--|--|--| ||1|(a) | Liver metastases 2 months | |||(b) | Bronchopneumonia 4 days | |||(c) | Stomach cancer and cerebral infarction 6 months |

Both liver metastases and bronchopneumonia can be caused by stomach cancer, which is the first condition mentioned on the lowest used line in Part 1. This means that stomach cancer is the tentative starting point, even though bronchopneumonia cannot cause liver metastases, and bronchopneumonia has a shorter duration than the liver metastases.

|Example 9:|||| |--|--|--| ||1|(a)| Liver metastases| |||(b)| Bronchopneumonia and stomach cancer|

Liver metastases cannot be due to bronchopneumonia. This means that no tentative starting point can be identified at Step SP3. Therefore, go to Step SP4.

Step SP4 – First cause on lowest used line does not explain all entries above, but a sequence ends with the terminal condition

If there is only one sequence ending with the terminal condition, find the starting point of this sequence. This is the new tentative starting point. Next, go to Step SP6. If there are two or more sequences of conditions or events ending with the terminal condition, identify the first-mentioned sequence as described in Section 6.1.3, and find the starting point of this first-mentioned sequence. Next, go to Step SP6. If there is no sequence ending with the terminal condition, go to Step SP5.

• As mentioned under Step SP3, always follow the instructions in Section 6.2.3, whether they appear to be medically correct or not.
• Stated relationships that are not listed as rejected in Section 6.2.3 should be accepted as far as possible. They reflect the certifier’s opinion about the causes leading to death and should not be disregarded lightly.
• If a stated relationship appears highly improbable, refer to internationally recognized decision tables for mortality coding.
• When evaluating a sequence, also remember that, according to Section 6.2.3, Special instructions on accepted and rejected sequences, a condition A can never be caused by a condition B if condition A has a longer duration than condition B.

|Example 10:|||| |--|--|--| ||1|(a)| Liver metastases 2 months| |||(b)| Cerebral infarction and stomach cancer 6 months|

Cerebral infarction cannot cause liver metastases, but liver metastases can be due to stomach cancer. Stomach cancer is the tentative starting point.

|Example 11:|||| |--|--|--| ||1|(a)| Bronchopneumonia 2 months| |||(b)| Cerebral infarction and liver metastases 6 months| |||(c)| Atherosclerosis and stomach cancer|

Atherosclerosis cannot cause liver metastases. However, there are three acceptable sequences on the certificate: (1) bronchopneumonia caused by cerebral infarction, in its turn caused by atherosclerosis; (2) bronchopneumonia caused by cerebral infarction, in its turn caused by stomach cancer; and (3) bronchopneumonia caused by liver metastases, in its turn caused by stomach cancer. But the first-mentioned sequence is bronchopneumonia caused by cerebral infarction, in its turn caused by atherosclerosis. Consequently, atherosclerosis is the tentative starting point.

Step SP5 – No sequence in Part 1

If there is no sequence ending with the terminal condition, then the terminal condition is also the tentative starting point. Next, go to Step SP6.

|Example 12:|||| |--|--|--| ||1|(a)| Liver metastases |||(b)| Cerebral infarction| |||(c)| Atherosclerosis| ||2|| Stomach cancer|

Atherosclerosis cannot cause liver metastases. Also, there is no sequence in Part 1 that ends with the terminal condition, because cerebral infarction cannot cause liver metastases. Because there is no sequence ending with the terminal condition, the terminal condition itself – liver metastases – is the tentative starting point.

Step SP6 – Obvious cause

Now check whether the tentative starting point you selected in Steps SP1 to SP5 obviously was caused by another condition on the certificate. If the tentative starting point is in Part 1, then this other condition must be either on the same line, further down in Part 1, or in Part 2. If the tentative starting point is in Part 2, this other condition must also be in Part 2.

Next, check whether there is another condition mentioned on the same line or further down on the certificate as the new tentative starting point you just identified that obviously caused this new tentative starting point. Continue looking for a new tentative starting point until you find a starting point that is not obviously caused by a condition reported on the same line or further down on the certificate. Then go to Step SP7. If there is no condition mentioned on the certificate that obviously caused the tentative starting point you selected in Steps SP1 to SP5, go to Step SP7.

• If the tentative starting point is in Part 1, look for an obvious cause of the tentative starting point first on the same line in Part 1, next on lower lines in Part 1, and finally in Part 2. Do not look for obvious causes on lines above the tentative starting point.
• If the tentative starting point is in Part 2, look for an obvious cause in Part 2. Do not look for obvious causes in Part 1.
• If a condition A has a longer duration than a condition B, then condition B cannot be the obvious cause of condition A.
• If there are several conditions that could be obvious causes of the tentative starting point, select the first-mentioned condition.
• ‘Obvious cause’ means that there must be no doubt that the tentative starting point was caused by the other condition mentioned on the certificate. It is not sufficient that the sequence would have been accepted if the tentative starting point had been reported as due to the other condition.
• Refer to Section 6.2.4, Special instructions on obvious cause (Step SP6), for further instructions. Note that whether a condition B is considered an obvious cause of a condition A may reflect importance for public health rather than what is motivated from a purely medical point of view. Therefore, always follow the instructions in Section 6.2.4, whether they appear to be medically correct or not.

|Example 13:|||| |--|--|--| ||1|(a)| Liver metastases |||(b)| Cerebral infarction| ||2|| Stomach cancer|

Cerebral infarction cannot cause liver metastases, and liver metastases is the tentative starting point. But stomach cancer is an obvious cause of liver metastases, and stomach cancer is the new tentative starting point.

|Example 14:|||| |--|--|--| ||1|(a)| Sepsis |||(b)| Peritonitis| ||2|| Necrosis of intestine, mesenteric infarction|

2 Necrosis of intestine, mesenteric infarction

Sepsis can be caused by peritonitis, and peritonitis is the tentative starting point. But necrosis of intestine is an obvious cause of peritonitis, so necrosis of intestine is the new tentative starting point. Next, mesenteric infarction is an obvious cause of necrosis of intestine, and mesenteric infarction is the final starting point.

|Example 15:|||| |--|--|--| ||1|(a)| Sepsis| |||(b)| Peritonitis| ||2|| Mesenteric embolism, ruptured appendicitis|

Sepsis can be caused by peritonitis, and peritonitis is the tentative starting point. Next, both mesenteric embolism and ruptured appendicitis are obvious causes of peritonitis. Because mesenteric embolism is mentioned first, it is the new tentative starting point.

Step SP7 – Ill-defined conditions

Now check whether the tentative starting point is listed in the table of ill-defined conditions (see Annex 14.3, List of ill-defined conditions). If it is, the tentative starting point is considered ill-defined. Then do as follows:

• If there are other conditions reported on the certificate, check whether they are all ill-defined. If all other conditions are ill-defined, go to Step M1.
• If there is at least one condition that is not ill-defined, then disregard the ill-defined condition. Go to Step SP1 and select another starting point, as if the ill-defined condition had not been mentioned on the certificate.
• If the tentative starting point is not ill-defined, go to Step SP8.

Note that Septic shock, Systemic inflammatory response syndrome of infectious origin without organ failure, Systemic inflammatory response syndrome of infectious origin with organ failure and Sudden infant death syndrome are not considered ill-defined. In some cases, the ill-defined condition may have an impact on how other conditions on the certificate are coded. If so, disregard the ill-defined condition when selecting the starting point, but take it into consideration when coding the other conditions on the certificate.

|Example 16:|||| |--|--|--| ||1|(a)| Respiratory failure| ||2|| Mesenteric embolism|

Respiratory failure is the only condition mentioned in Part 1 and it is the tentative starting point according to Steps SP2 and SP6. But respiratory failure is in the table of ill-defined conditions, so disregard respiratory failure and restart the selection procedure from Step SP1. Mesenteric embolism is the new starting point according to Step SP1.

|Example 17:|||| |--|--|--| ||1|(a)| Anaemia| |||(b)| Splenomegaly|

Splenomegaly, the tentative starting point according to Step SP3, is in the table of ill-defined conditions. Disregard splenomegaly and restart the selection procedure from Step SP1. Now, anaemia is the new starting point according to Step SP2. However, splenomegaly modifies the coding of anaemia (see the Alphabetical index). Code to ‘splenomegalic anaemia’.

Step SP8 – Conditions unlikely to cause death

Next, check whether the tentative starting point is listed in the table of conditions unlikely to cause death (see Annex 14.4, List of conditions unlikely to cause death). If it is, do as follows:

• If there are other conditions reported on the certificate, check whether they are all ill-defined or unlikely to cause death. If they are all ill-defined or unlikely to cause death, go to Step M1.
• If there are other conditions reported that are not ill-defined or unlikely to cause death, first check whether the death was caused by a reaction to treatment of the condition unlikely to cause death that you selected as the tentative starting point. If it was, then select the reaction to treatment as the starting point. Next, go to Step M1.
• If the death was not caused by a reaction to treatment of the condition unlikely to cause death, check whether the condition was the cause of another condition that is not on the list of conditions unlikely to cause death and that is not ill-defined. If it was, then the condition unlikely to cause death is still the tentative starting point. Next, go to Step M1.
• If there was no reaction to treatment and no complication of the condition unlikely to cause death, then disregard the condition unlikely to cause death. Go to Step SP1 and select another starting point, as if the condition unlikely to cause death had not been mentioned on the certificate.
• If the certificate mentions several treatments for the condition unlikely to cause death, select the initial treatment.
• ‘Complication’ means a condition that can be due to the condition unlikely to cause death, or due to the treatment of the condition unlikely to cause death.
• If the starting point is not a condition unlikely to cause death, then go to Step M1.

|Example 18:|||| |--|--|--| ||1|(a)| Hearing loss| ||2|| Ischaemic heart disease|

Hearing loss is the tentative starting point according to Step SP2, but hearing loss is in the table of conditions considered unlikely to cause death. There is another condition on the certificate, ischaemic heart disease, which is not in the table of conditions considered unlikely to cause death. Disregard hearing loss and restart the selection procedure from Step SP1. Ischaemic heart disease is the new starting point according to Step SP1.

|Example 19:|||| |--|--|--| ||1|(a)| Liver failure | |||(b)| Excessive use of paracetamol | |||(c) | Migraine type headache|

Migraine type headache is the tentative starting point according to Step SP3. It is in the table of conditions considered unlikely to cause death. The condition was treated with paracetamol and there was a reaction to the treatment, liver failure. Disregard the condition unlikely to cause death and select the reaction to the treatment, liver failure, as the starting point.

|Example 20:|||| |--|--|--| ||1|(a)| Sepsis | |||(b)| Submandibular abscess| |||(c)| Caries |

Caries is the tentative starting point according to Step SP3. It is in the table of conditions considered unlikely to cause death, but in this case it caused complications that are not considered unlikely to cause death. Because of that, select caries as the starting point.

|Example 21:|||| |--|--|--| ||1|(a)| Headache | |||(b)| Caries| ||2|| Ischaemic heart disease|

Caries is the tentative starting point according to Step SP3. It is in the table of conditions considered unlikely to cause death. A complication is reported, headache, but it is in the table of ill-defined conditions. Disregard both caries and headache and restart the selection procedure from Step SP1. Ischaemic heart disease is the new starting point according to Step SP1.

Check for modifications of the starting point (Steps M1 to M4)

The starting point you identified using Steps SP1 to SP8 is now considered the tentative underlying cause. There may be special coding instructions on this tentative underlying cause, or other reasons to modify the tentative underlying cause. Check whether the tentative underlying cause should be modified by applying the modification rules described in steps M1 to M3 (Modification rule 1 to Modification rule 3). Each step contains one modification rule. At each step, there is a description of the modification rule itself and what to do next. There are also bullet points with more detailed instructions and explanations.

Step M1 – Special instructions

Check whether special coding instructions apply to the tentative underlying cause. If a special coding instruction applies, assign a new tentative underlying cause according to the instruction.

Next, check whether any special instructions apply to this new tentative underlying cause. That is, reapply Step M1. Repeat until you have found a tentative underlying cause that is not affected by any further special coding instruction. Next, go to Step M2.

• Refer to Section 6.2.5, Special instructions on linkages and other provisions (Step M1), for detailed instructions on specific tentative underlying causes.
• According to some of these special instructions, the tentative underlying cause combines with another cause of death reported on the death certificate, into a new tentative underlying cause. If there are several such combinations that would apply to the tentative underlying cause, then apply the combination with the first-mentioned of these other conditions (the first-mentioned linkage).
• Note that some special instructions only apply under specific circumstances, for example where a condition A is reported as the cause of a condition B, or to deaths at a specific age.
• Sometimes Volume 1 or the Alphabetical index indicates a code for a combination of the tentative underlying cause with another cause mentioned on the certificate. Use the combination code only if the code title clearly indicates the aetiology of the condition. If no special coding instruction applies, then the starting point you found using Steps SP1 to SP8 is the tentative underlying cause. Next, go to Step M2.

|Example 1:|||| |--|--|--| ||1|(a)| Myocardial infarction | |||(b)|Ischaemic heart disease|

Ischaemic heart disease is the tentative starting point according to Step SP3. There is a special instruction on ischaemic heart disease reported with myocardial infarction, and, according to this instruction, myocardial infarction is the new tentative underlying cause.

|Example 2:|||| |--|--|--| ||1|(a)|Ischaemic heart disease| |||(b)| Atherosclerosis| ||2||Myocardial infarction|

Atherosclerosis is the tentative starting point according to Step SP3. There is a special instruction on atherosclerosis reported with ischaemic heart disease, and another one on atherosclerosis reported with myocardial infarction. Ischaemic heart disease is reported first on the certificate, so apply the instruction on atherosclerosis reported with ischaemic heart disease and select ischaemic heart disease as the new starting point. Next, there is a special instruction on ischaemic heart disease reported with myocardial infarction. Apply this instruction and select myocardial infarction as the new tentative underlying cause.

|Example 3:|||| |--|--|--| ||1|(a)|Ischaemic heart disease| |||(b)| Atherosclerosis| ||2| |Cerebral infarction|

Atherosclerosis is the tentative starting point according to Step SP3. There is a special instruction on atherosclerosis reported with ischaemic heart disease, and another one on atherosclerosis reported with cerebral infarction. Ischaemic heart disease is reported first on the certificate, so apply the instruction on atherosclerosis reported with ischaemic heart disease and select ischaemic heart disease as the new tentative underlying cause.

|Example 4:|||| |--|--|--| ||1|(a)|Cerebrovascular infarction| |||(b)| Atherosclerosis| |||(c)| Hypertension ||2||Myocardial infarction|

Hypertension is the tentative starting point according to Step SP3. There are special instructions on hypertension reported with cerebrovascular infarction and with myocardial infarction. Cerebrovascular infarction is reported first on the certificate, so apply the instruction on hypertension reported with cerebrovascular infarction and select cerebrovascular infarction as the new tentative underlying cause.

|Example 5:|||| |--|--|--| ||1|(a)|Dementia| |||(b)| Atherosclerosis|

Atherosclerosis is the tentative starting point according to Step SP3. There is a special instruction on atherosclerosis reported as the cause of dementia. Apply this instruction and select atherosclerotic dementia as the new tentative underlying cause.

|Example 6:|||| |--|--|--| ||1|(a)|Atherosclerosis| ||2||Dementia|

Atherosclerosis is the tentative starting point according to Step SP2. Although there is a special instruction on dementia reported as caused by atherosclerosis, this instruction does not apply here because dementia is reported in Part 2 and not as caused by atherosclerosis. In this case, atherosclerosis remains the tentative starting point.

|Example 7:|||| |--|--|--| ||1|(a)|Epilepsy| |||(b)| Alcoholism|

Alcoholism is the tentative starting point according to Step SP3. In Volume 1, a list of inclusion terms at G40.5, Special epileptic syndromes, mentions ‘epileptic seizures related to alcohol’. However, the code title for Special epileptic syndromes, does not mention alcohol. Therefore, keep alcoholism as the tentative starting point.

Step M2 – Specificity

If the tentative underlying cause describes a condition in general terms and a term that provides more precise information about the site or nature of this condition is reported on the certificate, this more informative term is the new tentative underlying cause. Next, check whether this new tentative underlying cause can be specified even further by other terms on the death certificate. That is, reapply Step M2. Repeat until you have found a tentative underlying cause that cannot be specified further.

• The more specific description must refer to the same condition as the tentative underlying cause. Do not disregard a generalized condition such as atherosclerosis because a more specific but unrelated condition is reported on the certificate (see also Example 9).
• Note that the new tentative underlying cause itself is sometimes specified further by the general term (see Example 10).
• If several other expressions on the certificate provide more precise information on the tentative underlying cause, start with the first-mentioned of these other conditions.
• Note that some instructions on specificity only apply under specific circumstances, for example where a condition A is reported as the cause of a condition B.

|Example 8:|||| |--|--|--| ||1|(a)| Cerebrovascular accident| |||(b)| Atherosclerosis| ||2| |Arterial embolism to brain stem|

Atherosclerosis is the tentative starting point according to Step SP3. There is a special instruction on atherosclerosis reported with cerebrovascular accident; apply this instruction and select cerebrovascular accident as the new starting point according to Step M1. The type of cerebrovascular accident is described more precisely in Part 2 as an arterial embolism to brain stem. This is the new tentative underlying cause.

Example 9:
	1	(a)	Cerebrovascular accident
		(b)	Atherosclerosis
	2		Oat cell cancer originating in upper right lobe

Atherosclerosis is the tentative starting point according to Step SP3. There is a special instruction on atherosclerosis reported with cerebrovascular accident; apply this instruction and select cerebrovascular accident as the new tentative underlying cause. There is no more specific description of the type of cerebrovascular accident on the certificate, and cerebrovascular accident remains the tentative underlying cause.

|Example 10:|||| |--|--|--| ||1|(a)| Meningitis| |||(b)| Tuberculosis|

Tuberculosis is the tentative starting point according to Step SP3. The manifestation is described as meningitis, and the two terms combine into tuberculous meningitis, which is the tentative underlying cause.

Step M3 – Recheck Steps SP6, M1 and M2

If, at this point, the tentative underlying cause is not the same as the starting point you selected using Steps SP1 to SP8, then go back to Step SP6. Repeat the procedures described in Steps SP6, M1 and M2. - Do not go back to Step SP6 if the cause selected in Step M1 or M2 is correctly reported as due to another condition, except when this condition is ill-defined. - Also, do not go back to Step SP6 if the tentative underlying cause is a reaction to treatment of a condition unlikely to cause death, as selected in Step SP8.

|Example 11:|||| |--|--|--| ||1|(a)| Sepsis | |||(b)|Arterial disease, arterial embolism of left leg| ||2|| Colon cancer |

Arterial disease is the tentative starting point according to Step SP3. Arterial embolism of left leg, reported as the second condition on line 1(b), is a specific type of arterial disease. Therefore, select arterial embolism of left leg as the tentative underlying cause in Step M2. Reapply Step SP6, because the tentative starting point is not the same as the one selected in Steps SP1 to SP8. But colon cancer is an obvious cause of arterial embolism, and colon cancer is the new starting point. No further modifications apply. Code colon cancer (Malignant neoplasm of colon, unspecified) as the underlying cause of death.

|Example 12:|||| |--|--|--| ||1|(a)| Sepsis | |||(b)|Arterial disease, arterial embolism of left leg| |||(c)|Atherosclerosis| ||2|| Colon cancer |

Atherosclerosis is the tentative starting point according to Step SP3. There is a special instruction on atherosclerosis reported as the cause of arterial disease, and, according to this instruction, arterial disease is the new starting point according to Step M1. Arterial embolism of left leg, reported as the second condition on line 1(b), is a more specific description of the type of arterial disease and is selected as the tentative starting point in Step M2. Do not reapply Step SP6, because arterial embolism of left leg is reported as due to atherosclerosis, and this is a correct causal relationship. No further modifications apply. Code arterial embolism of left leg, Embolism and thrombosis of arteries of lower extremities) as the underlying cause of death.

Step M4 – Instructions on medical procedures, poisoning, main injury and maternal deaths

Finally, apply the following instructions to the underlying cause you have arrived at:

• If the underlying cause you arrived at by applying Steps SP1 to SP8 and Steps M1 to M3 is surgery or another type of medical procedure, apply the instructions in Section 6.2.9, Special instructions on surgery and other medical procedures (Step M4).
• If the underlying cause you arrived at by applying the selection and modification rules in Steps SP1 to SP8 and Steps M1 to M3 is an injury or poisoning (a code in S00–T98), code the external cause of the injury or poisoning as the underlying cause of death.
• If the underlying cause is in Chapter 23, External causes of morbidity and mortality, also select a main injury. See the instructions in Section 6.2.6, Special instructions on main injury in deaths from external causes (Step M4).
• If the starting point you selected by applying Steps SP1 to SP8 and Steps M1 to M3 is poisoning, and more than one toxic substance is reported on the certificate, apply the instructions in Section 6.2.7, Special instructions on poisoning by drugs, medicaments and biological substances (Step M4), to identify the most important drug involved.
• If the decedent is a woman, and pregnancy, childbirth or puerperium is reported on the certificate, determine whether to code the underlying cause to Chapter 18, Pregnancy, childbirth and the puerperium, according to the instructions in Section 6.2.8, Special instructions on maternal mortality (Step M4). When you have found a cause of death that is not further changed in either Step SP6 or Steps M1 to M3, you have arrived at the underlying cause of death. Although the cause of death you identified is not further changed in Step SP6 or Steps M1 to M3, other restrictions may apply, for example that the cause is limited to one of the sexes or to a specific age range, or that the cause of death is improbable, considering the geographical setting. Therefore, always check whether any such restrictions apply to the underlying cause you selected.

Special instructions on accepted and rejected sequences (Steps SP3 and SP4)

This section lists sequences of causes of death that should be accepted or rejected when selecting the underlying cause of death. The purpose is to produce the most useful mortality statistics possible. Thus, whether a sequence is listed as ‘rejected’ or ‘accepted’ may reflect interests of importance for public health rather than what is acceptable from a purely medical point of view. Therefore, always apply these instructions, whether they can be considered medically correct or not. Individual countries should not correct what is assumed to be an error, since changes at the national level will lead to data that are less comparable to data from other countries, and thus less useful for analysis.

A. Accepted sequences

When applying Steps SP3 and SP4, accept the relationships listed below.

(a) Infectious diseases due to other conditions

Accept infectious diseases caused by other conditions, except for the infectious diseases listed in Section 6.2.3B, Rejected sequences, subsection (a), Infectious diseases due to other conditions.

(b) HIV reported as due to other conditions

Accept HIV as due to: - conditions necessitating blood transfusion, such as haemophilia, anaemia and major injuries - invasive procedures, such as surgery - drug abuse. Examples of such conditions are given in the Annex 14.5, Causes of HIV. Note that the list in Annex 14.5 is not complete.

(c) Infectious diseases due to HIV

Accept the following infectious diseases as due to HIV disease, malignant neoplasms and conditions impairing the immune system: 1. Typhoid and paratyphoid fevers, Other Salmonella infections, Shigellosis; Tuberculosis Sequelae of tuberculosis

(d) Malignancies and HIV

Accept the following malignant neoplasms as due to Human immunodeficiency virus [HIV] disease: - Malignant neoplasm of oropharynx - Malignant neoplasm of anus and anal canal - Kaposi sarcoma - Malignant neoplasm of vulva - Malignant neoplasm of vagina - Malignant neoplasm of cervix uteri, if specified as invasive - Malignant neoplasm of penis - Hodgkin lymphoma, if specified as primary in brain - Follicular lymphoma, if specified as primary in brain - Non-follicular lymphoma, if specified as primary in brain - Diffuse large B-cell lymphoma, if specified as immunoblastic - Burkitt lymphoma - Mature T/NK-cell lymphoma, if specified as primary in brain - Other and unspecified types of non-Hodgkin lymphoma, if specified as primary in brain - Other specified types of T/NK-cell lymphoma, if specified as primary in brain

(e) Diabetes due to other conditions

Accept Type 1 diabetes mellitus as due to conditions that cause autoimmune destruction of β-cells.

Accept Type 2 diabetes mellitus as due to conditions that cause insulin resistance.

Accept Other specified and unspecified diabetes mellitus as due to conditions that cause damage to the pancreas.

See Annex 14.6 for a list of conditions that can cause diabetes.

(f) Rheumatic fever due to other conditions

Accept Acute rheumatic fever and Chronic rheumatic heart diseases as due to: 1. Scarlet fever Sepsis due to Streptococcus, group A Streptococcal sore throat Streptococcal tonsillitis

(g) Hypertension due to other conditions

Accept a hypertensive condition as due to: 1. endocrine neoplasms renal neoplasms carcinoid tumours.

(h) Cerebrovascular diseases due to other conditions

Accept Intracerebral haemorrhage as due to Diseases of liver.

Accept cerebrovascular embolism, thrombosis and unspecified stroke as due to endocarditis.

(i) Congenital anomalies due to other conditions

Accept a congenital anomaly as due to a chromosome abnormality or a congenital malformation syndrome.

Accept pulmonary hypoplasia as due to a congenital anomaly.

(j) Accidents due to other conditions

Accept a Fall as due to a Disorder of bone density and structure or as due to a (pathological) fracture caused by a Disorder of bone density and structure.

Accept asphyxia and aspiration (W78–W80) caused by other causes.

(k) Acute or terminal circulatory diseases due to other conditions

Accept the following acute or terminal circulatory diseases as due to malignant neoplasm, diabetes or asthma:

*\<list to be edited after finalization of the ICD-11 MMS> *

B. Rejected sequences

When applying Steps SP3 and SP4, reject the relationships listed below.

(a) Infectious diseases due to other conditions

Do not accept the following infectious and parasitic diseases as due to any other causes, not even HIV/AIDS, malignant neoplasms or conditions impairing the immune system:

*\<list to be edited after finalization of the ICD-11 MMS> *

Do not accept the following infectious diseases as due to other causes, except HIV disease, malignant neoplasms and conditions impairing the immune system:

1. Typhoid and paratyphoid fevers, Other Salmonella infections, Shigellosis Tuberculosis Sequelae of tuberculosis.

(b) Malignant neoplasms due to other conditions

Do not accept a malignant neoplasm as due to any other cause, except the following malignant neoplasms as due to HIV:

• Malignant neoplasm of oropharynx
• Malignant neoplasm of anus and anal canal
• Kaposi sarcoma
• Malignant neoplasm of vulva
• Malignant neoplasm of vagina
• Malignant neoplasm of cervix uteri, if specified as invasive
• Malignant neoplasm of penis
• Hodgkin lymphoma, if specified as primary in brain
• Follicular lymphoma, if specified as primary in brain
• Non-follicular lymphoma, if specified as primary in brain
• Diffuse large B-cell lymphoma, if specified as immunoblastic
• Burkitt lymphoma
• Mature T/NK-cell lymphoma, if specified as primary in brain
• Other and unspecified types of non-Hodgkin lymphoma, if specified as primary in brain
• Other specified types of T/NK-cell lymphoma, if specified as primary in brain.

(c) Haemophilia due to other conditions

Do not accept haemophilia as due to any other cause.

(d) Diabetes due to other conditions

Do not accept Type 1 diabetes mellitus as due to any other cause except conditions causing autoimmune destruction of β-cells.

Do not accept Type 2 diabetes mellitus as due to any other cause except conditions causing insulin resistance.

Do not accept Other and Unspecified diabetes mellitus as due to any other cause except conditions causing damage to the pancreas.

See Annex 14.6 for a list of the conditions that can cause diabetes.

(e) Rheumatic fever due to other conditions

Do not accept rheumatic fever or rheumatic heart disease as due to other causes, except: 1. Scarlet fever 2. Streptococcal sepsis 3. Streptococcal sore throat 4. Acute tonsillitis

(f) Hypertension due to other conditions

Do not accept hypertensive conditions as due to a neoplasm, except: 1. endocrine neoplasms 2. renal neoplasms 3. carcinoid tumours.

(g) Chronic ischaemic heart disease due to other conditions

Do not accept Chronic ischaemic heart disease as due to a neoplasm.

(h) Atherosclerosis due to other conditions

Do not accept an atherosclerotic condition as due to a neoplasm.

(i) Influenza due to other conditionsDo not accept Influenza as due to any other cause.

(j) Congenital anomalies due to other conditions

Do not accept a congenital anomaly as due to any other cause, including immaturity, except: 1. congenital anomaly due to a chromosome abnormality or a congenital malformation syndrome 2. Pulmonary Hypoplasia due to a congenital anomaly.

(k) Conflicting durations

Do not accept a condition with a stated duration as due to a condition with a shorter duration (see Examples 6 and 8 in Section 6.2.1, Step SP3, for exceptions).

(l) Accidents due to other conditions

Do not accept accidents as due to causes coded in other chapters, except: 1. Fall as due to a Disorder of bone density and structure 2. Fall as due to a (pathological) fracture caused by a Disorder of bone density and structure 3. Asphyxia and aspiration as due to other causes.

(m) Suicide due to other conditions

Do not accept suicide as due to any other cause.

Special instructions on obvious cause (Step SP6)

This section lists conditions that should be considered an obvious cause of conditions selected as tentative starting point in Steps SP1 to SP5.

A. Complications of HIV

(a) Infectious diseases and HIV

Consider [HIV] disease stage 2-4 as an obvious cause of infectious diseases, except those listed in Section 6.2.3, Special instructions on accepted and rejected sequences, Section B, Rejected sequences, subsection (a), Infectious diseases due to other conditions.

Also consider HIV disease but not HIV-positive status as an obvious cause of Typhoid and paratyphoid fevers, Other Salmonella infections and Shigellosis, these are listed in the second part of Section 6.2.3 B, subsection (a).

Consider both HIV disease and HIV-positive status as an obvious cause of the following infectious diseases:

• Salmonella sepsis
• Cryptosporidiosis
• Isosporiasis
• Tuberculosis
• Infection due to other mycobacteria
• Progressive multifocal leukencephalopathy
• Herpes simplex] infections specified as chronic ulcers, bronchitis, pneumonia, or oesophagitis
• Cytomegalovirus infections, except for liver, spleen, lymph nodes
• Candidiasis of other sites, specified as of lung or oesophagus
• Coccidioidmycosis
• Histoplasmosis
• Cryptococcosis
• Pneumocystosis
• Sequelae of tuberculosis

(b) Malignant neoplasms and HIV

Consider both HIV disease as the obvious cause of the following malignant neoplasms:

1. Kaposi sarcoma
2. Cervix carcinoma, specified as invasive in Malignant neoplasm of cervix uteri
3. Lymphoma, specified as primary cerebral
4. Diffuse large B-cell lymphoma, specified as immunoblastic
5. Burkitt lymphoma

(c) Immune deficiency and HIV

Consider HIV disease as the obvious cause of immune deficiency.

(d) Pneumonia and HIV

Consider HIV disease stage 2-4 as an obvious cause of pneumonia.

(e) Wasting syndrome and HIV

Consider both HIV disease as an obvious cause of wasting syndrome.

B. Enterocolitis due to Clostridium difficile

Consider enterocolitis due to Clostridium difficile as an obvious consequence of antibiotic therapy.

C. Sepsis and systemic inflammatory response syndrome

Consider conditions that impair the immune system, wasting diseases (such as malignant neoplasms and malnutrition), diseases causing paralysis (such as cerebral haemorrhage and thrombosis), serious respiratory conditions and serious injuries (grade 1–4 according to the injury priority list in the Annex 14.7) as obvious causes of sepsis, and of Systemic inflammatory response syndrome [SIRS].

D. Complications of diabetes

Consider Diabetes mellitus (E10–E14) as the obvious cause of the following conditions:

*\<list to be edited after finalization of the ICD-11 MMS> *

E. Dehydration

Consider any intestinal infectious disease as an obvious cause of dehydration.

F. Dementia

Consider conditions that typically involve irreversible brain damage as obvious causes of dementia, if no other cause of the dementia is stated.

Consider Down syndrome as an obvious cause of Unspecified dementia and Alzheimer disease.

G. Mental retardation (F70–F79)

Consider the following conditions as obvious causes of mental retardation:

*\<list to be edited after finalization of the ICD-11 MMS> *

H. Heart failure and unspecified heart disease

Consider other heart conditions as the obvious cause of Heart failure and unspecified Heart disease.

I. Embolism

Consider venous thrombosis, phlebitis or thrombophlebitis, valvular heart disease, childbirth or any operation as the obvious cause of diseases described as ‘embolic’. However, there must be a clear route from the place where the thrombus formed and the place of the embolism.

J. Oesophageal varices Consider cirrhotic liver diseases as the obvious cause of Oesophageal varices.

K. Pneumonia

Consider Dependence syndrome due to use of alcohol as the obvious cause of Lobar pneumonia, unspecified.

Consider conditions that impair the immune system, wasting diseases (such as malignant neoplasms and malnutrition), diseases causing paralysis (such as cerebral haemorrhage and thrombosis), serious respiratory conditions, communicable diseases, conditions that affect the process of swallowing, other diseases that limit the ability to care for oneself, including dementia and degenerative diseases of the nervous system, poisoning and serious injuries (grade 1–4 according to the injury priority list in the Annex 14.7) as obvious causes of any pneumonia.

L. Pulmonary oedema

Consider the following conditions as obvious causes of Pulmonary oedema:

1. heart disease (including pulmonary heart disease)
2. conditions affecting the lung parenchyma, such as:
3. lung infections
4. aspiration and inhalation
5. respiratory distress syndrome
6. high altitude
7. circulating toxins
8. conditions causing fluid overload, such as:
9. renal failure
10. hypoalbuminaemia
11. congenital anomalies affecting the pulmonary circulation, such as:
12. congenital stenosis of pulmonary veins.

M. Nephritic syndrome

Consider any streptococcal infection (scarlet fever, streptococcal sore throat, etc.) as the obvious cause of Nephritic syndrome and Nephrotic syndrome.

N. Pyelonephritis

Consider any urinary obstruction from conditions such as hyperplasia of prostate or ureteral stenosis as the obvious cause of pyelonephritis.

O. Acute renal failure

Consider a urinary tract infection as the obvious cause of Acute renal failure, provided that there is no indication that the renal failure was present before the urinary tract infection developed.

P. Primary atelectasis of newborn

Consider congenital kidney conditions, premature rupture of membranes and oligohydramnios as obvious causes of Primary atelectasis of newborn.

Q. Premature rupture of membranes and oligohydramnios

Consider congenital kidney conditions as obvious causes of Fetus and newborn affected by premature rupture of membranes or oligohydramnios.

R. Haemorrhage

Consider anticoagulant poisoning or overdose as the obvious cause of haemorrhage. However, do not consider anticoagulant therapy, without mention of poisoning or overdose, as the obvious cause of haemorrhage. Further, consider treatment with steroid, aspirin, and nonsteroidal anti-inflammatory drugs (NSAIDs) as obvious causes of gastric haemorrhage.

Consider gastrointestinal haemorrhage as the obvious cause of secondary or unspecified anaemia.

S. Aspiration and inhalation

Consider conditions listed under Section 6.2.4 K, Pneumonia, as obvious causes of aspiration and inhalation.

T. Operations and other medical procedures

Consider surgery as the obvious cause of conditions that are considered common post-procedural complications, see Annex 14.2, List of conditions to be considered direct consequences of medical procedures.

Consider any surgical condition (such as malignant tumour or injury), reported anywhere on the certificate, as the obvious cause of an operation or other medical procedure performed on the same organ.

U. Common secondary conditions

Consider wasting diseases (such as malignant neoplasms and malnutrition), diseases causing paralysis (such as cerebral haemorrhage or thrombosis), communicable diseases, other disease that limits the ability to care for oneself, including dementia and degenerative diseases of the nervous system, and serious injuries as the obvious cause of the common secondary conditions listed in the Table 1. However, such secondary conditions should not be considered an obvious consequence of respiratory conditions.

Conditions in categories flagged with an ‘M’ (Maybe) should be considered obvious consequences of wasting and paralysing conditions only if they meet the prerequisite for code assignment noted in the final column of the table.

Table: Common secondary conditions

\<list to be edited after finalization of the ICD-11 MMS>

Special instructions on linkages and other provisions (Step M1)

Use the list in this section in Step M1.

The tentative underlying cause is listed in the left-hand column. If the conditions specified in the right-hand column apply, then use the code in bold as the new tentative underlying cause.

There are two types of combination:

‘with mention of’ means that the other condition may appear anywhere on the certificate;

‘when reported as the cause of’ means that the other condition must appear in a correct causal relationship or be otherwise indicated as being due to the tentative underlying cause.

For some conditions, there are further requirements, for example that a specific term has been used either for the tentative underlying cause or for the condition that may change the underlying cause code.

Summary of codes not to be used in underlying-cause mortality coding

*\<lists to be edited after finalization of the ICD-11 MMS> *

In addition to Extension codes

**Codes not to be used for underlying-cause mortality coding	(code to item in parentheses; if no code is indicated, code to other and unspecified causes of death)**
	(code to yyyy.y)

##### Codes not to be used if the underlying cause is known #####

\<list to be edited after finalization of the ICD-11 MMS>

Special instructions on main injury in deaths from external causes (Step M4)

If the underlying cause you arrived at by applying the selection and modification rules in Steps SP1 to SP8 and M1 to M3 is an external cause, code the external cause of the injury as the underlying cause of death. In addition to the underlying cause from Chapter 23, External causes of morbidity and mortality, also code a main injury. This applies to both body injuries and poisoning. For special instructions on how to identify the main injury in poisoning deaths, see Section 6.2.7. If more than one injury is reported on the death certificate, apply the following instructions:

(a) When the injuries reported include superficial and trivial injuries (as listed in the Annex 14.4, List of conditions unlikely to cause death), whether in Part 1 or Part 2, select the main injury as if the superficial or trivial injury had not been reported.

| Example 1:||| | |--|--|--| ||1|(a) | Contusion of arm and fracture of skull | || | (b) | Fall from scaffolding |

Fall from scaffolding is the underlying cause of death. Code underlying cause to W12, Fall on and from scaffolding. As main injury, code fracture of skull (Fracture of skull and facial bones, part unspecified). Disregard contusion of arm (Superficial injury of upper limb, level unspecified), as it is in the Annex 14.4, List of conditions unlikely to cause death.

(b) When serious (non-superficial and non-trivial) injuries are reported in both Part 1 and Part 2, select the main injury from Part 1. This applies even when the injuries mentioned in Part 2 have a higher rank in Annex 14.7, Priority ranking of ICD–11 nature-of-injury codes, than the injuries mentioned in Part 1.

| Example 2: |||| |--|--|--| ||1|(a) | Multiple intrathoracic injuries | |||(b) | Car driver, collision with bus | ||2 || Brain injuries |

Code to car driver injured in collision with bus as underlying cause of death Car occupant injured in collision with heavy transport vehicle or bus, driver injured in traffic accident. As main injury, code ‘Multiple injuries of thorax’. Unspecified brain injury has a higher rank in Annex 14.7 than multiple injuries of thorax, but multiple injuries of thorax are mentioned in Part 1 and take precedence over the injuries mentioned in Part 2.

When serious injuries are reported only in Part 2, select a main injury from Part 2.

(c) When more than one serious injury is reported in the relevant part of the certificate, select the main injury according to Annex 14.7, Priority ranking of ICD–11 nature-of-injury codes). Note that 1 is the highest priority rank and that 6 is the lowest.

| Example 3:|||| |--|--|--| || 1|(a) | Multiple intrathoracic injuries and brain injuries | ||| (b) | Car driver, collision with bus | |

Code to car driver injured in collision with bus as underlying cause of death Car occupant injured in collision with heavy transport vehicle or bus, driver injured in traffic accident. As main injury, code brain injury Intracranial injury, unspecified, which has a higher rank on the priority list than Multiple injuries of thorax.

(d) When more than one of the serious injuries reported in the relevant part of the certificate have the same and highest rank, select the first mentioned of these injuries. However, prefer a specific injury over an injury from the group Injuries involving multiple body regions, with the same priority rank.

| Example 4: |||| |--|--|--| ||1|(a) | Multiple injuries with rupture of aorta | ||| (b) | Car driver, collision with bus | |

Code to car driver injured in collision with bus as underlying cause of death Car occupant injured in collision with heavy transport vehicle or bus, driver injured in traffic accident. As main injury, code rupture of aorta Injury of thoracic aorta. Multiple injuries and rupture of aorta have the same rank on the priority list, but a specific injury takes precedence over injury from the group Injuries involving multiple body regions.

Special instructions on poisoning by drugs, medicaments and biological substances (Step M4)

A. Underlying cause

If the underlying cause you selected by applying Steps SP1 to SP8 and M1 to M3 is poisoning, there is more than one drug reported on the certificate and the drugs do not have the same external cause code, select a code for the underlying cause as follows:

(a) If one of the drugs is specified as the most important substance in bringing about the death, code the external cause code for that drug as the underlying cause of death.

| Example 5: |||| |--|--|--| || 1|(a) | Accidental heroin overdose | || 2 || Diazepam and amitriptyline present |

By placing heroin overdose in Part 1 and reporting the other substances as contributing causes of death in Part 2, the certifier has identified heroin as the most important substance in bringing about the death. Select accidental poisoning by heroin as underlying cause Accidental poisoning by and exposure to narcotics and psychodysleptics [hallucinogens], not elsewhere classified.

| Example 6: |||| |--|--|--| || 1|(a) | Poisoning by amphetamine | || 2 || Toxic levels of heroin and flunitrazepam |

By placing amphetamine poisoning alone in Part 1 and reporting the other substances as contributing causes of death in Part 2, the certifier has identified amphetamine as the most important substance in bringing about the death. Select accidental poisoning by amphetamine as underlying cause Accidental poisoning by and exposure to antiepileptic, sedative-hypnotic, antiparkinsonism and psychotropic drugs, not elsewhere classified.

| Example 7: |||| |--|--|--| || 1|(a) | Poisoning by alcohol | | | 2 | |Toxic levels of heroin and flunitrazepam |

By placing alcohol poisoning alone in Part 1 and reporting the other substances as contributing causes of death in Part 2, the certifier has identified alcohol as the most important substance in bringing about the death. Select accidental poisoning by alcohol as underlying cause.

(b) If none of the drugs is specified as the most important substance in bringing about the death, first try to get further information from the certifier. If no clarification can be obtained, code:

1. combinations of alcohol with a drug to the drug;
2. other multidrug deaths to the appropriate category for ‘Other’.

| Example 8: |||| |--|--|--| ||1|(a) | Toxic levels of heroin and amphetamine | |

Neither heroin nor amphetamine is identified as the most important substance in bringing about the death. Code to Accidental poisoning by and exposure to other and unspecified drugs, medicaments and biological substances.

| Example 9: |||| |--|--|--| ||1|(a) | Accidental poisoning by alcohol, heroin and diazepam ||

Neither of the substances is identified as the most important substance in bringing about the death. Poisoning by combinations of alcohol and drugs are coded to the drugs. Code to Accidental poisoning by and exposure to other and unspecified drugs, medicaments and biological substances.

Proceed by identifying the most dangerous drug and code it as the main injury.

B. Main injury

If the underlying cause is poisoning, use the code for poisoning in the Chapter 22, Injury, poisoning and certain other consequences of external causes chapter as main injury. If only one toxic substance is reported, code that substance as main injury. If several toxic substances are reported, identify the most dangerous substance and code it as main injury. To identify the most dangerous substance, apply the instructions that follow. (a) If one toxic substance is specified as the cause of death, code to that component substance.

| Example 10: |||| |--|--|--| ||1|(a) | Accidental overdose by heroin | | | 2 | |Diazepam and amitriptyline present |

By placing heroin overdose alone in Part 1 and reporting the other substances as contributing causes of death in Part 2, the certifier has identified heroin as the most important substance in bringing about the death. Select accidental poisoning by heroin as underlying cause Accidental poisoning by and exposure to narcotics and psychodysleptics [hallucinogens], not elsewhere classified. As main injury, code poisoning by heroin Poisoning by narcotics and psychodysleptics [hallucinogens], heroin.

| Example 11: |||| |--|--|--| ||1|(a) | Alcohol poisoning | ||2 | |Diazepam and amitriptyline present |

By placing alcohol poisoning alone in Part 1 and reporting the other substances as contributing causes of death in Part 2, the certifier has identified alcohol as the most important substance in bringing about the death. Select accidental poisoning by alcohol as underlying cause, Accidental poisoning by and exposure to alcohol. Code poisoning by alcohol as main injury Toxic effect of alcohol, unspecified.

(b) If no single toxic substance is indicated as the cause of death, code combinations of alcohol with a drug to the drug.

| Example 12: |||| |--|--|--| || 1|(a) | Toxic levels of alcohol and flunitrazepam | | | 2 || Diazepam and amitriptyline present |

By placing toxic levels of alcohol and flunitrazepam in Part 1 and reporting the other substances as contributing causes of death in Part 2, the certifier has identified alcohol and flunitrazepam as the most important substances in bringing about the death. Of these two, select poisoning by flunitrazepam because combinations of alcohol with a drug are coded to the drug. Select accidental poisoning by flunitrazepam as underlying cause, Accidental poisoning by and exposure to antiepileptic, sedative-hypnotic, antiparkinsonism and psychotropic drugs, not elsewhere classified). Code poisoning by flunitrazepam as main injury Poisoning by antiepileptic, sedative-hypnotic and antiparkinsonism drugs, benzodiazepines).

(c) If no appropriate combination category is available, select the main nature of injury code, in the following order of priority:

1. Opioid agonists and partial agonists and other and unspecified narcotics. Deaths that include multiple opioids classifiable should be prioritized as:
2. 1a. Heroin
3. 1b. Methadone
4. 1c. Opium
5. 1d. Other opioids
6. 1e. Other synthetic narcotics
7. 1f. Other and unspecified narcotics
8. Inhaled and intravenous anaesthetic agents

   Includes: Propofol

9. Tricyclic and tetracyclic antidepressants
10. Barbiturates
11. 4-Aminophenolderivatives

    Includes: APAP, acetaminophen, paracetamol

12. Antipsychotics and neuroleptics

    Includes: Phenothiazine antipsychotics and neuroleptics Butyrophenone and thioxanthene neuroleptics Other and unspecified antipsychotics and neuroleptics

13. Antiepileptic drugs, antiparkinsonism drugs and unspecified sedatives
14. Cocaine
15. Psychostimulants with abuse potential

    Includes: Amphetamines and derivatives

16. Monoamine oxidase inhibitor (MAO) antidepressants and other and unspecified antidepressants

    Includes: Selective serotonin reuptake inhibitors (SSRIs), venlafaxine

17. Benzodiazepines
18. Drugs and substances not listed above

If there is more than one drug in the same priority group, code to the first mentioned. Note that for poisonings, the selected underlying cause does not always match the code for main injury. For example, the underlying cause may express a combination of toxic substances, but the main injury code identifies the most dangerous component.

| Example 13: |||| |--|--|--| ||1|(a) | Toxic levels of cocaine, heroin, diazepam and amitriptyline |

Neither of the substances is identified as the most important substance in bringing about the death, and there is no specific code category for the combination of these substances. Code to Accidental poisoning by and exposure to other and unspecified drugs, medicaments and biological substances, as the underlying cause of death. As main injury, code to poisoning of heroin. On the priority list above, cocaine is in group 8, heroin is in group 1a, diazepam is in group 11 and amitriptyline is in group 10. Select heroin, the substance with the highest priority, Poisoning by narcotics and psychodysleptics [hallucinogens], heroin). Add the codes for the other substances to the cluster.

| Example 14: |||| |--|--|--| ||1|(a) | Heroin, cocaine, diazepam and amitriptyline overdose |

Neither of the substances is identified as the most important substance in bringing about the death, and there is no specific code category for the combination of these substances. Code to Accidental poisoning by and exposure to other and unspecified drugs, medicaments and biological substances, as the underlying cause of death. Add the codes for the individual substances to the cluster, if desired. Next, code poisoning by heroin as main injury. On the priority list above, heroin is in group 1a, cocaine is in group 8, diazepam is in group 11 and amitriptyline is in group 10. Select heroin, the substance with the highest priority, Poisoning by narcotics and psychodysleptics [hallucinogens], heroin. Add the codes for the other substances to the cluster.

| Example 15: |||| |--|--|--| ||1|(a) | Accidental poisoning by alcohol, heroin and diazepam |

Poisoning by combinations of alcohol and drug(s) is coded to the drug(s), see instruction in Section 6.2.7 B, subsection (b), above. Neither of the drugs reported in Part 1 is identified as the most important substance in bringing about the death, and there is no specific code category for the combination of these substances. Code to Accidental poisoning by and exposure to other and unspecified drugs, medicaments and biological substances, as the underlying cause of death. Next, code poisoning by heroin as main injury. On the priority list above, heroin is in group 1a and diazepam is in group 11. Select heroin, the substance with the highest priority Poisoning by narcotics and psychodysleptics [hallucinogens], heroin) and add codes for the other substances to the cluster.

Special instructions on maternal mortality (Step M4)

If pregnancy, childbirth, or puerperium is mentioned anywhere on the certificate, in most cases the underlying cause is coded to Chapter 18, Pregnancy, childbirth and the puerperium. This is either because the underlying cause you selected by applying Steps SP1 to SP8 and M1 to M4 is classified to Chapter 18 according to the Alphabetical index, or because there is a special code in Chapter 18 for the condition if it appears during pregnancy, childbirth and the puerperium. Apply the following instructions to determine whether an underlying cause that is indexed to other parts of the ICD should be classified to Chapter 18. Note that these instructions do not apply to conditions that are indexed to Chapter 18 in the Alphabetical index. If pregnancy, childbirth or puerperium is reported anywhere on the certificate but it is not clearly stated that pregnancy, childbirth or puerperium contributed to the death, first contact the certifier and ask for additional information.

• If the certifier states that the death was a complication of pregnancy, childbirth or puerperium, code the underlying cause to Chapter 18, Pregnancy, childbirth and the puerperium.
• If the certifier states that the death was not a complication of pregnancy, childbirth or puerperium, do not code the underlying cause to Chapter 18.
• If you cannot obtain any additional information, but pregnancy, childbirth or puerperium is mentioned in Part 1 or Part 2 of the certificate, code the underlying cause to Chapter 18.

If the underlying cause you selected is classifiable to Maternal infectious and parasitic diseases classifiable elsewhere but complicating pregnancy, childbirth and the puerperium and Other maternal diseases classifiable elsewhere but complicating pregnancy, childbirth and the puerperium), then add to the cluster the corresponding code from Chapter 01-19 as a multiple cause of death. This is important because otherwise relevant information on the death will not be retrievable. Note that some conditions are not coded to Chapter 18, even if they occurred during pregnancy, childbirth or puerperium, see the ‘Excludes’ note at the beginning of Chapter 18.

| Example 1: |||| |--|--|--| ||1|(a) | Amniotic fluid embolism |

The underlying cause, Amniotic fluid embolism, is indexed to Chapter 18.

| Example 2: |||| |--|--|--| ||1|(a) | Pulmonary oedema | || |(b) | Mitral regurgitation, pregnancy |

The underlying cause, mitral regurgitation, is coded to Chapter 18 because pregnancy is mentioned in Part 1. Code the underlying cause to Diseases of the circulatory system complicating pregnancy, childbirth and the puerperium. Also add the code for Mitral regurgitation to the cluster as a contributing cause of death.

| Example 3: |||| |--|--|--| ||1|(a) | Haemorrhage | | || (b) | Cervical cancer | | | 2 || Treatment delayed because of pregnancy |

The underlying cause, cervical cancer, is coded to Chapter 18 because pregnancy is mentioned in Part 2. Code the underlying cause to Other specified diseases and conditions complicating pregnancy, childbirth and the puerperium. Also add the code for cervical cancer as a contributing cause of death.

| Example 4: |||| |--|--|--| ||1|(a) | Hepatic failure | | || (b) | Dengue haemorrhagic fever | | | || 5 days | | |2 | |Additional information: 40 days postpartum|

Code the underlying cause to Other viral diseases complicating pregnancy, childbirth and the puerperium. Also add to the cluster the code for Dengue as a contributing cause of death.

Special instructions on surgery and other medical procedures (Step M4)

A. Reason for the surgery or procedure stated

If the tentative starting point you arrived at by applying Steps SP1 to SP7 and M1 to M4 is surgery or other medical procedure and the certificate states the reason for which the operation or procedure was performed, then select the reason for the operation or other procedure as the new starting point. Next, apply the instructions in Steps SP7 and M1 to M4 as already described.

B. Reason for the surgery or procedure not stated, complication reported

If the reason for the surgery or procedure is not stated and a complication is reported, proceed as described next. First check whether the Alphabetical index gives a default code for the reason of the surgery or procedure. If it does, this is the new starting point. Next, apply the instructions in Steps SP7 and M1 to M4 as already described. If the coding tool does not give a default code for the reason of the surgery or procedure, determine whether the type of surgery or procedure indicates a specific organ or site. If it does, then use the code for the residual category for the organ or site operated on as the new starting point. Next, apply the instructions in Steps SP7 and M1 to M4 as already described.

If the coding tool does not give a default code for the reason of the surgery or procedure, and the type of surgery or procedure does not indicate an organ or site, check whether the certificate mentions a misadventure at the time of the procedure. If it does, use the appropriate code from Complications during labour and delivery or Surgical and other medical procedures associated with injury or harm in therapeutic use and also Mode of injury or harm associated with a surgical or other medical procedure as underlying cause of death.

If the coding tool does not give a default code, the type of surgery or procedure does not indicate an organ or site and there is no mention of a misadventure at the time of the procedure, use the appropriate code from Complications during labour and delivery or Surgical and other medical procedures associated with injury or harm in therapeutic use as underlying cause of death.

C. Reason for the surgery or procedure not stated, no complication reported

If the reason for the surgery or procedure is not stated and no complication is reported, proceed as described next. First check whether the Alphabetical index gives a default code for the reason of the surgery or procedure. If it does, this is the new starting point. Next, apply the instructions in Steps SP7 and M1 to M4 as described earlier.

If the Alphabetical index does not give a default code for the reason of the surgery or procedure, determine whether the type of surgery or procedure indicates a specific organ or site. If it does, then use the code for the residual category for the organ or site operated on as the new starting point. Next, apply the instructions in Steps SP7 and M1 to M4 as described earlier.

If the Alphabetical index does not give a default code and the type of surgery or procedure does not indicate an organ or site, code to Other ill-defined and unspecified causes of mortality.

| Example 1:|||| |--|--|--| || 1|(a) | Pulmonary embolism | ||| (b) | Appendectomy |

The certificate does not specify the reason for the surgery, but the term appendectomy indicates appendix as the organ operated on. Code Disease of appendix, unspecified, as the underlying cause of death.

| Example 2: |||| |--|--|--| ||1|(a) | Accidental puncture of aorta | | ||(b)| Laparotomy |

The certificate does not specify the reason for the surgery and the term laparotomy does not indicate a specific organ. However, there is a mention of a misadventure at the time of the surgery. Code the misadventure, accidental puncture during laparotomy, as the underlying cause of death, Unintentional cut, puncture, perforation or haemorrhage during surgical and medical care, during surgical operation.

| Example 3: |||| |--|--|--| || 1|(a) | Postoperative haemorrhage | ||| (b) | Caesarean section | ||| (c) | Prolonged labour |

The certificate states the reason why the surgery was performed. Code the reason for the surgery, prolonged labour, as the underlying cause of death, Long labour, unspecified.

| Example 4: |||| |--|--|--| || 1|(a) | Laparotomy |

The certificate does not specify why the surgery was performed and the term laparotomy does not indicate a specific organ. There is no mention of a complication. Other ill-defined and unspecified causes of mortality, as the underlying cause of death.

D. Medical devices associated with adverse incidents due to external causes

If a death is caused by an incident involving a medical device, but the incident is due to an external cause and not to any breakdown or malfunctioning of the device itself, code the external cause as the underlying cause of death.

| Example 5:|||| |--|--|--| || 1| (a) | Inhalation pneumonia | | | |(b) | Haemorrhage of trachea | | | |(c) | Fell from bed while attached to respirator | | |2|| Respirator treatment following liver transplant

There is no mention of breakdown or malfunctioning of the respirator or the tracheal tube. Code Fall involving bed, the accident that caused the haemorrhage, as the underlying cause of death.

| Example 6: |||| |--|--|--| || 1|(a) | Pulmonary oedema | | || (b)|Intra-aortic balloon pump stopped | | ||(c)|Power cut due to hurricane | | ||(d)|Recent myocardial infarction with mitral insufficiency |

The balloon pump stopped working, not because of any malfunctioning or breakdown, but because of a power cut. Code the reason of the power cut, cataclysmic storm, as the underlying cause of death, Victim of cataclysmic storm.

If the external cause of the incident is not specifically classified, code to Exposure to unspecified factor causing other and unspecified injury.

| Example 7: |||| |--|--|--| ||1|(a) | Cardiac and respiratory failure | ||| (b) | Stopped administration of inotropic drugs | | || (c) | Accidental removal of subclavian line | | | 2 | |Surgery for acute rupture of gallbladder |

There is no mention of malfunctioning or breakdown of equipment. Since the accident that caused the removal of the subclavian line is not described, code to, Exposure to unspecified factor causing other and unspecified injury.

Coding instructions for mortality: multiple causes

Multiple-cause coding permits in-depth analysis of causes of death, for example of serious but avoidable complications of certain underlying causes, and the impact of coexisting conditions on the outcome of a disease process. Therefore, in mortality coding, both underlying cause and multiple causes should be recorded. Also, complete multiple-cause coding is essential for a correct application of the ICD instructions for selection and modification of the underlying cause of death (see Section 6.2). All possible detail should be retained in the multiple-cause coding, since records containing all multiple-cause conditions permit more thorough analysis than records with only a selection of the conditions reported on the certificate. In particular: - the position of the individual codes in the data record should reflect where on the certificate the corresponding diagnostic expressions were entered by the certifier, because some analyses may focus on the terminal cause of death, or on conditions reported in Part 2; - codes for common conditions, or for conditions regarded as symptomatic or less informative, should not be deleted or left out, since they may be of special interest in analysis of avoidable complications and may serve as markers of the seriousness of other conditions reported on the certificate; - multiple-cause data should be stored in two formats: one format that shows as clearly as possible which term the certifier used on the certificate and where on the certificate each term was reported; and one format that takes the stated or implied relationships between the reported conditions into consideration, and where the codes have been harmonized according to the instructions in the ICD volumes.

Uncertain diagnosis

Ignore expressions indicating doubt as to the certainty of the diagnosis, for example ‘apparently’, ‘presumably’, ‘probably’ or ‘possibly’. A tentative diagnosis, although uncertain, is of better use to mortality statistics than no diagnosis at all.

Either … or

The certifier might report alternative diagnoses, ‘either diagnosis A or diagnosis B’. In such cases, proceed as follows.

A. One condition, either one site or another

(a) If the sites are in the same anatomical system, code to the residual category for the group or anatomical system in which the reported sites are classified.

| Example 1: |||| |--|--|--| | |1|(a)| Cancer of kidney or bladder |

Code as Malignant neoplasm, urinary organ, unspecified.

(b) If the reported sites are in different anatomical systems, or if there is no residual category for the group or anatomical system, code to the residual category for the disease or condition specified.

| Example 2: |||| |--|--|--| ||1|(a)| Cancer of adrenal gland or kidney |

Code as Malignant neoplasm, primary site unspecified, since adrenal gland and kidney are in different anatomical systems.

B. One site or system, either one condition or another condition

(a) If the reported conditions are classifiable to different four-character subcategories of the same three-character category, code to the four-character subcategory for ‘unspecified’.

| Example 3: |||| |--|--|--| ||1|(a)| Arteriosclerotic heart disease or coronary aneurysm |

Code as Chronic ischaemic heart disease, unspecified.

(b) If the reported conditions are classifiable to different three-character categories but ICD–11 provides a residual category for the disease in general, code to the residual category.

| Example 4: |||| |--|--|--| ||1|(a)| Myocardial infarction or coronary aneurysm |

Code as the residual category for ischaemic heart disease.

(c) If the reported conditions are classifiable to different three-character categories and there is no residual category for the disease in general, code to the residual category relating to the disease of the anatomical site/system.

| Example 5: |||| |--|--|--| ||1|(a)| Tuberculosis or cancer of lung |

Code as Other disorders of lung. Both conditions involve the lung.

| Example 6: |||| |--|--|--| ||1|(a)| Stroke or heart attack |

Code as Other and unspecified disorders of circulatory system. Both conditions are in the circulatory system.

C. Either one condition or another, different anatomical systems When different diseases of different anatomical systems are reported as ‘either ... or’, code to Other specified general symptoms and signs.

| Example 7: |||| |--|--|--| ||1|(a)| Gallbladder colic or coronary thrombosis |

Code as Other specified general symptoms and signs.

D. Either disease or injury When death is reported as due to either a disease or an injury, code to Other ill-defined and unspecified causes of mortality.

| Example 8: |||| |--|--|--| ||1|(a)| Coronary occlusion or war injuries |

Code as Other ill-defined and unspecified causes of mortality.

Effect of connecting terms

When the certifier uses a connecting term, the codes assigned must be arranged to reflect the certifier intention. There are two types of connecting terms: those implying a causal relationship and those not implying a causal relationship between reported causes of death.

A. Connecting terms implying a causal relationship

A causal relationship can be expressed in two ways: ‘due to’ written or implied by a similar term; or ‘resulting in’ written or implied by a similar term.

(a) ‘Due to’ written or implied by a similar term

When one cause is certified with a connecting term implying it is due to another cause, enter the code for the first cause on the line where reported and the code for the other cause on the next lower line. Code any causes reported on the remaining lines in Part 1 on the next lower lines.

| Example 1: |||| |--|--|--| ||1|(a)| Heart failure due to ischaemic heart disease | ||| (b)| Diabetes |

Heart failure is the first cause on line (a), and code it to line (a). It is reported as due to ischaemic heart disease, so code ischaemic heart disease to line (b). Move diabetes, which is written on line (b), to line (c).

| Example 2: |||| |--|--|--| ||1|(a)| Heart failure because of hepatocellular carcinoma | | ||(b)| Ischaemic heart disease | | ||(c)| Diabetes |

Heart failure is the first cause on line (a), and code it to line (a). It is reported as due to hepatocellular carcinoma, so code hepatocellular carcinoma to line (b). Move ischaemic heart disease, which is reported on line (b), to line (c). Also move diabetes, which is reported on line (c), to line (d). This applies to other connecting terms or signs that indicate a ‘due to’ relationship, such as ‘caused by’, ‘because of’, or similar.

(b) ‘Resulting in’ written or implied by a similar term

When one cause is certified with a connecting term implying it resulted in another cause, enter the code for the cause following the connecting term on the line where reported, and the code for the cause preceding the connecting term on the next lower line. Code any causes reported on the remaining lines in Part 1 on the next lower lines.

| Example 3: |||| |--|--|--| ||1|(a)| Ischaemic heart disease resulting in heart failure | ||| (b)| Diabetes |

Code heart failure, which follows the connecting term ‘resulting in’, on line (a). Code ischaemic heart disease, which is reported before the connecting term, on line (b). Move diabetes, reported on line (b), one line down and code it on line (c).

| Example 4: |||| |--|--|--| ||1|(a)| Hepatocellular carcinoma causing heart failure | |||(b)| Ischaemic heart disease | |||(c)| Diabetes |

Code heart failure, reported after the connecting term ‘causing’, on line (a). Code hepatocellular carcinoma, which is reported before the connecting term, on line (b). Move ischaemic heart disease, reported on line (b), to line (c), and move diabetes, which is reported on line (c), to line (d). This applies to other connecting terms or signs that indicate a ‘resulting in’ relationship, such as ‘causing’, ‘leading to’, ‘developing into’, and similar.

B. Connecting terms not implying a causal relationship

(a) ‘And’ written or implied by a similar term first or last on a line

The connecting term ‘and’ does not imply a causal relationship, but it indicates that the terms before and after it both belong to an enumeration. Therefore, when a line ends with ‘and’, code the cause or causes on the next lower line last on the upper line, so that the coding reflects the enumeration implied by the connecting term. Similarly, when a line starts with ‘and’, consider this as a continuation of an enumeration starting on the line above, and code the cause or causes on that line last on the line above. Code any causes reported on the remaining lines in Part 1 where reported. This applies to other connecting terms or signs that indicate an enumeration but do not imply a causal relationship, such as ‘also’, ‘plus’, ‘besides’, ‘in addition’, ‘+’ or comma.

| Example 5: |||| |--|--|--| ||1|(a)| Heart failure and | ||| (b)| Ischaemic heart disease | | | | (c)| Diabetes |

Line 1(a) ends with ‘and’, so consider ‘ischaemic heart disease’, reported on line (b) as a part of the enumeration ‘heart failure and ischaemic heart disease’. Code accordingly, and place the codes for both heart failure and ischaemic heart disease on line 1(a). Code diabetes where it is reported, on line (c).

| Example 6: |||| |--|--|--| ||1|(a)|Heart failure | |||(b)| Ischaemic heart disease | |||(c)| and diabetes |

Line 1(c) starts with ‘and’. Consider diabetes, reported on line (c), as a part of the enumeration ‘ischaemic heart disease and diabetes’. Code accordingly, and place the codes for both ischaemic heart disease and diabetes on line 1(b).

(b) ‘And’ written or implied by a similar term but not first or last on a line

If a connecting term that does not imply a causal relationship is written on a line but not first or last, then treat is as a comma. Do not reformat the text and do not move any part of the causes to another line.

C. Diagnostic terms that do not stop at the end of the line

If a diagnostic term starts on one line in Part 1 and continues on the next line, code as if the entire diagnostic term had been written on the line where the diagnostic term starts. Code any causes reported on the remaining lines in Part 1 where reported.

| Example 7: |||| |--|--|--| ||1|(a)|Ischaemic | |||(b)| Heart disease | |||(c)| Diabetes type 2 |

‘Ischaemic heart disease’ is a diagnostic term reported on two lines. Code as if the complete term had been written on line (a). Code diabetes where it is reported, on line (c).

| Example 8: |||| |--|--|--| ||1|(a)| Pneumonia | |||(b)| Chronic kidney | |||(c)| disease, diabetes type 2 |

‘Chronic kidney disease’ is a diagnostic term reported on two lines. Reformat the certificate and code the complete term ‘chronic kidney disease’ on line (b). Also code diabetes on line (b), since it continues the line where ‘chronic kidney’ has been written.

Malignant neoplasms

To assign the correct multiple-cause code for a neoplasm, you must first determine behaviour (malignant, in situ, benign, uncertain or unknown) for each of the neoplasms reported on the death certificate. For malignant neoplasms, you must also determine whether to code them as primary or secondary. To that end, apply the instructions outlined in Sections 6.4 A and 6.4 B that follow.

A. Behaviour: malignant, in situ, benign or unknown/uncertain behaviour?

The four major types of behaviour are:

• malignant: the neoplasm invades surrounding tissue or disseminates from its point of origin and begins to grow at another site;
• in situ: the neoplasm is malignant but still fully confined to the tissue in which it originated;
• benign: the neoplasm grows in the place of origin without the potential for spread;
• uncertain or unknown behaviour: it is undetermined or unknown whether the neoplasm is benign or malignant.

Determine which code group to use as follows:

(a) The term itself indicates behaviour

Look in the ICD coding tool for the term used on the certificate to describe the neoplasm. If both morphology and location are stated, then look up the morphology term first. For specific morphologies, the coding tool gives either the ICD code to use, or directs you to the proper part of the list at ‘Neoplasm’ in the coding tool. If the morphology is not stated, go to the ‘Neoplasm’ list in the coding tool and code by site and behaviour.

(b) Other information on the certificate indicates behaviour

If the term used on the certificate does not indicate a specific behaviour, then look for other information indicating behaviour.

Code a neoplasm of unspecified behaviour, or described as ‘in situ’, as malignant if it is reported as the cause of, or together with, metastases or infiltration. See also Section 6.4 B, ‘Malignant neoplasms: primary or secondary?’, subsection (c), Other indication of primary malignant neoplasm.

| Example 1|||| |--|--|--| ||1|(a)|Colon tumour with liver metastases |

The colon tumour is reported with liver metastases and is considered malignant. Code the colon tumour as primary.

This also applies to other types of growths that are not indexed to Chapter 02, for example, certain polyps. If they are reported as the cause of metastases or secondary tumours, they should be considered malignant and coded as malignant neoplasms.

Also consider a neoplasm of unspecified behaviour as malignant if it is reported as due to a malignant neoplasm. To decide whether it is primary or secondary, see the instructions in section 6.4 B, Malignant neoplasms: primary or secondary?, subsection (c), Other indication of primary malignant neoplasm.

If a tumour is indexed to the Chapter 02 section for benign neoplasm but is reported as the cause of metastases or infiltration, check in the Alphabetical index and in Volume 1 whether there is a code for a malignant variety. If so, code it as malignant. If there is no code for a malignant variety, first try to obtain clarification from the certifier. If no further information is available, then accept the statement on the certificate and use the code for benign tumour.

If there is no indication of malignancy, code as uncertain or unknown behaviour.

B. Malignant neoplasms: primary or secondary?

If the neoplasm is coded to malignant neoplasms, next decide whether it is primary or secondary.

The primary site is the anatomical location where the malignant neoplasm originated. A malignant neoplasm may spread to other parts of the body, and these sites are referred to as secondary or metastases. It is most important to determine the primary site. When the death certificate is ambiguous as to the primary site, every effort should be made to obtain clarification from the certifier. The instructions that follow should be applied only when clarification cannot be obtained. The ICD provides the following blocks for primary malignant neoplasms:

• Malignant neoplasms, stated or presumed to be primary, of specified anatomical site. This group does not include lymphoid, haematopoietic and related tissues and does not include Neoplasms of brain and central nervous system

• Malignant neoplasms of ill-defined sites

• Malignant neoplasm, without specification of site

• Neoplasms of haematopoietic and lymphoid tissues – includes malignant neoplasms

• Neoplasms of brain and central nervous system – includes malignant neoplasms

For secondary malignant neoplasms, the ICD provides the block:

• Secondary and unspecified malignant neoplasms, stated or presumed to be metastatic spread from another site

For malignant neoplasms of unspecified site not stated or presumed to be primary or secondary, the ICD provides a code for, Malignant neoplasm, primary site unspecified.

(a) Common sites of metastases

When choosing between codes for primary and secondary malignant neoplasms, refer to the following list of common sites of metastases:

|| |--|--| |bone|mediastinum| |brain|meninges| |diaphragm|peritoneum| |ill-defined site|pleura| |liver|retroperitoneum| |lung|spinal cord| |lymph nodes|

See below for further instructions on how to use this list.

(b) Malignant neoplasm reported as primary

If the certifier describes a malignant neoplasm as ‘primary’, ‘primary in’, ‘originating in’, or with similar terms, then use a code for primary malignant neoplasm.

If the morphology has been stated, always look up the morphology in Volume 3 first, because for some morphologies there are specific ICD codes that are different from the code given in the ‘Neoplasm’ table by site and behaviour.

(c) Other indication of primary malignant neoplasm

Also code a malignant neoplasm as primary, although not described as primary by the certifier, if:

• all other malignant neoplasms on the certificate are described as secondary or as metastases;
• it is in the code range ‘Neoplasms of haematopoietic and lymphoid tissues’:
• A primary neoplasm of haematopoietic and lymphoid tissues may occur simultaneously together with another primary neoplasm in the same range. Code all malignant neoplasms classifiable to Neoplasms of haematopoietic and lymphoid tissues as primary, unless the certifier specifies them as secondary;
• the site is not on the list of common sites of metastases.

If the site is on the list of common sites of metastases, code the malignant neoplasm as primary if:

• the morphology indicates that it is primary of the reported site;
• it is described as caused by a known risk factor for malignant neoplasms of the stated site;
• it is the only malignant neoplasm mentioned on the death certificate, and it is not described as ‘metastatic’:

exception: code malignant neoplasm of lymph nodes as secondary, even if it is the only reported neoplasm on the certificate, unless it is stated that the lymph node neoplasm is primary; note: if the only malignant neoplasm reported on the certificate is malignant neoplasm of liver, and it is not specified as either primary or secondary, then use the code, Malignant neoplasm of liver, unspecified;

• it is malignant neoplasm of lung, and all other malignant neoplasms mentioned on the certificate are on the list of common sites of metastases: code lung as secondary only if another malignant neoplasm is reported in the same part of the certificate (Part 1 or Part 2 of frame A) and this other malignant neoplasm is coded as a primary malignant neoplasm.

• Always code lung as primary if the malignant neoplasm is described as bronchogenic or of bronchus

Code a neoplasm that is not indexed as malignant, for example meningioma, as primary malignant if it is reported as causing secondary or metastatic spread and a code for a malignant variety of the neoplasm is available. See also above, Section 6.4 A, Behaviour: malignant, in situ, benign or unknown/uncertain, subsection (b), Other information on the certificate indicates behaviour.

Exceptions are listed next.

• If durations are stated, the secondary neoplasms must not have a longer duration than the presumed primary malignant neoplasm.
• If morphologies are stated, the secondary and presumed primary malignant neoplasms must have the same morphology.
• If a neoplasm that would not be coded as malignant is reported as the cause of another neoplasm that would not be coded as malignant, then code both neoplasms according to the Alphabetical index. Do not assume malignancy or metastatic spread.

| Example 2: |||| |:--|:--|:--| || 1|(a)|Brain metastasis | || |(b)| Lung tumour|

The lung tumour has caused metastatic spread and is considered malignant. It is also considered primary, since the other site mentioned (brain) is a metastasis. Code the lung tumour as primary of lung.

| Example 3: |||| |:--|:--|:--| ||1|(a)|Cancer of pancreas | ||| (b) | Cancer of stomach|

Pancreas and stomach are not on the list of common sites of metastases. Code both cancers as primary.

| Example 4: |||| |:--|:--|:--| ||1|(a)| Cancer of liver and lung | ||| (b) | Chronic hepatitis |

Chronic hepatitis increases the risk of primary liver cancer. Therefore, consider the liver cancer primary and code to malignant neoplasm of liver, unspecified. Do not use the code for Secondary malignant of liver. Code the lung cancer as Secondary, because the only other malignant neoplasm on the certificate is primary.

| Example 5: |||| |:--|:--|:--| ||1|(a)| Kidney cancer and lung cancer | |

Code the kidney cancer as primary, since it is not on the list of common sites of metastases. Code lung cancer as secondary, since it is reported in the same part of the certificate as the kidney cancer and the kidney cancer is considered primary.

| Example 6: |||| |:--|:--|:--| ||1|(a)| Lung cancer | ||2|| Kidney cancer |

Code the lung cancer as Primary. There is no other primary malignant neoplasm in the same part of the certificate as where lung cancer is reported, and the code for lung cancer is not influenced by neoplasms mentioned in another part of the certificate. Code the kidney cancer as Primary, since it is not on the list of common sites of metastases.

| Example 7: |||| |--|--|--| ||1|(a)| Brain tumour | ||2|| Lung tumour, probably secondary|

Consider both tumours as malignant, since the certifier described one of the two as secondary, which is evidence of malignant behaviour. See Section 6.4 A, Behaviour: malignant, in situ, benign or unknown/uncertain, subsection (b), Other information on the certificate indicates behaviour. Code the brain tumour as primary, since the other malignant neoplasm on the certificate is described as secondary. The qualification ‘probably’ is ignored; see Section 6.3.2, Uncertain diagnosis.

| Example 8: |||| |--|--|--| ||1|(a)|Metastatic involvement of chest wall| |||(b)| Carcinoma in situ of breast|

Code the carcinoma in situ of breast as primary Malignant neoplasm of breast. Since the breast tumour has spread to the chest wall it is no longer in situ.

| Example 9: |||| |--|--|--| ||1|(a)|Secondary malignant neoplasm of lung and brain| |||(b)|Polyp of stomach|

Code the polyp as primary malignant neoplasm of stomach. Since the polyp is reported as the cause of secondary spread, it is considered malignant.

| Example 10: |||| |--|--|--| ||1|(a)| Brain cancer|

Brain is on the list of common sites of metastases, but in this case it is the only malignant neoplasm mentioned on the certificate. Use the code for primary Malignant neoplasm of brain

| Example 11: |||| |--|--|--| ||1|(a)| Cancer of cervical lymph nodes|

Code the cancer of cervical lymph nodes as secondary. It is considered secondary to an unspecified primary malignant neoplasm.

| Example 12: |||| |--|--|--| ||1|(a)| Bladder cancer| |||(b)|Primary in prostate|

The prostate cancer is described as primary. Code it to the group of primary malignant neoplasms. Code bladder cancer as secondary, since the certificate states that the cancer was primary in another site. See also Section 6.4 B, Malignant neoplasms: primary or secondary, subsection (e), Other indication of secondary malignant neoplasm.

| Example 13: |||| |--|--|--| ||1|(a)| Bladder tumour | |||(b)|Lung tumour|

None of the tumours is specified as malignant or benign. Therefore, do not assume malignancy or metastatic spread. Use codes from the group of Neoplasms of uncertain or unknown behaviour, bladder and lung.

(d) Malignant neoplasm reported as secondary

If the certifier describes a neoplasm as secondary, then code to the appropriate subcategory in Malignant neoplasms of ill-defined, secondary and unspecified sites.

(e) Other indication of secondary malignant neoplasm

If a malignant neoplasm is not described as primary or secondary but the morphology is stated, first look up the morphology in the Alphabetical index. If the morphology is incompatible with the stated site of the neoplasm (i.e. the neoplasm cannot be primary of the stated site according to textbooks and other reference literature), then assign a code for a malignant neoplasm of unspecified site for the morphology indicated.

Code a malignant neoplasm as secondary if the neoplasm is:

• specified as secondary by the certifier;
• unspecified whether primary or secondary, and the site is on the list of common sites of metastases:

  exception: if there is only one malignant neoplasm mentioned and the site is on the list of common sites of metastases, then code the neoplasm as primary although it is on the list of common sites of metastases. This does not apply to lymph nodes, which are always coded as secondary. See also section 6.4 B, Malignant neoplasms: primary or secondary?, subsection (b), Other indication of primary malignant neoplasm;

exception: code lung as primary, if all other sites in the same part of the certificate (Part 1 or Part 2) are on the list of common sites of metastases;

• unspecified whether primary or secondary, and the certifier states that the cancer is primary in another site. This applies whether the site is on the list of common sites of metastases or not:

• regardless of site, do not code a neoplasm as secondary if it is of a different morphology from another neoplasm stated to be primary. See also Section 6.3.5C, More than one primary malignant neoplasm;

• unspecified whether malignant, in situ or benign, and it is reported as due to a malignant neoplasm:

  exception: if durations are stated, do not code the unspecified neoplasm as secondary if it has a duration that is longer than the durations of the malignant neoplasm reported as the cause of the unspecified neoplasm;

• the morphology indicates that the neoplasm cannot be primary of the stated site.

Do not use order of entry to determine whether a neoplasm specified as malignant is primary or secondary. Code a malignant neoplasm reported as due to another malignant neoplasm as secondary only if it is described as secondary, metastatic spread or similar, or if it is on the list of common sites of metastases.

Do not confuse ‘primary’ with ‘primary in’. Whereas ‘primary in’ identifies one of several malignant tumours as the primary tumour, ‘primary’ simply means that the malignant neoplasm was not secondary. It does not necessarily mean that all other malignant neoplasms mentioned on the certificate were secondary.

| Example 14: |||| |--|--|--| ||1|(a)| Carcinoma of adrenal glands | ||2||Primary in kidney|

The malignant neoplasm of adrenal glands is considered secondary, since the certificate states that the cancer was primary in kidney. Code the adrenal carcinoma as Secondary and the primary in kidney as Malignant Primary neoplasm of kidney.

| Example 15: |||| |--|--|--| ||1|(a)| Prostate cancers | |||(b)|Primary site unknown|

The primary site is described as unknown. Code to Malignant neoplasm of unknown primary site. Code prostate cancer as Secondary, since the primary malignant neoplasm clearly was in another site.

| Example 16: |||| |--|--|--| ||1|(a)| Brain tumour | |||(b)|Lung cancer|

The brain tumour is considered malignant, since it is reported as due to lung cancer. Also, it is considered secondary, since it is on the list of common sites of metastases and reported together with lung cancer. Code the brain tumour as Secondary malignant. Code the lung cancer as primary, since the only other reported neoplasm is on the list of common sites of metastases.

| Example 17: |||| |--|--|--| ||1|(a)| Cancer growth in liver and lymph nodes | ||2||Malignant neoplasm of stomach|

The cancer growth in liver and lymph nodes is considered secondary, since they are both on the list of common sites of metastases. Code as Secondary malignant neoplasm of liver and lymph node, and as Malignant primary neoplasm of stomach.

| Example 18: |||| |--|--|--| ||1|(a)| Cancer of lung, pleura and chest wall |

Code the cancer of lung as primary, since the other sites mentioned on the certificate, pleura and chest wall, are on the list of common sites of metastases. Code cancer of pleura and chest wall as secondary.

|Example 19: |||| |--|--|--| ||1|(a)| Mesothelioma of pleura and lymph nodes|

Mesothelioma of pleura is in the code range for primary malignant neoplasms. The malignant neoplasm of lymph nodes is considered secondary, since lymph nodes is on the list of common sites of metastases.

|Example 20: |||| |--|--|--| ||1|(a)|Lung cancer| ||2||Stomach cancer |

Code both lung cancer and stomach cancer as primary. Although lung is on the list of common sites of metastases, it is the only malignant neoplasm mentioned in Part 1 of the certificate, and the coding of lung cancer is not influenced by neoplasms mentioned in another part of the certificate.

|Example 21: |||| |--|--|--| ||1|(a)|Cancer of bladder | |||(b)|Cancer of kidney |

Code both cancer of bladder and cancer of kidney as primary, since neither is on the list of common sites of metastases, and neither is described as primary.

|Example 22: |||| |--|--|--| ||1|(a)|Osteosarcoma of sacrum| |||(b)|Clear cell cancer of kidney |

Code both malignant neoplasms as primary. Bone is on the list of common sites of metastases, but osteosarcoma is indexed as a primary cancer of bone. Also, it is of different morphology than clear cell cancer of kidney.

|Example 23: |||| |--|--|--| ||1|(a)|Osteosarcoma of lung|

The morphology indicates a primary neoplasm of bone, and the reported site (lung) is incompatible with the morphology. Code to osteosarcoma of unspecified site, also add a code for Secondary malignant neoplasm of lung.

If all sites are on the list of common sites of metastases, then code all sites as secondary. It is recommended that you also add a code for unknown primary. Code ‘primary malignant neoplasm of unspecified site’, if no morphology is stated. If the morphology is stated, then code to the ‘unspecified site’ code for the morphology involved.

C. More than one primary malignant neoplasm

More than one primary malignant neoplasm may be reported on the same certificate. Code each primary malignant neoplasm. Indications of several primary malignant neoplasms are:

• different morphologies;
• a site-specific morphology reported with a malignant neoplasm of another site that is not on the list of common sites of metastases;

• the sites are not on the list of common sites of metastases:

  • if one morphology term is less specific and covers a more specific term that is also used on the certificate, then consider the two as referring to the same neoplasm;

  • do not consider ‘cancer’ or ‘carcinoma’ as morphologic terms, but as synonyms to ‘malignant neoplasm’.

|Example 24: |||| |--|--|--| ||1|(a)|Transitional cell carcinoma of bladder| ||2| |Osteosarcoma, primary in knee |

Bladder on 1(a) is not on the list of common sites of metastases. The malignant neoplasm reported in Part 2 is specified as primary. Further, the two neoplasms are of different morphology and both are considered primary. Code as Malignant neoplasm of bladder and primary osteosarcoma of knee.

||Example 25: |||| |--|--|--| ||1|(a)|Hepatoma| |||(b) |Cancer of breast|

The morphology ‘hepatoma’ indicates a primary malignant neoplasm of liver. The breast cancer is also considered primary, since breast is not on the list of common sites of metastases. Code as Hepatoma and primary malignant neoplasm of breast.

|Example 26: |||| |--|--|--| ||1|(a)|Oat cell carcinoma | |||(b) |Cancer of breast|

The morphology ‘oat cell carcinoma’ indicates a primary malignant neoplasm of lung. The breast cancer is also considered primary, since breast is not on the list of common sites of metastases. Code as primary, although lung is on the list of common sites of metastases, and primary Malignant neoplasm of breast.

D. Site not clearly indicated

If a malignant neoplasm is described as in the ‘area’ or ‘region’ of a site, or if the site is prefixed by ‘peri’, ‘para’, ‘pre’, ‘supra’, ‘infra’ or similar expressions, then first check whether this compound term is included in the Alphabetical index.

If the compound term is not in the Alphabetical index, then code morphologies classifiable to one of the categories: - Malignant melanoma of skin - Other malignant neoplasms of skin - Mesothelioma, - Kaposi sarcoma - Peripheral nerves and autonomic nervous system - Connective and soft tissue - Meninges - Brain - Other parts of central nervous system

to the appropriate subdivision of that category. If the compound term is not in the Alphabetical index and the morphology is not classifiable to the categories above, or the morphology is not stated, then code to the appropriate subdivision of Neoplasm of other and ill-defined sites.

|Example 27: |||| |--|--|--| ||1|(a)| Fibrosarcoma in the region of the pancreas |

Code as Malignant neoplasm of connective and soft tissue of abdomen.

|Example 28: |||| |--|--|--| ||1|(a)| Carcinoma in the lung area |

Code as Malignant neoplasm of other and ill-defined sites, within the thorax. When the site of a primary malignant neoplasm is not specified, do not make any assumption of the primary site from the location of other reported conditions such as perforation, obstruction or haemorrhage. These conditions may arise in sites unrelated to the neoplasm. For example, intestinal obstruction may be caused by the spread of a malignant neoplasm of ovary.

|Example 29: |||| |--|--|--| ||1|(a)| Obstruction of intestinea | | | | (b) | Carcinoma|

Code the carcinoma as Malignant neoplasm, without specification of site.

E. Primary site unknown

If the certificate states that the primary site is unknown and does not mention a possible primary site, code to the category for unspecified site for the morphological type involved. For example, code adenocarcinoma to ‘primary site unknown’.

|Example 30: |||| |--|--|--| ||1|(a)| Secondary carcinoma of live | | | | (b) | Primary site unknown|

The certificate states that the primary site is unknown. For line 1(b), use the code for primary carcinoma without specification of site.

|Example 31: |||| |--|--|--| ||1|(a)|Generalized metastases| | | | (b) |Melanoma | | | | (c) | Primary site unknown |

The certificate states that the primary site is unknown. Code as primary malignant melanoma of unspecified site.

However, if the certificate mentions a probable or possible primary site, disregard the expression indicating doubt and code to that site. See also Section 6.3.2, (Uncertain diagnosis).

| Example 32: |||| |--|--|--| ||1|(a) | Secondary carcinoma of liver | ||| (b) | Primary site unknown, possibly stomach |

The certificate states that the primary site is unknown, but it also mentions stomach as a possible primary site. Ignore ‘possibly’ and code line 1(b) as Primary malignant neoplasm of stomach.

If the certificate mentions several possible primary sites, select a code according to the instructions in Section 6.3.3 A, (One condition, either one site or another) above.

| Example 33: |||| |--|--|--| ||1|(a) | Secondary carcinoma of liver | || |(b) | Primary site unknown, probably stomach or colon. |

The certificate states that the primary site is unknown, but it also mentions stomach or colon as a possible primary site. Code line 1(b) as primary Malignant neoplasm of ill-defined sites within the digestive system.

F. Overlapping sites

The introduction to Chapter 02 describes the contents and the intended use of subcategory .8 for malignant neoplasms of overlapping sites. In mortality coding, however, the codes for malignant neoplasms of overlapping sites should be used only if the lesion has been expressly described as overlapping, or if the anatomical term used on the death certificate indicates an overlapping site. Do not use the codes for overlapping lesions if a malignant neoplasm has spread from one part of an organ or organ system to another part of the same organ or organ system.

| Example 34: |||| |--|--|--| ||1|(a) | Overlapping malignant neoplasm of tongue and floor of mouth |

Code as Overlapping lesion of lip, oral cavity and pharynx. The neoplasm is described as overlapping.

| Example 35: |||| |--|--|--| || 1|(a) | Malignant neoplasm of rectosigmoid colon |

Code as Malignant neoplasm of rectosigmoid junction. The term ‘rectosigmoid’ indicates an overlapping site. It is not sufficient that the certificate enumerates contiguous sites. In that case, code the sites one by one according to the instructions given above.

| Example 36: ||||| |--|--|--| ||| 1|(a) | Malignant neoplasm of colon and gallbladder |

There is no statement that ‘colon and gallbladder’ refers to an overlapping neoplasm. None of the sites is on the list of common sites of metastases, and consequently they are considered as two independent primary sites. Code as primary Malignant neoplasm of colon, and primary Malignant neoplasm of gallbladder.

G. ‘Metastatic’ cancer

Note: The expression ‘metastatic’ is a problem mainly in the English language. Other countries should translate only as much as needed of Section 6.3.5 G. Neoplasms qualified as metastatic are always malignant, either primary or secondary. However, the adjective ‘metastatic’ is used in two ways, sometimes meaning a secondary from a primary elsewhere and sometimes denoting a primary that has given rise to metastases.

(a) Malignant neoplasm ‘metastatic from’ a specified site

If a malignant neoplasm is described as ‘metastatic from’ a specified site, or if a ‘due to’ relationship implies a spread from a specified site, that site should be considered primary. This also applies to sites on the list of common sites of metastases.

(b) Malignant neoplasm ‘metastatic to’ a specified site

If a malignant neoplasm is described as ‘metastatic to’ a specified site, or if a ‘due to’ relationship implies a spread to a specified site, that site should be considered secondary, whether the site is on the list of common sites of metastases or not. However, if a morphology classifiable to categories primarily subdivided by histopathology is reported, code to the ‘unspecified site’ subcategory of that morphological type.

(c) Malignant neoplasm metastatic of site A to site B

A malignant neoplasm described as metastatic of site A to site B should be interpreted as primary of site A and secondary of site B.

(d) ‘Metastatic’ malignant neoplasm on the list of common sites of metastases

Except for lung, code a ‘metastatic’ neoplasm of a site on the list of common sites of metastases as secondary, even if no other neoplasm is mentioned on the certificate. For ‘metastatic’ neoplasm of lung, see Section 6.3.5 G, Metastatic cancer, subsection (f), ‘Metastatic’ cancer of lung).

Exception: code a neoplasm of a site on the list of common sites of metastases as primary when it is reported as due to a condition that increases the risk of a malignant neoplasm of that site or tissue.

Exception: code a neoplasm of a site on the list of common sites of metastases as primary if it is the only malignant neoplasm mentioned on the certificate.

(e) ‘Metastatic’ malignant neoplasm not on the list of common sites of metastases

If the only malignant neoplasm is specified as ‘metastatic’ and the site is not on the list of common sites of metastases, then code as primary malignant neoplasm of that particular site.

If one or more neoplasms specified as ‘metastatic’ are reported on the certificate and there is also another malignant neoplasm that is not specified as ‘metastatic’, then code the neoplasm not specified as ‘metastatic’ as primary and the ones specified as ‘metastatic’ as secondary. This applies also to neoplasms not on the list of common sites of metastases, if specified as metastatic.

| Example 37: |||| |--|--|--| ||1|(a) | Bladder cancer | | | |(b) | Metastatic prostate cancer |

Code as Secondary prostate cancer and primary bladder cancer. The order of entry does not impact on the coding.

(f) ‘Metastatic’ cancer of lung

If the only malignant neoplasm mentioned is ‘metastatic’ neoplasm of lung, code to primary Malignant neoplasm of lung. Also code a ‘‘metastatic’ neoplasm of lung as Primary malignant neoplasm of lung, if all other neoplasm sites reported on the death certificate are on the list of common sites of metastases. If another malignant neoplasm is mentioned that is not on the list of common sites of metastases, then code a ‘metastatic’ malignant neoplasm of lung as Secondary malignant neoplasm of lung.

(g) ‘Metastatic’ neoplasm of a specific morphology

If the certificate reports a malignant neoplasm specified as ‘metastatic’ of a morphological type classifiable to a cancer category that mentions a specific histopathology only, and the site reported is consistent with the morphological type, then code to a primary malignant neoplasm of the specified morphological type. Use the appropriate site subcategory for the specified morphological type or site.

If the ‘metastatic’ cancer reported on the certificate and the site is not consistent with the morphological type, then code to a secondary malignant neoplasm of the specified site. Also add a code for a primary malignant neoplasm of unspecified site for the stated morphological type.

| Example 38: |||| |--|--|--| ||1|(a) | Osteosarcoma of sacrum, metastatic |

The site sacrum is consistent with a primary cancer of bone. Code as primary osteosarcoma of sacrum.

| Example 39: |||| |--|--|--| ||1|(a) |Osteosarcoma of kidney, metastatic |

Code osteosarcoma of kidney as a Secondary malignant neoplasm, because the specified site (kidney) is not consistent with osteosarcoma, which is primary in bone. Also code in the cluster, Osteosarcoma of unspecified site.

Sequelae

A. Sequelae of tuberculosis

Sequelae of tuberculosis include conditions specified as such or as arrested, cured, healed, inactive, old or quiescent, unless there is evidence of active tuberculosis. It does not include chronic tuberculosis, which should be coded as active infectious disease.

B. Sequelae of trachoma

Sequelae of trachoma include residuals of trachoma specified as healed or inactive and certain specified sequelae, such as blindness, cicatricial entropion and conjunctival scars, unless there is evidence of active infection. It does not include chronic trachoma, which should be coded as active infectious disease.

C. Sequelae of viral encephalitis

Sequelae of viral encephalitis include conditions specified as such, and late effects present one year or more after onset of the causal condition. It does not include chronic viral encephalitis, which should be coded as active infectious disease.

D. Sequelae of other infectious and parasitic diseases

Sequelae of other infectious and parasitic diseases include conditions specified as such or as arrested, cured, healed, inactive, old or quiescent. Sequelae also include conditions present one year or more after onset of conditions classifiable to categories, unless there is evidence of active disease. It does not include chronic infectious and parasitic diseases, which should be coded as active infectious and parasitic disease.

E. Sequelae of rickets

Sequelae of rickets include conditions stated to be a sequela or late effect of rickets, or previous rickets as the cause of conditions present one year or more after onset of rickets. It does not include chronic malnutrition or nutritional deficiency, which should be coded to current malnutrition or nutritional deficiency.

Consistency between sex of patient and diagnosis

Most categories of ICD–11 apply to persons of both sexes. However, some diseases are more likely to occur in one sex than in the other. A list of those conditions is given in the Annex.

If there is an apparent inconsistency between the sex of the deceased and the code selected for a cause of death reported on the certificate, then the coder should check the information and make sure that no reporting error occurred.

Follow any additional information provided by the certifier. If it turns out that the code is in fact correct, in spite of the apparent inconsistency, then the code should be kept. In such cases, it might be useful to add a note to the statistics that the reported cause of death has been verified and is correctly reported and coded. If no additional information can be obtained and the reported cause of death is fully incompatible with the sex of the deceased and there is no indication of sex-change treatment, then code, Other ill-defined and unspecified causes of mortality. In such cases, a note can be added to the statistical publication, specifying the number of cases recoded to that category because of sex and cause inconsistencies that could not be verified.

Specific instructions on other ICD categories

A. Rheumatic fever with heart involvement

If there is no statement that the rheumatic process was active at the time of death, assume activity if the heart condition (other than terminal conditions and bacterial endocarditis) that is specified as rheumatic, or stated to be due to rheumatic fever, is described as acute or subacute. In the absence of such description, the terms ‘carditis’, ‘endocarditis’, ‘heart disease’, ‘myocarditis’ and ‘pancarditis’ can be regarded as acute if either the interval from onset is less than one year or, if no interval is stated, at ages under l5 years. ‘Pericarditis’ can be regarded as acute at any age.

B. Pneumonia and immobility

Code pneumonia, organism unspecified reported with immobility or reduced mobility to Hypostatic pneumonia, unspecified.

C. Obstetric death of unspecified cause, Obstetric deaths 42 days 1 year after delivery, sequelae of direct obstetric causes

These categories classify obstetric deaths according to the time elapsed between the obstetric event and the death of the woman. These categories should be complemented in the cluster by a code indicating the specific cause of death with a code from the appropriate chapter of ICD.

D. Perinatal deaths

Use a code from Chapter 19, Certain conditions originating in the perinatal period, if:

1. the condition is indexed to a code in Chapter 19;
2. there is an index entry for the specified condition as congenital/perinatal/newborn, and the duration of the condition indicates that the condition developed in the neonatal or perinatal period. This applies even if the condition is not specified as neonatal or perinatal on the certificate.

For some conditions diagnosed below a specific age, it is assumed that the condition was congenital, see the following section, Congenital malformations, deformations and chromosomal abnormalities.

Further, for children less than 28 days old, assume that a reported condition developed in the perinatal period, unless the duration is stated and the onset was after the first completed week of life.

Note that some types of conditions are excluded from Chapter 19, such as:

• Tetanus neonatorum
• Congenital gonococcal infection
• Congenital syphilis
• HIV disease
• Infectious diseases acquired after birth
• Intestinal infectious diseases
• Neoplasms
• Hereditary haemolytic anaemia
• Transient hypogammaglobulinaemia of infancy
• Endocrine, nutritional and metabolic diseases
• Certain congenital diseases of the nervous system
• Congenital cardiomyopathy
• Intestinal obstruction or paralytic ileus
• Pemphigus neonatorum and Staphylococcal scalded skin syndrome (L00)
• Cradle cap
• Diaper [napkin] dermatitis
• Developmental anomalies
• Injury, poisoning and certain other consequences of external causes

E. Developmental anomalies 

Conditions classified as Developmental anomalies should be coded as such if the duration of the condition indicates that it existed from birth, even if the condition is not specified as congenital on the certificate. Further, the following conditions should be coded congenital at the ages stated, provided there is no indication that they were acquired after birth. 1. Under l year: aneurysm, aortic stenosis, atresia, atrophy of brain, cyst of brain, deformity, displacement of organ, ectopia, hypoplasia of organ, malformation, pulmonary stenosis, valvular heart disease. Under 4 weeks: heart disease NOS, hydrocephalus NOS.

F. Complications of surgical and medical care

Whenever a complication of a procedure is not indexed or is not a synonym of an inclusion or indexed term, code early complications and mechanical complications to the appropriate section in chapter 20, ‘Injury or harm arising from surgical and medical care, not elsewhere classified’. Code late complications and longstanding complications of organ function to the appropriate system chapter.

Routine use and special cases

Routine cause of death

In routine cause of death reporting systems, every individual death is certified by a qualified medical doctor who carries out an accurate post mortem examination, collects history from relatives, and has access to all pre-existing medical information about the defunct. The medical certification of the cause of death is usually the responsibility of the attending physician, and should be in line with international recommendations. Administrative procedures should ensure confidentiality of data from death certificates or other medical records.

In the case of deaths certified by coroners or other legal authorities, the medical evidence supplied to the certifier should be stated on the certificate in addition to any legal findings.

Routine cause of death reporting is usually embedded in the certification of death process. Death certificates are a legal requirement for burial and for inheritance.

Verbal autopsy

Verbal autopsy (VA) is a method used to ascertain the cause of a death based on an interview with next of kin or other caregivers. This is done using a standardized questionnaire that elicits information on signs, symptoms, medical history, and circumstances preceding death. The cause of death, or the sequence of causes that led to death, are assigned based on the data collected by the questionnaire and other available information. Rules and guidelines, algorithms or computer programs, may assist in evaluating the information to determine the cause of death (11). The main objective of the VA is to describe the causes of death at the community or population level in areas, where civil registration and death certification systems are weak and where most people die at home without having had contact with the health system. A standard VA instrument comprises a VA questionnaire, cause of death or mortality classification system, and diagnostic criteria (either expert or data derived algorithms) for deriving causes of death. The VA process consists of interviews, data recording, and identification of the cause of death from the reports. At any step, factors can influence the cause-specific mortality fractions estimated throughout the process. Besides research, VA is a viable method for causes of death identification in settings where no physician can evaluate the deceased.

Maternal mortality

A maternal death is defined as the death of a woman while pregnant or within 42 days of termination of pregnancy, irrespective of the duration and site of the pregnancy, from any cause related to or aggravated by the pregnancy or its management, but not from accidental or incidental causes. Maternal deaths are subdivided into two groups.

• Direct obstetric deaths: those resulting from obstetric complications of the pregnant state (pregnancy, labour, and puerperium), and from interventions, omissions, incorrect treatment, or from a chain of events resulting from any of the above.
• Indirect obstetric deaths: those resulting from previous existing disease or disease that developed during pregnancy and not due to direct obstetric causes, but were aggravated by the physiologic effects of pregnancy.

In order to improve the quality of maternal mortality data and provide alternative methods of collecting data on deaths during pregnancy or anything related to pregnancy, as well as to encourage the recording of deaths from obstetric causes occurring more than 42 days following termination of pregnancy, the Forty-third World Health Assembly in 1990 adopted the recommendation that countries consider including questions regarding current pregnancy and pregnancy within one year preceding death on death certificates.

Maternal mortality is part of the Millennium Development Goals (MDGs), and of the Sustainable Development Goals (SDG) that serve to monitor the impact of the joint work of the international community in this field.

If pregnancy, childbirth, or puerperium is mentioned anywhere on the certificate, in most cases the underlying cause is coded to Chapter 18 (Pregnancy, childbirth and the puerperium). This is either because the underlying cause selected by applying Step SP1–M4 is classified to Chapter 18 according to the Alphabetical index, or because there is a special code in Chapter 18 for the condition if it appears during pregnancy, childbirth and the puerperium. Apply the following instructions to determine if an underlying cause which is indexed to other parts of the ICD should be classified to Chapter 26, Extension codes. Note that these instructions do not apply to conditions that are indexed to Chapter 18 in the Alphabetical index.

• If pregnancy, childbirth or puerperium is reported anywhere on the certificate but it is not clearly stated that pregnancy, childbirth or puerperium contributed to the death, first contact the certifier and ask for additional information.
• If the certifier states that the death was a complication of pregnancy, childbirth or puerperium, code the underlying cause to Chapter 18, Pregnancy, childbirth and the puerperium.
• If the certifier states that the death was not a complication of pregnancy, childbirth or puerperium, do not code the underlying cause to Chapter 18, Pregnancy, childbirth and the puerperium.
• If you cannot obtain any additional information, but pregnancy, childbirth or puerperium is mentioned in Part 1 or Part 2 of the certificate, code the underlying cause to Chapter 18, Pregnancy, childbirth and the puerperium.
• If the underlying cause you selected is classifiable to ‘Maternal infectious and parasitic diseases classifiable elsewhere but complicating pregnancy, childbirth and the puerperium’ and ‘Other maternal diseases classifiable elsewhere but complicating pregnancy, childbirth and the puerperium’, then add to the cluster the corresponding code from Chapter 1-18 as a multiple cause of death. This is important because otherwise crucial information on the death will not be retrievable.

Note that some conditions are not coded to Chapter 18, Pregnancy, childbirth and the puerperium, even if they occurred during pregnancy, childbirth or puerperium, see the Excludes note at the beginning of Chapter 18, Pregnancy, childbirth and the puerperium.

| Example 1: |||| |--|--|--| ||1|(a)| Amniotic fluid embolism|

The underlying cause, amniotic fluid embolism, is indexed to Chapter 18

| Example 2: |||| |--|--|--| ||1|(a)| Pulmonary oedema| ||||caused by| |||(b)| Mitral regurgitation, pregnancy|

The underlying cause, mitral regurgitation, is coded to Chapter 18 because pregnancy is mentioned in Part 1. Code the underlying cause to diseases of the circulatory system complicating pregnancy, childbirth and the puerperium. Also add to the cluster the code for mitral regurgitation as a contributing cause of death.

| Example 3: |||| |--|--|--| ||1|(a)| Haemorrhage| ||||caused by| |||(b)| Cervical cancer Treatment delayed because of pregnancy|

The underlying cause, cervical cancer, is coded to Chapter 18 because pregnancy is mentioned in Part 2. Code the underlying cause to other specified diseases and conditions complicating pregnancy, childbirth and the puerperium. Also add to the cluster the code for cervical cancer as a contributing cause of death.

| Example 4: |||| |--|--|--| ||1|(a)| Hepatic failure| ||||caused by| |||(b)| Dengue haemorrhagic fever 5 days| |||Additional information:|40 days postpartum|

Code the underlying cause to other viral diseases complicating pregnancy, childbirth and the puerperium. Also add to the cluster the code for dengue haemorrhagic fever as a contributing cause of death.

Infant mortality

Infant mortality refers to death of children under the age of 1 year and older than 28 days. It is an indicator for quality of life and health infrastructure.

Child and infant mortality

Child mortality (Under-5 mortality rate – probability of dying by age of 5 years, and infant mortality rate – probability of dying by age of 1 year) is a leading indicator of the level of child health, quality of life, health infrastructure, and overall development in countries. It is also the MDG indicator number 4.

Perinatal mortality

The legal requirements for the registration of fetal deaths and live births vary from country to country and between countries. If possible, all fetuses and infants weighing at least 500 g at birth, whether alive or dead, should be included in the statistics. When information on birth weight is unavailable, the corresponding criteria for gestational age (22 completed weeks) or body length (25 cm crown-heel) should be used. The criteria for deciding whether an event has taken place within the perinatal period should be applied in the order:

1. birth weight,
2. gestational age,
3. crown-heel length.

The inclusion of fetuses and infants weighing between 500 g and 1000 g as stated in national statistics is recommended both because of its inherent value and because it improves the coverage of reporting at 1000 g and over. For perinatal mortality statistics, full-scale multiple-cause analysis of all conditions reported will be of the greatest value. Whenever possible, a separate certificate of cause of perinatal death should be completed, in which the causes are to be set out as follows:

(a) main disease or condition in foetus or infant

(b) other diseases or conditions in foetus or infant

(c) main maternal disease or condition affecting foetus or infant

(d) other maternal diseases or conditions affecting foetus or infant

(e) other relevant circumstances.

The reliability of the neonatal mortality estimates depends on accuracy and completeness of reporting and recording of births and deaths. Under-reporting and misclassification are common, especially for deaths occurring in newborns.

Perinatal mortality: guidelines for certification

With the update of the International form of medical certificate of cause of death in 2016, just one certificate is used for all cases (see Annex 14.1). Care needs to be taken to correctly fill in the specific section for perinatal deaths on the certificate.

The previously recommended perinatal death certificate should be replaced by the form in Annex 14. 1..

---

Main Uses of the ICD: Morbidity

Morbidity data are used for statistical reporting mostly on national or local levels. While some of this statistical reporting is conducted within an academic research context, it is commonly conducted in applied settings to inform health system and public health agency decision-making. ICD coded data also form the basis of different casemix systems such as different varieties of Diagnosis Related Groups (DRGs). Coded morbidity data can also be used to inform a variety of clinical guidelines through provision of foundational information on burden of disease.

What is coded: Conditions of patient

The health care practitioner responsible for the patient's treatment is also responsible for documenting the patient's health conditions. This information should be organized systematically by using standard recording methods. A properly completed record is essential for good patient management. It is also an essential prerequisite to the creation of a valid coded record of patient diagnoses, derived through a coding process from written information describing a patient's medical condition. When a sound written record of patient conditions is available, successful coding of this information in ICD and associated classifications produces a valuable source of epidemiological and other statistical data on morbidity and other health care problems. The person transforming the information on the stated condition to codes (the ‘coder’) may be the health care practitioner or a clinical coder who is not responsible for the patient’s treatment). In the latter situation, which is quite common among member countries, the coder depends on the adequacy of clinical documentation of patient condition by health care practitioners in the medical record. The primary importance of clinical documentation by health care practitioners as the starting point for coded health data cannot be overstated, and needs to be underlined as being a matter of key importance within countries and internationally – with implications for health information and clinical documentation teaching within health care practitioner training programs.

Documentation principles related to morbidity coding

Morbidity data are increasingly being used in the formulation of health policies and programs, and in their management, monitoring and evaluation, in epidemiology, in identification of risk populations, and in clinical research (including studies of disease occurrence in different socioeconomic groups). Record as the main condition the condition that is determined to be the reason for admission, established at the end of the episode of health care. That definition of main condition is to be applied for both, inpatients and outpatients. (Importantly, and as mentioned earlier, this is a change in the WHO's main condition definition that existed in ICD–10.) In addition to the main condition, the record should, whenever possible, also list separately other conditions or problems dealt with during the episode of health care. Other conditions are defined as those conditions that coexist or develop during the episode of health care and affect the management of the patient. It is recommended, where practicable, to carry out multiple-condition coding and analysis to supplement the routine data.

Guidelines for recording diagnostic information for morbidity data analysis

The health-care practitioner responsible for the patient’s treatment should select the main condition to be recorded, as well as any other conditions, for each episode of health care.The term episode is used for all settings, including hospital admissions. It is acknowledged that the definition may be different in each country, though it is most often considered to be a contiguous hospital care period, which begins on the first day of a person’s admission to a care facility and ends on the day upon which they are discharged from that facility. (Some countries consider sequential care periods on different wards within the same hospital to be distinct episodes of care.)

In the context of these morbidity coding rules, the term practitioner is used throughout the morbidity rules to mean physician or any qualified health care practitioner who is legally accountable for establishing the patient’s diagnosis. This information should be organized systematically by using standard recording methods. A properly completed record is essential for good patient management and is a valuable source of epidemiological and other statistical data on morbidity and other health-care problems.

Specificity and detail

Each diagnostic statement should be as informative as possible in order to classify the condition to the most specific ICD category. Examples of such diagnostic statements include: - transitional cell carcinoma of trigone of bladder - acute appendicitis with localized peritonitis - meningococcal pericarditis - pregnancy-induced hypertension - diplopia due to reaction to antihistamine taken as prescribed - osteoarthritis of hip due to an old hip fracture - fracture of neck of femur following a fall at home - full thickness burn of palm of left hand due to grilling accident

Unconfirmed diagnoses

If no definite diagnosis has been established by the end of an episode of healthcare, then the information that permits the greatest degree of specificity and knowledge about the condition that necessitated care or investigation should be recorded. This should be done by stating a symptom, abnormal finding or problem, and also reporting the unconfirmed diagnosis qualifying it as ‘rule out’, ‘or’provisional diagnosis’, when it has been considered but not established. In case the unconfirmed diagnosis has resulted in treatment, it should be mentioned first – together with the extension code ‘provisional’. In case the unconfirmed diagnosis has in monitoring only, it should be mentioned second – together with the extension code ‘rule out’

Contact with health services for reasons other than illness

Episodes of health care or contact with health services are not restricted to identification, treatment or investigation of current illness or injury. Episodes may also occur when someone who may not currently be sick requires or receives limited care or services; the details of the relevant circumstances should be recorded as the ‘main condition’. Examples include: - monitoring of previously treated conditions - immunization - contraceptive management, antenatal and postpartum care - surveillance of persons at risk because of personal or family history - examinations of healthy persons, e.g. for insurance or occupational reasons - seeking of health-related advice - requests for advice by persons with social problems - consultation on behalf of a third party. Chapter 24 (Factors influencing health status and contact with health services) provides a broad range of categories (QA00–QF9Z) for classifying these circumstances; reference to this chapter will give an indication of the detail required to permit classification to the most relevant category.

Conditions due to external causes

When a condition such as an injury, poisoning or other effect of external causes is recorded, it is important to document fully both the nature of the condition and the circumstances that gave rise to it. For example: ‘fracture of neck of femur caused by fall due to slipping on pavement’; ‘cerebral contusion caused when patient lost control of car, which hit a tree’; ‘accidental poisoning –patient drank disinfectant in mistake for soft drink’; or ‘severe hypothermia –patient fell in her garden in cold weather’.

Documentation of sequelae

Where an episode of care is for the treatment or investigation of a residual condition (sequela) of a disease that is no longer present, the sequela should be fully described and its origin stated, together with a clear indication that the original disease is no longer present. For example: ‘deflected nasal septum– fracture of nose in childhood’, ‘contracture of Achilles tendon – late effect of injury to tendon’, or ‘infertility due to tubal occlusion from old tuberculosis’. Where multiple sequelae are present and treatment or investigation is not directed predominantly at one of them, a statement such as ‘sequelae of cerebrovascular accident’ or ‘sequelae of multiple fractures’ is acceptable.

Guidelines for selecting ‘main condition’ and ‘other conditions’ for coding purposes

The main condition is the condition that is determined to be the reason for admission, established at the end of the episode of health care by the practitioner. If more than one condition is recorded as the main condition, see coding rule MB1 in the section below on reselection of a main condition when the original selection is incorrect. Whenever possible, a record with an obviously inconsistent or incorrectly recorded main condition should be returned for clarification.

Where an episode of health care concerns a number of related conditions present at admission and contributing to the need for admission (e.g. multiple injuries, multiple sequelae of a previous illness or injury, or multiple conditions occurring in human immunodeficiency virus [HIV] disease), the one that is mentioned first should be recorded as the ‘main condition’ and the others as ‘other conditions’.

The ‘main condition’ and ‘other conditions’ relevant to an episode of health care should have been recorded by the responsible health-care practitioner. If clarification of potential erroneous documentation is not possible, Rules MB1 to MB3 (Section on morbidity rules below) will help the coder to deal with some of the more common causes of incorrect recording. The guidelines given below are for use when the coder may be unclear as to which code should be used.

For clinical and resource allocation purposes, in many instances, the manifestation of a disease (kind and severity e.g. ulcer stage 3) may be more relevant during a specific treatment episode than the underlying disease (e.g. Diabetes mellitus). For prevention programs at national levels, knowledge about the underlying aetiology may be more important. Quality and safety will require reporting additional detail related to the stay. For comprehensive analysis and use of morbidity data, it is crucial to have a dataset with multiple fields covering all the aspects above.

Type 2 extension codes (a new section of codes in ICD–11) will provide a number of distinct codes that serve as concept modifying flags for marking how the diagnosis is to be used and/or interpreted. These extension code modifiers include:

• Reason for admission
• Most resource intensive condition
• Tentative (provisional) diagnosis
• Differential diagnosis
• Rule-out diagnosis
• Present on Admission
• Arising after admission during hospital stay

For more detail about the use of extension codes, which identify additional characteristics about the diagnoses, see section 4.2.7.

Example 1:

A patient is admitted with a myocardial infarction, for which he stays in hospital for 5 days without any notable complications. Myocardial infarction & ‘reason for admission as determined after the stay’ (Note that the above cluster using an ‘&’ to cluster the myocardial infarction stem code with the extension code for ‘reason for admission determined after the stay’.)

Example 2:

A patient is admitted to hospital with myocardial infarction and then develops a stroke that leads to a one month hospitalization. Myocardial infarction is the main condition because it was the reason for admission. Stroke is then marked with a cluster-linked diagnosis-type extension code flag indicating that the diagnosis arose after admission. i.e. ‘Myocardial infarction’ & ‘reason for admission as determined after the stay’ ‘Stroke’ & ‘developed after admission’

Example 3:

A patient is admitted to the hospital with pneumonia and congestive heart failure. The final diagnosis recorded by the responsible health care practitioner is:

Pneumonia

Congestive Heart Failure

There are two possible main conditions. Pneumonia is selected as the main condition because it was mentioned first. As mentioned earlier, a coder must attempt to judge which of the two candidate conditions is most severe or consuming a greater amount of resources, based on the provider's documentation, to thus be selected as the main condition. If presence at admission, severity or resource use are difficult to discern between possible main conditions, the first one is selected.

Such a system with diagnosis flags meets the objectives of countries that want a reason-for-admission coding rule, while also meeting the objectives of countries that want to be able to make inferences regarding complications of care and resource consumption (of relevance to casemix systems). As mentioned earlier, there are some health care episodes that are not related to the treatment of or investigation of current illness of injury (e.g., monitoring of previously-treated conditions, immunization visits, seeking of health-related advice). In such situations, a code for the main condition can potentially be found in ICD–11 Chapter 24 (Factors influencing health status and contact with health services).

The ‘main condition’ and ‘other conditions’ relevant to an episode of health care should be recorded by the most responsible health care practitioner. Straightforward coding should follow. Whenever possible, a record with an inconsistent or incorrectly recorded main condition should be returned for clarification. Failing clarification, Rules MB1 to MB3 will help the coder to manage some of the more common causes of incorrect recording. The guidelines given below are for use when the coder may be unclear as to which code should be used.

It has been recommended that ‘other conditions’, in relation to an episode of care, should be recorded in addition to the main condition, even for single-cause analysis, since this information may assist in choosing the correct ICD code for the main condition.

Coding of conditions with cluster coding

A significant new feature in ICD–11 is an embedded functionality for the linking (or clustering) of related diagnostic concepts to better capture the clinical narrative surrounding an episode of care. This so-called post-coordination of diagnostic concepts is now possible in ICD–11. The post-coordinated coding of conditions is shown as an example here. For additional detail see also section 5.

Case is reported with:

Condition 1. diabetes mellitus type 2

Condition 2. diabetic retinopathy

Condition 3. arterial hypertension.

Code condition 1 as cluster as follows: 6A11 Type 2 diabetes mellitus / ME90.Z Diabetic retinopathy, unspecified 

Code condition 2 as follows: CA10 Essential hypertension

Coding of suspected conditions, symptoms, abnormal findings and non-illness situations

If the period of health care was for an inpatient, the coder should be cautious about classifying the main condition to Chapters 21 and 24. If a more specific diagnosis has not been made by the end of the inpatient stay, or if there was truly no codable current illness or injury, then codes from the above chapters are permissible (see also Rules MB3 and MB5, Section 8.2.2.9). The categories can be used in the normal way for other episodes of contact with health services. If, after an episode of health care, the main condition is still recorded as ‘suspected’, ‘questionable’, etc., and there is no further information or clarification, the suspected diagnosis must be coded as if established.

The Category for ‘Medical observation and evaluation for suspected diseases and conditions’ applies to suspected diagnoses that were ruled out after investigation.

Example 4:
	Main condition:	Suspected acute cholecystitis
	Other conditions:	-

Code to acute cholecystitis as ‘main condition’.

Example 5:
	Main condition:	Admitted for investigation of suspected malignant neoplasm of cervix – ruled out

Code to observation for suspected malignant neoplasm as ‘main condition’, add to the cluster the prefix code ‘ruled out’ from chapter 26, Extension codes.

Example 6:
	Main condition:	Severe epistaxis
	Other conditions:	-

Patient in hospital one day. No procedures or investigations reported.

Code to epistaxis. This is acceptable since the patient was obviously admitted to deal with the immediate emergency only.

Coding of multiple conditions

Where multiple conditions are recorded in the documentation as ‘Multiple ...’, and no single condition predominates, the code for the ‘Multiple ...’ category should be used as the preferred code, where data system can retain one code only. Additional codes are added to the cluster for individual conditions listed where the data system can handle the cluster coding. Ideally all the multiple conditions are coded as a cluster (see cluster coding). Such coding applies mainly to conditions associated with HIV disease, to injuries and sequelae.

Coding of combination categories

The ICD provides certain categories where two conditions or a condition and an associated secondary process can be represented by a single code. Such combination categories should be used as the main condition where appropriate information is recorded. The Alphabetical Index indicates where such combinations are provided for, under the indent ‘associated with’, which appears immediately after the lead term. Two or more conditions recorded under ‘main condition’ may be linked if one of them may be regarded as an adjectival modifier of the other.

Example 7:
	Main condition:	Renal failure
	Other conditions:	Hypertensive renal disease

Code to hypertensive renal disease with renal failure and add to the cluster a code for chronic renal failure, stage unspecified.

Example 8:
	Main condition:	Glaucoma secondary to eye inflammation
	Other conditions:	-

Code to glaucoma due to eye infection.

Example 9:
	Main condition:	Intestinal obstruction
	Other conditions:	Left inguinal hernia

Code to unilateral or unspecified inguinal hernia and add to the cluster a code for intestinal obstruction.

Example 10:
	Main condition:	Diabetic cataract. Type 1 diabetes mellitus
	Other conditions:	Hypertension
	Specialty:	Ophthalmology

Code the Type 1 diabetes mellitus and add to the cluster a code for ‘diabetic cataract’.

Example 11:
	Main condition:	Type 2 diabetes mellitus
	Other conditions:	Hypertension
		Rheumatoid arthritis
		Cataract
	Specialty:	General medicine

Code to type 2 diabetes mellitus as ‘main condition’. Note that in this example the linkage of cataract with diabetes must not be made since they are not both recorded under ‘main condition’.

Coding of external causes of morbidity

For injuries and other conditions due to external causes, both the nature of the condition and the circumstances of the external cause should be coded. The preferred ‘main condition’ code should be that describing the nature of the condition. This will often, but not always, be classifiable to Chapter 22. The code from Chapter 23 indicating the external cause would be used as an optional additional code.

Example 12:
	Main condition:	Fracture of neck of femur caused by fall due to tripping on uneven pavement
	Other conditions:	Contusions to elbow and upper arm

Code to fracture of neck of femur (S72.0) as ‘main condition’. The external cause code for fall on same level from slipping, tripping or stumbling on street or highway (W01, place of occurrence 4) may be used as an optional additional code linked to the fracture code through clustering in a string code representation.

Example 13:
	Main condition:	Severe hypothermia- patient fell in her garden in cold weather
	Other conditions:	Senility

Code to hypothermia as ‘main condition’ and add to the cluster the external cause code for exposure to excessive natural cold and for site ‘at home’.

Example 14:
	Main condition:	Diplopia due to reaction to antihistamine taken as prescribed
	Other conditions:	-

Code to diplopia as the ‘main condition’ and add to the cluster the external cause code for antiallergic and antiemetic drugs causing adverse effects in therapeutic use and the code for mode of harm ‘medication taken as prescribed at correct dose’.

Example 15:
	Main condition:	Haemoglobinuria caused by training for marathon run (training on outdoor track at stadium)
	Other conditions:	-

Code to haemoglobinuria due to haemolysis from other external causes as ‘main condition’ and add to the cluster the external cause code for overexertion and strenuous, repetitive movements at sports and athletics area.

Coding of acute and chronic conditions

Where the condition is recorded as being both acute (or subacute) and chronic, and the ICD provides separate categories or subcategories for each, but not for the combination, the category for the acute condition should be sequenced first, followed by the chronic condition. When an appropriate combination code is provided for both the acute and chronic condition, assign only the combination code.

Example 16:	Main condition:	Acute and chronic cholecystitis
	Other conditions:

Code to acute cholecystitis as the ‘main condition’. The code for chronic cholecystitis is added to the cluster.

Example 17:	Main condition:	Acute exacerbation of chronic obstructive bronchitis
	Other conditions:

Code to chronic obstructive pulmonary disease with acute exacerbation as the ‘main condition’ since the ICD provides an appropriate single pre-coordinated code for the combination.

Coding of postprocedural conditions and complications

For certain acute complications of surgical and other procedures Chapter 21 provides categories, for example surgical wound infections, mechanical complications of implanted devices, or shock. Most body-system chapters also contain categories for permanent conditions that occur either as a consequence of specific procedures and techniques or as a result of the removal of an organ, e.g. postmastectomy lymphoedema syndrome, post-irradiation hypothyroidism. Some conditions (e.g. pneumonia, pulmonary embolism) that may arise in the postprocedural period are not considered unique entities and are, therefore, coded in the usual way, but an optional additional code from PC70-PD5Z may be added to identify the relationship to a procedure, and an extension code indicating the timing of onset of a condition diagnosed during a hospital stay (e.g. 'diagnosis arose after admission'). When postprocedural conditions and complications are recorded as the main condition, reference to modifiers or qualifiers in the Alphabetical Index is essential for choosing the correct code.

Example 18:	Main condition:	Hypothyroidism since thyroidectomy 1 year ago
	Other conditions:	-
	Specialty:	General medicine

Code to postsurgical hypothyroidism as the ‘main condition’ and add to the cluster a code for ‘endocrine procedure as the cause of injury’. (Note: A third mode/mechanism code associated with surgical procedure as a cause of injury can also be included in the cluster, if a discrete mode of injury is apparent in the medical record. Mode/mechanism codes associated with health care interventions are discussed in detail below in section 8.4.3.)

Example 19:	Main condition:	Excessive haemorrhage after tooth extraction
	Other conditions:	Pain
	Specialty:	Dentistry

Code to haemorrhage resulting from a procedure as the ‘main condition’. And add to the cluster a code for ‘Dental procedure as the cause of injury’. (A third mode/mechanism code associated with surgical procedure as the cause of injury can also be included, if a specific mode is apparent.)

Example 20:	Main condition:	Postoperative psychosis after plastic surgery
	Other conditions:	-
	Specialty:	Psychiatry

Code to psychosis as the ‘main condition’ and add to the cluster a code for other specified surgical procedures [as the cause of abnormal reaction of the patient]) to indicate the postprocedural relationship. The related diagnoses can then be linked through the available clustering mechanism.

Rules for reselection when the main condition is incorrectly recorded

The responsible health-care practitioner indicates the ‘main condition’ to be coded, and this should normally be accepted for coding subject to the guidelines above and in the chapter-specific notes. However, certain circumstances or the availability of other information may indicate that the health- care practitioner has not followed the correct procedure. In this instance, clarification of the main condition should be obtained from the responsible health care practitioner. If it is not possible to obtain clarification from the health-care practitioner, one of the following rules may be applied and the ‘main condition’ reselected.

Rules for reselection of main condition

Rule MB1. Several conditions recorded as ‘main condition’

If several different conditions (they cannot be classified to a single stem code) are recorded as the ‘main condition’, and other details on the record point to one of them as the ‘main condition’ for which the patient received care, select that condition. If there is the ability to record more than one main condition (for example, in addition to the ‘reason for admission’ reporting also ‘main resource condition’), a code(s) from Chapter 26, Extension codes, should be assigned to indicate the different types of main condition that are reported. If the "main condition" cannot be determined based on documentation, query the provider. If the setting does not allow for provider query, then select the condition first mentioned.

Example 21:
	Main condition:	Cataract. Staphylococcal meningitis. Ischaemic heart disease
	Other conditions:	-
	Specialty:	Neurology
> Patient in hospital for five weeks.

Code staphylococal meningitis as the 'main condition'.

Example 22:
	Main condition:	Chronic obstructive bronchitis. Hypertrophy of prostate. Psoriasis vulgaris
	Other conditions:	-
	Specialty:	Outpatient in the care of a dermatologist

Code psoriasis vulgaris as the 'main condition'.

Example 23:
	Main condition:	Mitral stenosis. Acute bronchitis. Rheuematoid arthritis
	Other conditions:	-
	Specialty:	General medicine
> No information about therapy.

Code mitral stenosis, the first mentioned condition, as the 'main condition'.

Example 24:
	Main condition:	Chronic gastritis. Secondary malignancy in axillary lymph nodes. Carcinoma of breast
	Other conditions:	-
	Procedure:	Mastectomy

Code malignant neoplasm of breast as the 'main condition'.

Example 25:
	Main condition:	Premature rupture of membranes. Breech presentation. Anaemia
	Other conditions:	-
	Procedure:	Spontaneous delivery

Code premature rupture of membranes unspecified, the first mentioned condition, as the 'main condition'.

Rule MB2. Condition recorded as ‘main condition’ is presenting symptom of diagnosed, treated condition If a symptom or sign (usually classifiable to Chapter 21), or a problem classifiable to Chapter 24, is recorded as the ‘main condition’ and this is obviously the presenting sign, symptom or problem of a diagnosed condition recorded elsewhere and care was given for the latter, reselect the diagnosed condition as the ‘main condition’.

Example 26:	Main condition:	Haematuria
	Other conditions:	Varicose veins of legs
		Papillomata of posterior wall of bladder
	Treatment:	Diathermy excision of papillomata
	Specialty:	Urology

The haematuria is caused by the papillomata of the bladder, which is the focus of treatment by excision. Reselect papillomata of posterior wall of bladder as the ‘main condition’ and code accordingly.

Example 27:	Main condition:	Coma
	Other conditions:	Ischaemic heart disease
		Otosclerosis
		Type I diabetes mellitus
	Specialty:	Endocrinology
	Care:	Establishment of correct dose of insulin

Reselect type 1 diabetes mellitus as the ‘main condition’ and add a code for diabetic coma as the first entry in the cluster. The information provided indicates that the coma was due to diabetes mellitus and coma is taken into account as it modifies the coding.

Example 28:	Main condition:	Abdominal pain
	Other conditions:	Acute appendicitis
	Procedure:	Appendectomy

Reselect acute appendicitis as the 'main condition' and code accordingly.

Example 29:	Main condition:	Febrile convulsions
	Other conditions:	Anaemia
	Procedure:	No information about therapy

Code febrile convulsions as the ‘main condition’. Rule MB2 does not apply since the ‘main condition’ as reported is not a presenting symptom of the other reported condition.

Rule MB3. Signs and symptoms Where a symptom or sign is recorded as the ‘main condition’ with documentation that it may be due to either one condition or another, select the symptom as the ‘main condition’. Where two or more conditions are recorded as diagnostic options for the ‘main condition’, select the first condition recorded.

See also 'Coding of multiple conditions and coding of combination categories.'

Example 30:	Main condition:	Headache due to either stress and tension or acute sinusitis
	Other conditions:

Select headache as the ‘main condition’ and code accordingly.

Unlikely main condition

Where a condition, or an incidental problem, is recorded as the ‘main condition’, and a more life threatening or contagious condition, relevant to the treatment given and/or the specialty that cared for the patient, is recorded as an ‘other condition’, query the provider and if the reason for admission was reported incorrectly, you will need to reselect the latter as the ‘main condition’.

| Example 31: |Main condition: | Acute sinusitis | |----------|--------| | | Other conditions: | Carcinoma of endocervix| | | | Hypertension | | Patient in hospital for three weeks | | | Procedure: | Total hysterectomy| | | Admission: | Acute sinusitis, carcinoma of cervix accidentally discovered during stay | | | Admission specialty: | Ear, nose and throat | | | Treating specialty: | Gynaecology |

Keep Acute sinusitis as the ‘main condition’. The case has been reported correctly, in view of timing of diagnosis. Even though the carcinoma of the cervix is a more life threatening condition and was treated during this episode, the main condition remains as acute sinusitis as that was the reason for admission. The notes indicate the carcinoma of the cervix was discovered accidentally during the stay.

| Example 32: |Main condition: |Rheumatoid arthritis| | ------ |----- | | | Other conditions: | Diabetes mellitus| | | | Strangulated femoral hernia | | | | Generalized arteriosclerosis | | Patient in hospital for two weeks. | | | Procedure: | Herniorrhaphy| | | Specialty: | Surgery – no other specialty involved. |

Clarify if the reason for admission is definitely the rheumatoid arthritis, and if so, code rheumatoid arthritis as the main condition and cluster code the strangulated femoral hernia with a code from the extension code chapter to identify it as a condition arising during the stay. However, in this example, if the likely reason for admission was the hernia, then reselect strangulated femoral hernia as the ‘main condition’.

Example 33:	Main condition:	Epilepsy
	Other conditions:	Otomycosis
	Specialty:	Ear, nose and throat
		Specialty:

Reselect otomycosis as the ‘main condition’. The case has been reported incorrectly, in view of the treating specialty and the documented procedure.

Example 34:	Main condition:	Congestive heart failure
	Other conditions:	Fracture neck of femur due to fall during hospitalization
	Patient in hospital for four weeks
	Procedure:	Internal fixation of fracture Internal medicine for 1 week then
	Specialty:	transfer to orthopaedic surgery for treatment of fracture

Keep the congestive heart failure as the main condition (reason for admission) and cluster code the fractured neck of femur with a code from the extension code chapter to identify it as a condition arising during the stay. Additional coding may be necessary to identify all dimensions of this case for the purposes of patient safety analysis.

Example 35:	Main condition:	Cerebrovascular accident
	Other conditions:	Diabetes mellitus
		Hypertension
		Cerebral haemorrhage

Reselect cerebral haemorrhage as the ‘main condition’.

Example 36:	Main condition:	Congenital heart disease
	Other conditions:	Ventricular septal defect

Reselect ventricular septal defect as the ‘main condition’ and add to the cluster a code for congenital heart disease.

Example 37:	Main condition:	Enteritis
	Other conditions:	Crohn disease of ileum

Reselect Crohn disease of ileum as the ‘main condition’.

Example 38:	Main condition:	Dystocia
	Other conditions:	Hydrocephalic foetus
		Fetal distress
	Procedure:	Caesarean section

Reselect obstructed labour due to other abnormalities of foetus.

Chapter-specific notes

Guidance is given below for specific chapters where problems may be encountered in selecting preferred ‘main condition’ codes. The preceding general guidelines and rules apply to all chapters unless a specific chapter note states otherwise.

Chapter 1: Infectious and parasitic diseases

Human immunodeficiency virus [HIV] disease

A patient with a compromised immune system due to HIV disease may sometimes require treatment during the same episode of care for more than one disease, for example mycobacterial and cytomegalovirus infections. Only subcategories for stage, tuberculosis and malaria are provided in this block for HIV disease. Code the appropriate subcategory as selected by the health-care practitioner, and add to the cluster a code for the HIV caused disease.

Example 1:	Main condition:	HIV disease and Kaposi sarcoma
	Other conditions:	-

Code to HIV disease, stage unspecified and add to the cluster a code for the Kaposi sarcoma.

Example 2:	Main condition:	Toxoplasmosis and cryptococcosis in HIV patient
	Other conditions:	-

Code to HIV disease, stage unspecified and add to the cluster the codes for toxoplasmosis and cryptococcosis.

Example 3:	Main condition:	HIV disease Stage 4 with Pneumocystis carinii pneumonia, Burkitt lymphoma and oral candidiasis
	Other conditions:	-

Code to HIV disease clinical stage 4. Additional codes for Pneumocystis carinii pneumonia, Burkitt lymphoma and oral candidiasis may be added to the cluster, if desired.

Chapter 2: Neoplasms

When coding neoplasms, refer to the instructions at the level of the individual categories regarding code assignment and the use of additional morphological or site descriptions from the extension codes.

A neoplasm, whether primary or metastatic, that is the focus of care during a relevant episode of health care, should be recorded and coded first in the cluster. When the ‘main condition’ as recorded by the health-care practitioner is a primary neoplasm that is no longer present (having been removed during a previous episode of care), code first the neoplasm of the secondary site, the current complication, or the appropriate circumstance codable to Chapter 24, Factors influencing health status or contact with health services for reasons other than illness that was the focus of the treatment or investigation during the current episode of care, add an appropriate code from Chapter 24 for personal history of neoplasm and the code for the removed primary neoplasm at the end of the cluster..

Example 1:	Main condition:	Carcinoma of prostate
	Other conditions:	Chronic bronchitis
	Procedure:	Prostatectomy

Code to malignant neoplasm of prostate as the ‘main condition’. Code Chronic bronchitis in a separate cluster.

Example 2:	Main condition:	Carcinoma of breast resected two years ago
	Other conditions:	Secondary carcinoma in lung
	Procedure:	Bronchoscopy with biopsy

Code to Malignant neoplasm metastasis in lung and add to the same cluster the code for Personal history of malignant neoplasm of breast in conjunction with the code for Carcinoma of the breast.

Example 3:	Main condition:	Previously excised bladder cancer – admitted for follow-up examination by cystoscopy
	Other conditions:	-
	Procedure:	Cystoscopy

Code to Follow-up examination after surgery for malignant neoplasm. The code for Personal history of malignant neoplasm of urinary tract is added to the cluster as an additional code and a code for the Bladder cancer.

Metastatic malignant neoplasms, except of lymphoid, haematopoietic, central nervous system or related tissues, unspecified

The code should be used alone for coding only when the malignancy is described as 'disseminated metastases' or 'metastatic carcinoma' (or other similar terms as described in the inclusion list of the code) and the specific sites are not documented.

*Unspecified malignant neoplasms of ill-defined or unspecified sites *

This code should be used only when the health care practitioner has clearly recorded the neoplasm as an unknown primary site or as an unspecified malignancy, assumed primary.

Malignant neoplasms of independent, primary multiple sites

Malignant neoplasms of independent (primary) multiple sites should be used when the health-care practitioner records as the main condition two or more independent primary malignant neoplasms, none of which predominates. Additional codes to identify the individual neoplasms are added to the cluster to identify the individual malignant neoplasms listed. Codes from the extension codes may be added to the cluster to identify additional detail of the histopathology and the site.

Example 4:	Main condition:	Carcinomatosis
	Other conditions:	-

Code to Metastatic malignant neoplasms, except of lymphoid, haemoatopoietic, central nervous system or related tissues, unspecified and add a code for Unspecified malignant neoplasms of ill-defined or unspecified sites in the same cluster. An appropriate code from Chapter 24 for personal history of neoplasm should be used for a primary neoplasm that is no longer present.

Example 5:	Main condition:	Multiple myeloma and primary adenocarcinoma of prostate

Code to Malignant neoplasms of independent, primary multiple sites and add the codes for Plasma cell myeloma and Malignant neoplasm of prostate. A code from the extension codes to identify the multiple myeloma may be added as optional additional code to the cluster in conjunction with the Plasma cell myeloma.

Chapter 3: Diseases of the blood or blood-forming organs

Certain conditions classifiable to this chapter may result from drugs or other external causes. Codes from Chapter 20 may be used as optional additional codes.

|Example 1: |Main condition:| Trimethoprim-induced folate deficiency anaemia| |-------------|----------| ||Other conditions:| - |

Code Drug-induced folate deficiency anaemia and add a code for ‘Drugs medicaments and biological substances associated with injury or harm in therapeutic use, Other systemic anti-infectives and antiparasitics, Antimalarials and drugs acting on other blood protozoa’ in the same cluster together with a code for ‘Drugs medicaments and biological substances associated with injury or harm in therapeutic use, Other systemic anti-infectives and antiparasitics’. A code to identify trimethoprim may be used as an optional additional code from the extension codes in the same cluster.

Chapter 5: Endocrine, nutritional or metabolic diseases

Certain conditions classifiable to this chapter may result from drugs or other external causes. Codes from Chapter 23 may be used as optional additional codes.

Diabetes mellitus

In coding the selection of an appropriate additional category that describes the complication should be based on the ‘main condition’ as recorded by the health-care practitioner. Code all reported complications individually in the same cluster.

Example 1:	Main condition:	Kidney failure due to diabetic glomerulonephrosis
	Other conditions:	-

Code to Unspecified diabetes mellitus and add a code for the Kidney failure, unspecified in the same cluster.

Example 2:	Main condition:	Type 1 diabetic with nephropathy, gangrene and diabetic cataract
	Other conditions:

Code to Type 1 diabetes mellitus and add codes for the Kidney failure unspecified, Gangrene and Diabetic cataract in the same cluster.

Carcinoid syndrome

This code is not to be used alone if a carcinoid tumour is recorded, unless the episode of care was directed predominantly at the endocrine syndrome itself and the tumor is not reported.

Chapter 6: Mental, behavioural or neurodevelopmental disorders

In some categories there is provision for optional additional codes.

Chapter 8: Diseases of the nervous system

Certain conditions classifiable to this chapter may result from the effects of drugs or other external causes. Codes from Chapter 23 may be used as optional additional codes.

Late effect of cerebrovascular disease

These codes are not to be used as the preferred code for the ‘main condition’ if the nature of the residual condition is recorded.

Paralytic symptoms

These codes are not to be used alone if a current cause is recorded, unless the episode of care was mainly for the paralysis itself and the cause is not recorded.

Example 1:	Main condition:	Cerebrovascular accident with hemiplegia
	Other conditions:
	Specialty:	Neurology

Code Stroke not known if ischaemic or haemorrhagic and add the code for (Hemiplegia, unspecified) may be used as an optional additional code.

Example 2:	Main condition:	Cerebral infarction three years ago
	Other conditions:	Paralysis of left leg
		Patient receiving physical therapy

Code Monoplegia of lower extremity, unspecified as ‘main condition’. Late effect of cerebral ischaemic stroke may be used as an optional additional code.

Chapter 9: Diseases of the visual system

Visual impairment) These codes are not to be used alone if the cause is recorded, unless the episode of care was mainly for the blindness itself and the cause of blindness is not recorded.

Chapter 10: Diseases of the ear or mastoid process

Acquired hearing impairment

These codes are not to be used alone if the cause is recorded, unless the episode of care was mainly for the hearing loss itself and the cause was not recorded.

Chapter 11: Diseases of the circulatory system

Secondary hypertension

This code is not to be used alone if the cause is recorded. When coding to the cause, secondary hypertension is used as additional code to indicate that this manifestation has been relevant in the context of a treatment.

Chapter 15: Diseases of the musculoskeletal system or connective tissue

Many musculoskeletal conditions are treated without knowing the underlying disease. In such cases only the musculoskeletal condition is coded. Wherever the underlying condition is known, it is coded at the second place in the coding cluster.

Chapter 18: Pregnancy, childbirth or the puerperium

Complications following abortion and ectopic and molar pregnancy

These codes are not to be used alone, except where a new episode of care is solely for treatment of a complication, e.g. a current complication of a previous abortion. It may be used as an optional additional code to identify associated complications and to give fuller details of the complication.

Example 1:	Main condition:	Ruptured tubal pregnancy with shock
	Specialty:	Gynaecology

Code ruptured tubal pregnancy and in the same cluster the shock following abortion and ectopic and molar pregnancy.

Example 2:	Main condition:	Incomplete abortion with perforation of uterus
	Specialty:	Gynaecology

Code Unspecified abortion with Other specified complications following abortion, ectopic or molar pregnancy and add in the same cluster codes indicating complication of medical care and the resulting injury, the perforation of the uterus.

Example 3:	Main condition:	Disseminated intravascular coagulation following abortion performed two days ago at another facility
	Specialty:	Gynaecology

Code Delayed or excessive haemorrhage following abortion, ectopic ord molar pregnancy. No other code is required since the abortion was performed during a previous episode of care.

Delivery

Use of these codes to describe the ‘main condition’ should be limited to cases where the only information recorded is a statement of delivery or the method of delivery. These codes may be used as additional codes to indicate a method or type of delivery where no separate data item or procedural classification is being used for this purpose.

Example 4:	Main condition:	Pregnancy
	Other conditions:
	Procedure:	Low forceps delivery

Code Single delivery by forceps or vacuum extraction as ‘main condition’ since no other information is provided.

Example 5:	Main condition:	Pregnancy delivered
	Other conditions:	Failed trial of labour
	Procedure:	Caesarean section

Code Failed trial of labour, unspecified as the ‘main condition’. The code for Single delivery by caesarean section, unspecified, is used as an additional code in the same cluster.

Example 6:	Main condition:	Twin pregnancy delivered
	Other conditions:
	Procedure:	Spontaneous delivery

Code Twin pregnancy as the ‘main condition’. Multiple delivery, all spontaneous may be added as an optional additional code.

Example 7:	Main condition:	Term pregnancy delivered of dead foetus, 2800 g
	Other conditions:	-
	Procedure:	Spontaneous delivery

For the mother, code to Maternal care for intrauterine death if no specific reason for the fetal death can be determined.

Certain maternal diseases classifiable elsewhere but complicating pregnancy, childbirth or the puerperium

The subcategories provided should be used as ‘main condition’ codes in preference to categories outside Chapter 18 when the conditions being classified have been indicated by the health-care practitioner to have complicated the pregnant state, to have been aggravated by the pregnancy, or to have been the reason for obstetric care. The pertinent codes from other chapters may be used as optional additional codes to allow specification of the condition.

Example 8:	Main condition:	Toxoplasmosis
	Other conditions:	Pregnancy undelivered
	Specialty:	High-risk antenatal clinic

Code Protozoal diseases complicating pregnancy, childbirth or the puerperium as the main condition. A code for Toxoplasmosis, unspecified is used as an additional code in the cluster to identify the specific organism.

Chapter 21: Symptoms, signs or clinical findings, not elsewhere classified

Categories from this chapter should not be used as ‘main condition’ codes unless the symptom, sign or clinical finding was clearly the main condition treated or investigated during an episode of care and was unrelated to other conditions recorded by the health-care practitioner. See also Rule MB3 and the introduction to Chapter 21 in Volume 1 for further information.

Chapter 22: Injury, poisoning or certain other consequences of external causes

Where multiple injuries are recorded and no one of these has been selected as the ‘main condition’, code to one of the categories provided for statements of multiple injuries of:

• same type to the same body region;
• different types to the same body region; and
• same type to different body regions

and code in the same cluster, after the code for ‘multiple’ the individual injuries separately

Note the following special cases:

• for internal injuries recorded with superficial injuries and/or open wounds only, code to internal injuries first in the cluster that has all the injuries;
• for fractures of skull and facial bones with associated intracranial injury, code to the intracranial injury first in the cluster that has all the injuries;
• for intracranial haemorrhage recorded with other injuries to the head only, code to intracranial haemorrhage first in the cluster that has all the injuries; and
• for fractures recorded with open wounds of the same location only, code to fracture first in the cluster that has all the injuries.

When the multiple injury categories are used, codes for any individual injuries listed are used as additional codes in the same cluster.

Example 1:		Injury of bladder and urethra
	Other conditions:	-

Code to Injury of multiple pelvic organs and add codes for Injury of bladder and Injury of urethra in the same cluster.

Example 2:		Open intracranial wound with cerebellar haemorrhage
	Other conditions:	-

Code to Traumatic intracerebellar haemorrhage and add the code for Open wound of head, unspecified

Categories are provided in Chapter 22 for Injury or harm arising from surgical or medical care, not elsewhere classified, e.g. surgical wound infections, complications of implanted devices, shock, etc. Most body-system chapters also contain categories for conditions that occur either as a consequence of specific procedures and techniques or as a result of the removal of an organ, e.g. postmastectomy lymphoedema syndrome, post-irradiation hypothyroidism. Some conditions (e.g. pneumonia, pulmonary embolism) that may arise in the postprocedural period are not considered unique post procedural entities and, therefore, are not found with the codes for postprocedural disorders. An optional additional code from the block Causes of healthcare related harm or injury may be added to identify the relationship to a procedure.

Chapter 23: External causes of morbidity or mortality

These codes are not to be used as ‘main condition’ codes. They are intended for use as optional additional codes to identify the external cause of conditions classified in Chapter 22, and may also be used as optional additional codes with conditions classified in any other chapter but having an external cause.

Special cases

The morbidity special tabulation list is intended as a basis for national lists and for inter-country comparison. National lists can be constructed by either condensing or expanding the core classification as appropriate. The list is suitable for data on inpatient care and, with suitable adaptation – notable aggregation of some items and expansion of items relating to Chapter 21 (Symptoms, signs or clinical findings, not elsewhere classified) and Chapter 24 (Factors influencing health status and contact with health services) -- for information from other sources, such as ambulatory care and surveys. When a local list is constructed, the key to the condensed categories should contain the three (or four) character codes of the core classification. The list has been designed for international comparisons of hospital morbidity statistics. This concise list allows for comparison of hospital activity, independent of health systems, and based on the version of the ICD in use. The conditions have been selected in a way that they can always be treated in an admission of at least 24 hours. If, after examination of the frequencies of the ICD four-character rubrics, it is necessary to expand the list, some of the items within ICD categories can be subdivided according to the core classification or even to the five-character level. If the recommended list is too detailed or if a shorter list is required, selection can be made based on national or local health concerns. Depending on a country's ‘epidemiological profile’, categories may be combined to shorten the list.

Morbidity for clinical care

Clinical care comprises different levels of treatment, all of which mean a level of diagnostic capacity that is higher than in primary care. The ICD addresses this level of detail primarily through multidimensional coding. Secondary care refers to the health care services provided by medical specialists and other health professionals who generally do not have first contact with patients, for example, cardiologists, urologists, or dermatologists. It includes acute care, necessary treatment for a short period of time for a brief but serious illness, injury or other health condition, such as in a hospital emergency department. It also includes skilled attendants during childbirth, intensive care, and medical imaging services. ‘Secondary care’ is sometimes used synonymously with ‘hospital care’. However, many secondary care providers do not necessarily work in hospitals, such as psychiatrists, clinical psychologists, or physiotherapists, and some primary care services are delivered within hospitals. Depending on the organization and policies of the national health system, patients may be required to see a primary care provider for a referral before they can access secondary care. Tertiary care refers to specialized consultative health care, usually for inpatients and following a referral from a primary or secondary health professional, in a facility that has personnel and facilities for advanced medical investigation and treatment, such as a tertiary referral hospital.

Morbidity for epidemiology

Epidemiology is the study of the distribution and determinants of health-related states or events (including disease) and the application of this study to the control of diseases and other health problems. Various methods can be used to carry out epidemiological investigations: surveillance and descriptive studies are used to study distribution and analytical studies are used to study determinants. ICD coded data, either from morbidity and mortality sources, contribute to the understanding of the health of a population.

Morbidity for quality and patient safety

Coded health information is used to measure and report on various aspects of quality of care and patient safety (e.g. reporting on in-hospital mortality or adverse event rates for various conditions, or reporting on patient safety indicators). Users of this kind of health information are health system payers (e.g. ministries of health, or in privately-funded health care systems, health insurance companies) and other stakeholders, such as health quality councils, hospital administrators, clinical leaders/groups, or public advocacy organizations.

The quality and safety use case for ICD–11

The quality and safety use case of the ICD is based on the availability of large numbers of methodological tools that are originally based on ICD–10. Specific examples include the Charlson and Elixhauser co-morbidity indices, AHRQ Patient Safety (Agency for Healthcare Research and Quality) Indicators, the Hospital Standardized Mortality Ratio, and various other administrative data quality indicators. Most of these tools were actually first developed as ICD-9 or ICD-9-CM coding algorithm methodologies, and have only recently been translated, through rigorous research, to ICD–10. ICD–11 development involved restructuring to coding logic or structure (relative to ICD–10), and pre-existing tools for quality or patient safety reporting need to undergo similar translation work so that they can be applied in ICD–11. WHO recommendations on coding rules for hospital separation episodes improve comparability of records across hospitals and jurisdictions. Specific examples of coding rules include: a) rules for specifying the most responsible diagnosis, b) numbers of codes per record, c) code clustering mechanisms, and d) use of a status display system that distinguishes diagnoses arising during a hospital stay from those present at admission. Quality and patient safety reporting is often focused not only on diagnostic information available in the International Classification of Diseases, but also on procedure information, that is currently coded in various country-specific procedure coding systems. The harmonization of ontological concepts in international procedure coding systems will be important going forward. The available medical complication codes of ICD–11 are in line with current knowledge in the domain of safety and adverse events.

Reporting on indicators of quality of care and patient safety

This use case relates to the use of coded health information to measure and report on various aspects of quality of care and patient safety. (e.g., reporting on in-hospital mortality or adverse event rates for various conditions, or reporting on patient safety indicators). The initiating actor may be a health quality council, hospital administrators, clinical leaders/groups, a health system payer (e.g., ministries of health, or in privately-funded health care systems, health insurance companies) or a public advocacy ‘watch-dog’ organizations. The participating actors are hospital administrators, clinicians, health system decision makers, public representatives, patients and their families, and sometimes even the media. Preconditions are: - Availability of person-level data on episodes of health care delivery (e.g., hospitalizations, physician visits) - Identifiers that permit attribution of the health care delivery episode to a provider, provider group, or a given health facility/hospital. - Clinical information on diagnoses present and procedures performed during a health care delivery episode. - Clinical information on relevant outcomes such as mortality, length of stay, and specific adverse events. - Analytical expertise among initiating actors so that attention is paid to data validity considerations, knowledge of ‘best’ indicators (e.g., the most valid patient safety indicators), risk adjustment methodology, etc. The outputs are reports containing information on dimensions of system quality. These can either provide global information on system performance, or comparative information stratified by provider unit (e.g. physician-level, hospital-level, or regional reporting).

1 Functionality:

An ideal course of events for such use would include:

• The initiating actor communicates desire to conduct quality/safety measurement and reporting to relevant stakeholders.
• Appropriate applications are made to secure access to the data needed to conduct the planned reporting.
• Appropriate methodological and clinical expertise is enlisted to ensure that best methodological practices are incorporated into the planned reporting, and that clinical face validity and acceptability are considered.
• Data analysis and reporting are undertaken.
• Broad dissemination and knowledge translation to stakeholders is undertaken. A continuous quality improvement process is undertaken in response to reports (with consideration given to quality improvement interventions, and repeat measurement after intervention). Exceptions: Quality/safety reporting that does not follow the sequence of steps described above can be compromised. Indeed, there are many historical instances of failed or suboptimal quality/safety reporting from administrative data because of skipped steps. (e.g., 1. quality reporting without valid indicators, or appropriate methodologies for risk adjustment, 2. quality reporting without good clinical face validity, 3. quality reporting without a Continuous Quality Improvement(CQI) mindset, etc.). Examples of sub-use cases (addressing the quality dimensions of effectiveness, efficiency, safety, access)
• reporting on global mortality by facility (e.g., the hospital standardized mortality ratio - HSMR)
• reporting on condition-specific mortality
• reporting on patient safety indicators
• reporting on global or condition-specific length of stay
• reporting on readmission rates after hospitalization
• reporting on global or condition-specific costs of care (e.g., cost per hospitalization)
• reporting on waiting times
• reporting on small area variability in utilization

2 Additional information:

Requirements: See ‘Preconditions’ section above. There needs to be ongoing development and refinement of quality reporting methodologies (in essence, ongoing research around the development of new administrative data quality indicators and new methodologies for risk adjustment in outcome/quality reporting). Assumptions: An underlying assumption in quality or patient safety reporting from administrative data is uniformity of data format and data validity across comparator units (i.e., across provides, hospitals, or jurisdictions). Uniformity in data format and validity is not always present, and has been a common reason for criticism of quality or patient safety reports derived from administrative data. In this regard, all ongoing WHO efforts towards achieving directive coding rules help to facilitate comparative reporting by reducing data variability across comparator units (e.g., rules on factors such as the definition of the ‘most responsible diagnosis’, numbers of possible codes per record, and the implementation of diagnosis-timing codes). See also the separate use case description for international comparative reporting.

Conceptual model for quality and patient safety

Exposure to health care events sometimes has unintended and undesired consequences. Health care, the people to whom it is provided and the complications that can arise in the course of care are highly diverse and complex. Representing them comprehensively in an information system is challenging, and is presently beyond the bounds of practicality for routine administrative information systems of the types that are intended to make use of the ICD. The conceptual model has three components: -harm to the patient: With what main consequence for the patient’s health? -cause or source of harm: What caused the harm? -mode or mechanism: In what way? How did the source of harm actually produce harm?

Overview of code-set in ICD–11 for quality and patient safety

A key feature of the Quality and patient safety code-set in ICD–11 is that a cluster of codes is required to represent a case. Provision for cluster-coding has been an aspect of the ICD since the 6th Revision, which introduced separate chapters for coding injuries and their external causes. It has also been prominent in ICD-9 and ICD–10 clinical modifications, often as a requirement to ‘code also’ a second concept to provide additional relevant information on a case. Use of the term ‘cluster’ is novel in ICD–11 and so is the extent and formalisation of requirement for cluster-coding. The cluster required to code a Quality and safety case has three codes, one for each of the three components of the model given above. The first component, Quality and patient safety Harm, is usually represented by a standard ICD–11 diagnosis code, from [almost] any chapter of the classification. Some forms of harm that can result from Quality and safety events are not adequately represented by a standard ICD–11 diagnosis code. A special set of categories to represent these forms of harm are provided in the injury chapter of ICD–11, under the heading Complications of Surgical and Medical Care, not elsewhere classified. Quality and patient Safety causes and sources of harm fall into four types of causes at the top level that capture events caused by:

1. substances (drugs and medicaments, etc.),
2. procedures,
3. devices, and
4. a mixed bag of other types of causes (e.g. problems associated with transfusions, or problems associated with diagnosis, including missed diagnosis, incorrect diagnosis, etc.).

The full Quality and safety cause code-lists are part of the Chapter 23, External causes.

Quality and safety Mode or Mechanism (‘Mode’ is used here) refers to the main way in which the Quality and safety Cause leads to the harm represented in the third concept, Quality and safety Harm. Quality and safety Modes are specific to the types of Quality and safety Cause. Examples are:

	Examples of corresponding Quality and safety Mode or Mechanism
Cause or Source of Harm	Mode or Mechanism
Substance	Overdose, under-dose, wrong substance.
Procedure	Accidental perforation of an organ during a procedure.
Device	Dislodgement. Malfunction.
Other cause	Mismatched blood. Patient dropped during transfer from OR table.

The third component, Quality and safety Harm, is usually represented by a standard ICD–11 diagnosis code, from [almost] any chapter of the classification. Some forms of harm that can result from Quality and safety events are not adequately represented by a standard ICD–11 diagnosis code. A special set of categories to represent these forms of harm are provided in the injury chapter of ICD–11, under the heading Complications of Surgical and Medical Care, not elsewhere classified.

Examples for the ICD–11 Quality and safety coding model

The ICD–11 Quality and safety coding model is demonstrated by the examples in the following table.

|| |---------|--------------------|------------------------------|------------------|---------| |Example 1 |Case | A woman has been admitted to hospital for stabilization of diabetes. She is erroneously prescribed three times the usual dose of a medication. The abnormally high dose is given and the patient has a hypoglycaemic episode. ||| | | Harm | Diabetes mellitus with hypoglycaemia ||| | | Mode or Mechanism| Overdose of substance ||| | | Cause | Exposure to a drug, medicament or biological substance - Insulin and oral hypoglycaemic [anti-diabetic] drugs ||| | Example 2 | Case | A man attends a primary care physician for removal of a skin lump, mainly to exclude the possibility of malignancy. The lesion is excised and the wound is sutured. It later becomes known that the physician had Hepatitis C, and the patient has now contracted this disease. ||| | | Harm | Acute hepatitis C ||| | |Mode or Mechanism | Failure of sterile precautions ||| | | Cause | Procedures associated with injury or harm therapeutic use - Biopsy procedure ||| | Example 3 | Case | An elderly woman is admitted due to a fractured neck of femur. Surgical fixation is undertaken. The operative site bleeds heavily the day after surgery, requiring return to theatre. ||| | | Harm | Haemorrhage | | | | | Mode or Mechanism | Mode of injury or harm associated with surgical procedure - unspecified ||| | | Cause | Medical or surgical procedure - Orthopaedic surgical procedure ||| | Example 4 |Case |A 63 year old man had a knee-replacement less than a year ago, because of arthritis. The implanted device has come loose, resulting in pain and reduced function.| | | Harm|Loosening of internal joint prosthesis | | | Mode or Mechanism| Device failure | | | Cause|Orthopaedic devices associated with adverse incidents prosthetic and other implants, materials and accessory devices | | Example 5 | Case|A man has bowel cancer. Abdominal surgery was done several days ago to resect the affected part of the colon and re-join the preserved part of the colon. The anastomosis has leaked, and required surgical revision.| | | Harm|Leaking anastomosis | | | Mode or Mechanism| Other specified mode. [Because none of the more specific types appears to provide for failure of anastomosis.]| | | Cause|Medical or surgical procedure - Gastrointestinal surgical procedure |

Note that in each of these examples, a mode/mechanism of harm is coded alongside the cause of harm code for all cases. This is true, even when a mode of harm is not apparent. In the latter situations, a code for ‘mode or mechanism of injury unspecified’ should be selected, for any of substance-related harm, procedure-related harm, or device-related harm. The ‘other aspects of care’ category of causes of healthcare related harm is the exception to this, where there is typically only a need to code the ‘other aspects of care codes’ from Chapter 23 along with the actual harm or injury code from anywhere in the classification.

Recommendations for data capture and organization

Information systems must be capable of: - capturing the three components - marking the three codes as belonging to the same cluster (see also instructions for cluster coding)

Recommendations for use and interpretation of coded data

These recommendations apply to the use of records in which data were captured and organised as recommended in the previous section. - Select records involving a quality or patient safety event: these are all records with any quality or patient safety harm code. - Summarise types of quality or patient safety harm represented in a set of records: select records with any quality or patient safety harm code. Summarise the distribution of quality or patient safety Harm codes present in the selected set. - Summarise quality or patient safety causes of harm in a set of records. - Summarise quality or patient safety mechanisms in a set of records. - Summarise quality or patient safety harm in a set of records.

Morbidity for research purposes

The morbidity use case for ICD–11 includes a number of situations where the primary goal is to work in an academic research paradigm to extract information from ICD–11 coded data to study burden of disease, clusters of disease, geographic distribution of diseases, and health impacts associated with various diseases. The research paradigm is of course most relevant when it has translational relevance to either health system policy or public health policy, in which case the research paradigm, labelled as such, becomes indistinguishable from applied morbidity analyses conducted for the purposed of health planning. Nevertheless, explicit mention is made here of the widespread use of ICD–11 coded data in a research paradigm, recognizing that this is one of the significant drivers for developing a clinically rich and detailed classification system, with novel features and coding rules that enhance the classification's potential as a research tool.

Morbidity in primary care

Primary care has been defined as essential front line health care based on practical, scientifically sound, and socially acceptable methods and technology made universally accessible to individuals and families in the community through their full participation and at a cost that the community and the country can afford to maintain. Of relevance to primary care, ICD–11 includes many diagnostic and disease entities that are common reasons for searching contact with the health system at the first level of health services. ICD–11 has various primary-care tabular lists depending on the resource level. A primary care tabular list for low resource-settings enables simple reporting of broader concepts. A high and middle resource-setting tabular list is used when more sophisticated diagnoses and treatments are available. An International Classification of Disease for Primary Care (ICD-PCI) has been developed by the World Organization of National Colleges, Academies and Academic Associations of General Practitioners/Family Physicians (WONCA), through its WONCA International Classification Committee (WICC). WONCA and the WHO have collaborated in the updates of their classifications, and for ICD–11 have collaborated in filling previous gaps in the ICD for primary care use. As a result, new versions of WONCA's International Classification of Primary Care and ICD–11 share a common subset of categories. Both classifications are now fully derived classifications of the ICD and allow cross sectoral comparability, such as between primary care and hospital activity.

The ICD–11 has versions shaped for different primary care settings:

• Low resources setting - a simplified version.
• Middle resource setting.
• High resource setting, the tabular list for mortality and morbidity statistics contains elements relevant to primary care and is thus able to be used in high resource environments for primary care, as well as for secondary and tertiary care. The middle resource version shares all concepts with ICPC, which allows the ICD or ICPC to be used interchangeably for primary care medicine settings.

Casemix groupings

In casemix grouping systems such as the Diagnosis Related Group (DRG) system, ICD based data are used for reimbursement or resource allocation. Such systems are used in systematic fashion (nationwide) in over 22 countries for reimbursement or resource allocation. The assignment of patient cases to groups is based on an algorithm using, in addition to coded diagnosis information, coded procedures and a number of other variables. The scientific basis of the casemix systems is grounded in healthcare economics and in the theory of medicine. Since casemix systems are an essential part of administration in countries that use them, smooth transfer to the new revision of the ICD in these systems is essential for the approval and implementation of the new revision. ICD–11 has been developed to accommodate the different levels of detail that are required in diagnosis-related casemix groupings, in close collaboration with the custodians of the diverse casemix systems. Joint use in a specific casemix system is driven by the relevant grouper algorithms, and partly also by national legislation. For matters of international comparability of hospital activity, it is recommended that countries adopt the new WHO definition of main diagnosis and that country implementations of ICD–11 apply the new extension codes for the type of diagnosis that are provided with ICD–11. For international tabulations, the resulting diagnoses are listed with the aid of the International Shortlist for Hospital Morbidity Tabulation.

Use of functioning properties

Functioning embedded in ICD serves to allow a first documentation of functioning of an individual. Using selected subsets for the assessment of functioning has proven a useful method that is also used in the WHO Disability Assessment Scale (WHODAS). Where possible, the full ICF should be used for a complete reporting of functioning.

Description of functioning properties

FPs refer to aspects of functioning that are potentially related to a particular health condition, e.g. fine hand use is likely to be present in people with rheumatoid arthritis^1. In the ICD, FPs serve as an alignment of disease entities with these aspects as represented in the ICF (Selb et al., 2015). More specifically, FPs are predefined ICF categories from the activities and participation (A&P) component of the ICF likely to be most relevant for a given health condition. This does not preclude that a person may also experience problems in other aspects of functioning. In the ICD, these ICF categories are grouped into the following domains:

Domains	Description of domain
Understanding Impact	Cognitive, mental or psychological processes involved in comprehending, thinking and making interpretations.
Communication Impact	Verbal or non-verbal processes including transmission or sharing of information between two or more people.
Life Management Activities Impact	Activities involving everyday actions, tasks and routines in different contexts such as home, school and work place
Mobility Impact	An ability to move easily and freely by changing body position or from one location to another
Self-Care Impact	Activities necessary to care for oneself to maintain health and hygiene
Household Activities Impact	Activities involving everyday domestically-related activities such as cleaning the living area, washing and drying clothes, or shopping
Interpersonal Relations Impact	Interactions between people (e.g. family members, friends, relatives, strangers) in socially appropriate manners.
School Activities Impact	Activities related to and involvement in school activities
Work Activities Impact	Activities related to and involvement in work life
Social Participation Impact	Engagement and involvement in social and political life
Children and Youth Impact	Aspects of functioning that consider specific developmental characteristics of children and youth

The inclusion of functioning domains in the ICD-11 Content Model through FPs are designed to assist health professionals and professionals in related fields - to comprehensively describe aspects of functioning that are related to a person’s health condition. - in their daily decision-making, e.g. making a diagnosis, in treatment planning, monitoring outcomes and treatment response over time and along the continuum of care. - in examining eligibility for disability pension and other social welfare benefits or specific health services and programs. The use of FPs by stakeholders in the health system who are responsible for program planning and resource allocation is intended to enable the systematic collection of information about the health and health-related problems and needs of persons living with a specific health condition. This data could be used to - plan programs and interventions - inform resource allocation - provide data for risk adjustment to inform optimal care planning. For all stakeholders in the health system, including researchers, functioning is a relevant outcome variable in the comparison of different treatment modalities. Noteworthy, FPs point to the use of ICF in the context of ICD. It is only in this context that a given health condition triggers the opening up of a preselected set of ICF categories to be coded within the context of ICD. The use of ICF in this context does not override the conceptual model underpinning the ICF which recognizes the multiple interactions among the domains of functioning and health.

Implementation of functioning properties

In the ICD-11, there are three predefined options available for specifying FPs for a given health condition. These specifications define the aspects of functioning to be coded. More specifically, the problems of a person with respect to these aspects at the given time point are coded. The setting will need to specify beforehand which option(s) should be available in the respective setting; only these options would then open up once the user enters a health condition.

Option 1 is the default for any health condition. Option 2 applies for 100 rehabilitation-relevant health conditions, such as rheumatoid arthritis, obesity, or depression. Option 3 constitutes an extension of Option 1 and 2.

Option1: Default for any health condition:

The default set for any health condition is a set of 4 ICF categories. These four ICF categories are a selection of a minimal generic set to best describe variations in functioning across people with various health conditions^2. The (G) after the category title always indicates that these are the 4 ICF categories from the A&P component derived from the minimal generic set, so-called ‘generic set’.

Life Management Activities Impact
d230	Carrying out daily routine (G)	Described as carrying out simple or complex and coordinated actions in order to plan, manage and complete the requirements of day-to-day procedures or duties, such as budgeting time and making plans for separate activities throughout the day.
Mobility Impact d450	Walking (G)	Described as moving along a surface on foot, step by step, so that one foot is always on the ground, such as when strolling, sauntering, walking forward, backwards, or sideways.
d455	Moving around (G)	Described as moving the whole body from one place to another by means other than walking, such as climbing over a rock or running down a street, skipping, scampering, jumping, somersaulting or running around obstacles.
Work Activities Impact
d850	Remunerative employment (G)	Described as engaging in all aspects of work, as an occupation, trade, profession or other form of employment, for payment, as an employee, full or part time, or self-employed, such as seeking employment and getting a job, doing the required tasks of the job, attending work on time as required, supervising other workers or being supervised, and performing required tasks alone or in groups.

Option 2: Further specification for rehabilitation-relevant health conditions

For 100 selected rehabilitation-relevant health conditions, e.g. rheumatoid arthritis, obesity, and depression there is a tailored set of FPs available in the ICD-11. If the user codes one of these 100 health conditions, a drop-down field will appear that allows the user to decide whether to continue with the generic FPs set as indicated in option 1 or to add to this generic set the FPs specifically tailored for the given health condition. Always including the four generic set categories in the FPs ensures that there is a minimal set of functioning information available at any point in time across the continuum of care. The full list of these 100 health conditions can be found in the Appendix.

Option 3: Further specification for any health condition in the ICD-11

To best describe varying levels of functioning across health conditions and along the continuum of care, ICD-11 users can decide to code an additional 13 ICF categories beyond the four generic set categories defined as the default in option 1^3. This more comprehensive set of 17 FPs, that includes the 4 FPs from option 1 (indicated with a (G) for generic set) can be found in the table below. This comprehensive set can be selected from the drop-down field that will appear when coding.

Life Management Activities Impact
d230	Carrying out daily routine (G)
d240	Handling stress and other psychological demands
Mobility Impact
d410	Changing basic body position
d415	Maintaining a body position
d420	Transferring oneself
d450	Walking (G)
d455	Moving around (G)
d465	Moving around using equipment
d470	Using transportation
Self-Care Impact
d510	Washing oneself
d520	Caring for body parts
d530	Toileting
d540	Dressing
d550	Eating
d570	Looking after one's health
Household Impact
d640	Doing housework
d660	Assisting others
Interpersonal Relations Impact
d710	Basic interpersonal interactions
d770	Intimate relationships
Work Activities Impact
d850	Remunerative employment (G)
Social Participation Impact
d920	Recreation and leisure

Coding functioning properties

Once a user selects an ICD code, the list of functioning properties opens up. Two coding rules are available for FPs. Each setting would have to specify beforehand which coding rule should be applied. Details about implementing FPs for morbidity are described in Section 3.2.5.

Binary rule for coding

The binary rule for coding implies that each function property (FP) is pre-coordinated with a ‘.0’, indicating with a zero that the person experiences no limitation in the given ICF category, or a ‘.8’ indicating that the person experiences limitations in the given ICF category. The codes of the FPs are reported in a separate data field.

ICF qualifier rule for coding

The ICF qualifier coding implies that the user is directed outside the ICD-11 to the ICF. The user would follow the coding instructions outlined in the ICF (WHO, 2001). A generic scale from 0 'no problem' to 4 'complete problem' is used in this context for coding the extent of functional impairments and restrictions. As with coding rule a) the coding of the FPs are documented in a distinct data field. Note: For certain purposes, such as predicting mortality, re-admission rates, length of stay, or monitoring disease activity and functioning over time, a single score on people’s functioning may be important. The creation of such a score would need to satisfy distinct attributes of measurement, such as uni-dimensionality and invariance across groups. More research is needed to establish such scores. It can be expected that in the future an impact score can be created by integrating information already collected in routine practice based on relevant psychometrical methods. Once such systems for score transformation are available, they can be integrated into the electronic format of the ICD-11.

Examples

In the following section, the coding instructions for FPs are illustrated with four case examples. For each example both coding rules, including the extension for the binary coding rule, are outlined. The cases apply different options for specifying FPs: Example 1 deploys option 1; Example 2 constitutes a rehabilitation-relevant health condition and draws upon option 2 for the specification of FPs; Example 3 refers to option 3, and Example 4 illustrates how FPs can be specified in the case of a multi-morbid patient.

Example 1: Influenza

Six days ago Mr. Mburu, 35 year old small shop owner, started experiencing extreme fatigue, 38 degree fever, severe headache and joint pain. These symptoms increasingly got worse as the days past, and in addition coughing, hot flashes and chills also appeared. Despite feeling very ill, he continued to work since his livelihood depends on keeping the shop open. However, at the end of the second day Mr. Mburu had major difficulties taking care of daily activities and moving around was extremely strenuous. He thought that he acquired malaria again, since these symptoms were similar to a bout of malaria he had experienced a few years before. The local doctor gave him Chloroquine, an anti-malarial medication, with an initial dose of 10 mg/kg followed by 5 mg/kg the following 2 days. However, the symptoms did not subside, and he started experiencing pneumonia-like symptoms - chest pains and difficulty breathing. His wife decided to bring him to the hospital. He was almost unable to walk into the clinic from the parking lot. The clinic doctor diagnosed his symptoms as influenza with pneumonia, and recommended inpatient treatment. However, due to lack of health insurance and limited finances for inpatient care, Mr. Mburu decided to go home with a prescription for Cloxacillin 500 mg orally every 6 hours that the clinic doctor gave him. Since it was too late to treat the influenza virus itself, the doctor did not prescribe any anti-viral medication.

Specification of FPs

Mr. Mburu has been diagnosed with Influenza. Option 1, the default for any health condition has been used to specify the FPs.

Binary rule for coding

In the example of Mr. Mburu, the FPs have been coded as follows:

Life Management Activities Impact
Carrying out daily routine (G)	d230.8
Mobility Impact
Walking (G)	d450.8
Moving around (G)	d455.8
Work Activities Impact
Remunerative employment (G)	d850.8

The codes listed in the far right column are then documented in the distinct data field for FPs in the ICD-11.

ICF qualifier rule for coding

Life Management Activities Impact
Carrying out daily routine (G)	0	1	2	3	4	d230.3
Mobility Impact
Walking (G)	0	1	2	3	4	d450.2
Moving around (G)	0	1	2	3	4	d455.3
Work Activities Impact
Remunerative employment (G)	0	1	2	3	4	d850.3

The codes listed in the far right column are then documented in the distinct data field for FPs in the ICD-11.

Example 2: Recurrent Depressive Disorder

Ms. Dupont, 70 year-old former bookkeeper, has been living on and off with depression since her late teenage years. Unrecognized as clinical depression and feeling misunderstood by her doctor at the time, she did not return to see him, leaving her depression untreated. When she became pregnant her partner was supportive, but a few months later they separated. Not able to work, financial worries and inability to handle the stress exacerbated Ms. Dupont’s existing sleep problems. Despite seeing various health professionals her underlying depression continued to be undertreated. After the birth of her son she tried unsuccessfully to contain the symptoms, thinking that her emotional problems were a result of her situation rather than due to an illness. She often had difficulties in making decisions and juggling her daily routine, especially with a newborn. When her son started school, her depression subsided. However, after a few years, her depression returned. She regularly experienced migraines and back pain, increasingly took time off from work. Facing reproach from her colleagues, Ms. Dupont felt so alone. While her general practitioner found no organic cause for her health problems, another doctor diagnosed her with depression and prescribed medication (Prozac). Despite some improvement, she lost her job. Afterward she discontinued Prozac, she started having suicidal thoughts. Consequently Ms. Dupont’s family arranged for therapy with a psychologist in a combined practice with a psychiatrist. The psychiatrist prescribed amitriptyline 60 mg/day. Pharmacological therapy with psychological counselling, Ms. Dupont has been able to control her depression for many years. Stabilized on medication, she ended psychological counselling 10 years ago. Her son is now grown and has moved away, but they still have regular phone contact – a weekly highlight. Like many elderly people, she spends much of her time alone. She sometimes struggles to carry out her daily routine, including self-care, and decision-making is often difficult. Her psychiatrist suggested increasing her dosage of amitriptyline from 60 mg to 75 mg/day. However, there was concern that the daytime drowsiness and periodic dizziness that she had been experiencing for years will consequently increase. Due to these symptoms Ms. Dupont recently fell down a short flight of stairs. Additionally, she has been getting more and more constipated, almost every other day.

Specification of FPs

Ms. Dupont has been diagnosed with recurrent depressive disorder. Depression has been identified as one of the 100 rehabilitation-relevant health conditions and an evidence-based specification of the FPs is available in the ICD-11. This set consists of 16 FPs, including the 4 FPs specified in Option 1.

Binary rule for coding

In the example of Ms. Dupont, the FPs have been coded as follows:

Understanding Impact
Solving problems	d175.0
Making decisions	d177.8
Thinking	d163.0
Life Management Impact
Carrying out daily routine (G)	d230.8
Managing daily routine	d2301.8
Managing one's own activity level	d2303.8
Handling stress and other psychological demands	d240.0
Communication Impact
Conversation	d350.0
Mobility Impact
Walking (G)	d450.0
Moving around (G)	d455.8
Self-Care Impact
Washing oneself	d510.0
Looking after one's health	d570.0
Interpersonal Relations Impact
Family relationships	d760.0
Intimate relationships	d770.0
Work Activities Impact
Acquiring, keeping and terminating a job	d845.0
Remunerative employment (G)	d850.0

The codes listed in the far right column are then documented in the distinct data field for FPs in the ICD-11.

ICF qualifier rule for coding

Understanding Impact
Solving problems	d175.0
Making decisions	d177.8
Thinking	d163.0
Life Management Impact
Carrying out daily routine (G)	d230.8
Managing daily routine	d2301.8
Managing one's own activity level	d2303.8
Handling stress and other psychological demands	d240.0
Communication Impact
Conversation	d350.0
Mobility Impact
Walking (G)	d450.0
Moving around (G)	d455.8
Self-Care Impact
Washing oneself	d510.0
Looking after one's health	d570.0
Interpersonal Relations Impact
Family relationships	d760.0
Intimate relationships	d770.0
Work Activities Impact
Acquiring, keeping and terminating a job	d845.0
Remunerative employment (G)	d850.0

The codes listed in the far right column are then documented in the distinct data field for FPs in the ICD-11.

Example 2: Rheumatoid Arthritis

Mrs. Baker, a 42 year-old woman, has been suffering from rheumatoid arthritis (Rheumatoid factor seropositive, erythrocyte sedimentation rate (ESR) 25 mm/h, haemoglobin 10g/dl) for about 10 years. Both hands, her right shoulder and right knee are affected by the disease (Subluxation MCP joints 2 and 3 right, commencing ulnar drift of right hand; degenerative changes of proximal and distal interphalangeal joints right hand, dig. 2 and 3). The physical examination revealed an ulnar drift, swan-neck deformity of the little finger and Boutonnière deformity of the index finger of her right hand and her right knee was moderately swollen. The range of motion (ROM) in her right should (140°/0/30°), hand joints (volar flexion/dorsiflexion right 30°/0/0° and left 35°/0/10°; radial flexion/ulnar flexion right 0°/5/20° and left 0°/0/20°) and knee (flexion/extension right 95°/10/0° and left 120°/0/0°) were reduced. Additionally, Mrs. Baker experiences pain with every movement especially in the morning. The restricted movement due to the swelling and pain, especially in her fingers and shoulder, make dressing by herself difficult and complicates other self-care activities. Additionally, carrying out daily routine such as cleaning and cooking is problematic due to Mrs. Baker’s right shoulder and hands. Besides difficulties experienced in activities of daily living, Mrs. Baker had to give up her job as a secretary, since major job tasks such as typing was extremely difficult due to the swelling and deformation of her fingers and pain that ensued while typing. Treatment and interventions included medication and physical therapy once a week to mobilize the joints. During a hospital visit, the occupational therapist recommended using an electric opener for opening up cans and a special grip for knives to facilitate cooking, and provided Mrs. Baker with a wrist splint for joint protection and to help quiet inflammation. Mrs. Baker was still unable to work.