The International Statistical Classification of Diseases and Related Health Problems (ICD) is a comprehensive classification of causes of morbidity and mortality, and is published by the World Health Organisation. The previous 9th revision (ICD-9) was published in 1975 and came into use in hospital information systems in 1979. It was superseded by the 10th revision (ICD-10) from April 1995.

ICD permits the systematic analysis, interpretation, and comparison of morbidity data collected in different areas and at different times. Thus, the specified purpose of the ICD is to provide a means of classifying medical terminology and is defined as a system of categories to which morbid entries are assigned according to established criteria.

Categories have been chosen and logically sequenced to facilitate the statistical study of disease phenomena. A specific disease entity that is of particular public health importance, or that occurs frequently, has its own category. Otherwise, categories are assigned to groups of separate but related conditions.

Each category has a title reflecting its composition, and an alphanumeric code as a means of unique identification.

ICD-10 is published as a three volume set:

• Volume 1 is the Tabular List, comprising a descriptive classification of diseases and injuries, supplementary classifications, and appendices;
  
• Volume 2 is the Instruction Manual which contains the World Health Organisation's guidelines for coding;
  
• Volume 3 is the Alphabetical Index to diseases and nature of injury. It covers external causes of injury and provides a Table of drugs and chemicals.

In addition, NHS Data Standards & Products also distributes the following files:

• codes and titles file, containing a list of ICD-10 codes together with abbreviated titles;
  
• metadata file, containing codes, titles and validation criteria for each ICD-10 code;
  
• tables of equivalence files, identifying the relationship between ICD-9 and ICD-10 codes.

Further details regarding the use of individual diagnostic fields within the minimum data set Hospital Episode Statistics record format will be found within each listed Hospital Episode Statistics record, see Commissioning Data Sets Menu.

Coding Standards

The following standards have been developed to provide specific instruction in areas where ICD-10 allows a variety of approaches to be taken. Readers may wish to refer to Volume 2 of ICD-10 to obtain further information. Further detailed standards are included in the Clinical Coding Instruction Manual. The Coding Clinic insert of the Data Quality Review Newsletter publishes changes to national standards and these are included in the next following amendment set to the Instruction manual. The Data Quality Review Newsletter is available from NHS Classifications Service.

Adoption of basic coding guidelines (cf: Vol 2, 3.3 p28)

The basic principles of the classification, particularly those relating to basic coding guidelines and the structure of the classification, as detailed at 2.3 and 2.4, pages 11-13 of Volume 2, are to be adopted as the standard for the implementation of ICD-10.

Priority of Allocation to Special Groups (cf: Vol 2, 2.4, p13)

Where a diagnosis could be coded to either a Special Group Chapter or any of the other chapters, then it should always be coded to the appropriate Special Group Chapters.

Special Group Chapters include those chapters relating to epidemic diseases, constitutional and general diseases, developmental diseases and injuries. Therefore, these should be used in preference to chapters relating to local disease arranged by site.

Use of U and other unused codes (cf: Vol 2, 2.4.7, p17)

Codes U00 - U49 are reserved by the World Health Organisation for the future provisional assignment of diseases of uncertain etiology. Codes U50-U99 are potentially available for local use, for instance in research. However, no U codes may be used unless specifically designated for use by the NHS Classifications Service.

Filler "X" character for three character codes

Insertion of the character "X" into the fourth field of codes where only three characters exist is mandatory. This provides a standard "filler" character. The fourth character, "X", is only applicable to UNDIVIDED three-digit codes for which no valid fourth character exists; it is not an option for completing diagnoses which have not been coded to 4 - characters.

Use of fifth character

The fifth character will continue to be available for use within the entire Musculo-skeletal Chapter. Its use is recommended when the data is present in the source document, and where doing so adds information.

The fifth character for open fractures/internal injuries will continue to be available for use within the entire "Injury" section within the "Injury, poisoning and certain consequences of external causes" Chapter. Use of fifth character is recommended.

Use of Specialty-based adaptations

The World Health Organisation have provided an undertaking to ensure that any such classification endorsed by them will be fully compatible with the main classification at the fourth character level, but will not necessarily be compatible either at the fifth character level or with each other. For instance the meaning of a five character code in one Specialty adaptation can be quite different in another.

Use of such classifications, including the Royal College of Paediatrics & Child Health (formerly British Paediatric Association (BPA)) Classification of Diseases, and the application of the ICD to Dentistry and Stomatology is optional.

UK extensions to ICD-10

It is not intended to issue or support extensions to ICD-10. The other extensions previously issued for use with ICD-9 have either been incorporated into ICD-10, are no longer relevant, or are covered by separate arrangements.

Guidelines on coding "main" and "other" conditions, etc

Guidelines for coding "main condition" and "other" conditions, rules for reselection when the main code is incorrectly recorded, and chapter-specific notes are provided in Vol 2, 4.4.2-4.

Primary Diagnosis has been defined and was mandated for use in England from 1 April 1996. It is:

• the main condition treated or investigated during the relevant episode of healthcare,
• 
  and
  
• where there is no definitive diagnosis, the main symptom, abnormal findings or problem.

Using Secondary Diagnosis fields

The secondary diagnoses fields follow the primary diagnosis.

Status of "are for use with" characters (cf: Vol 1 pp276, 320, 680, 724, 1013, 1026, 1027, 1029, 1032, 1034, 1037, 1039, 1042, 1052)

Where the phrase "are for use with" is seen, this instruction is mandatory, and the codes referred to must be used. Note: these characters represent the fourth character subdivision of the related categories.

Status of "if desired" codes (cf: Vol 1 p496)

Where the phrase "if desired" is seen, in relation to coding additional information, where that information is present in source documentation this instruction is not optional.

Multiple body sites

Single codes identifying multiple body sites must not be used where the information is available to enable use of individual codes, with the exception of those identifying bi-laterality of the same limbs and apart from those specific instances detailed in other paragraphs in this section.

Dagger and asterisk system (Vol 2, 3.1.3, p20)

Provision has been made for dual classification by aetiology (denoted by a "†" or dagger symbol), and manifestation (denoted by a "*" or asterisk symbol).

Where conditions to be coded correspond to dagger/asterisk pairs, both should be recorded, in succeeding fields, with the dagger code first. Multiple asterisk codes with one dagger code are not allowed, due to validational difficulties. Each asterisk code must, therefore, be preceded by its own dagger code, even where this means repeating dagger codes.

Alpha characters are to be used instead of daggers and asterisks because of possible confusion with other characters (not all keyboards carry a uniform dagger symbol, but all carry uniform alpha characters). Upper case "A" (asterisk) and upper case "D" (dagger) should be used.

For those codes which include a dagger or an asterisk, the appropriate symbol should always be entered at the sixth position in the code.

Identification of Infecting Organism (cf: Vol 2, 3.1.3, p22 and Vol 1, p178)

Where the name of the Infecting Organism is not identified in the title of the three character rubric, an additional code must be used when the infecting organism has been identified.

Functional Activity of Neoplasms (cf: Vol 2, 3.1.3, p22, and Vol 1, pp182, 271)

These codes should be used as appropriate i.e. as defined in national standard training.

Organic, including Symptomatic, Mental Disorders (F00-F09) (cf: Vol 2, 3.1.3, p22 and Vol 1, p312)

The use of these categories is mandatory when the data is present in source documents and when recording it helps to identify all the elements of the diagnostic statement as presented in the case note.

Toxic Agents (cf: Vol 2, 3.1.3, p22)

The use of external cause codes (Chapter XX: External causes of morbidity and mortality) is mandatory when the data is present in the source documents.

Multiple coding of multiple injuries (cf: Vol 1, p892)

The single code for "multiple injuries" is only to be used where no further detail is present in source documents.

Nature of Injury (cf: Vol 2, 3.1.3, p22 and Vol 1, p1011)

Coding of both the external cause and the actual injury is mandatory where the information is present in source documents.

The code for the injury and the associated external cause code should be recorded in succeeding fields on the patient data record. In the case of more than one injury caused by the same event, one external cause code will serve all injuries, and must be sequenced after the final injury.

Alcohol Involvement (cf: Vol 1, p1122)

The use of these codes is recommended when the data is present in the source documents. These codes must not be used in isolation.

Use of Morphology codes

The use of morphology codes is optional. They can be used for local purposes, or where systems are provided for their transfer to Cancer Registers, but must be kept distinct from ICD-10 diagnosis codes.

Use of ICD-0

Where data is collected using ICD-0 (Oncology) classification, recording of both topographical and morphological codes is mandatory, as is translation of the data to ICD-10 using the appropriate conversion tables. Conversion tables are obtainable from the National Cancer Institute, EPN343J, Bethesda, Maryland, 2089, USA.

Unknown/unspecified causes of morbidity/mortality (cf: Vol 1 pp879, 890)

ICD-9 code 799.9 was used in cases of undiagnosed disease where the site or system involved had not been specified, and for unknown causes of morbidity and mortality. These categories must not be used where further information is available from any source e.g. test results, admission books, casualty records, X-ray records.

ICD-10 has two codes for these purposes:

• R69X Unknown and unspecified causes of morbidity;

• R99X Other ill-defined and unspecified causes of mortality.

Diarrhoea and Gastroenteritis (cf: Vol 1, p112)

Currently in the UK Diarrhoea and Gastroenteritis are presumed non-infectious in origin and must be coded to K52.9 in the absence of a clinical description of the disease as infectious.

HIV Coding

Use of the fourth character sub-divisions of B20-B23 to identify specific conditions is recommended where the information exists. Conditions classifiable to two or more subcategories of the same category should be coded to the .7 subcategory of the relevant category; the resulting conditions should then be coded as "other conditions". Double coding should only occur where value and specific detail is added. The exceptions to the rule are malignant neoplasms, where there is a need to identify malignant neoplasms for Cancer Registers. It is therefore necessary to double code these conditions with the appropriate B21 code plus the neoplasm code from Chapter II.