A System to Convert ICD Diagnostic Codes for Alcohol Research

The coding of medical diagnoses for morbidity and mortality in the United States and elsewhere is based on a standard system of disease classification developed by the World Health Organization. This system has been updated every 10 years since 1900 to reflect current medical knowledge. The most recent edition, the International Classification of Diseases Ninth Revision (ICD-9), was placed in use to classify causes of death beginning with 1979 data. ICD-9 replaced the International Classification of Diseases Adapted for Use in the United States Eighth Revision (ICDA-8), which was used from 1969 through 1978.

The ICD-9 diagnostic categories are more specific than those in ICDA-8. Ninth Revision changes related to alcoholism include: (1) discrimination of four types of chronic alcoholic liver disease (ICDA-8 recognized only alcoholic liver cirrhosis); (2) introduction of alcoholic cardiomyopathy and excessive blood level of alcohol as conditions not previously differentiated; and (3) separation of alcoholism into two categories--alcohol dependence syndrome and nondependent abuse of alcohol.

Investigators and policy planners who use data on alcohol-related mortality need to accommodate to the transition from ICDA-8 to ICD-9 in two ways. First, when studying trends or combining data from years before and after the transition (between 1978 and 1979), they should match the ICD-9 categories with ICDA-8 categories to make the data comparable. Second, they should identify alcohol-related conditions in ICD-9 that were not distinguished in ICDA-8 to provide more complete coverage of alcohol-related health problems.

A conversion table or crosswalk has been developed that alcohol investigators can use to recode alcohol-related underlying cause-of-death data from ICD-9 to ICDA-8.

Several steps are necessary in the recoding process: (1) reviewing the diagnoses included in each alcohol-related category in the two systems and identifying categories--or groups of categories--in ICD-9 that match those in ICDA-8; (2) examining the average yearly number of deaths in each ICDA-8 category in the interval 1975-1978 and comparing the average number of deaths with the number of deaths in corresponding ICD-9 categories in 1979; and (3) adjusting the matched categories to reconcile discrepancies in the data that might result from differences between coding rules and actual practices.

Table 3 presents a list of the new alcohol-related categories in ICD-9. With the exception of "excessive blood level of alcohol," these conditions were included in other ICDA-8 classifications but were not assigned specific codes. The number of deaths attributed to each of these causes in 1979 is shown in column 3. Studies of alcohol-related mortality using data coded in ICD-9 should include these new ICD-9 categories, as well as those considered in earlier studies.

The completed crosswalk of ICD-9 categories to those in ICDA-8 is shown in Table 4. Persons who use mortality data may wish to apply or modify this crosswalk or to meet their requirements. Table 5 presents data related to each of the major groups of alcohol-related categories under the two systems. To allow for comparison, the number of deaths in 1979, based on the crosswalk categories for ICD-9, is shown along with the average yearly number of deaths in the corresponding ICDA-8 categories for the interval 1975-1978.

The number of deaths in the matched categories are generally similar from ICD-9 to ICDA-8, suggesting the crosswalk is valid. However, a detailed review reveals some discrepancies: somewhat more alcoholic psychosis deaths under ICD-9 (1979) than under ICDA-8 (1975-1978); fewer deaths from alcoholic liver disease and from other specified chronic liver diseases. However, there were more deaths in 1979 from unspecified chronic liver disease and from alcohol poisoning. These differences may reflect three factors: (1) actual changes in patterns of mortality from one period to another; (2) differences in diagnostic concepts applied in the Eighth and Ninth Revisions; and/or (3) discrepancies between interpretations and applications of codes. Reported by JD Colliver, PhD, P Van Natta, MA, Alcohol Epidemiologic Data System, CSR, Incorporated; M Dufour, MD, D Bertolucci, MA, H Malin, MA, Div of Biometry and Epidemiology, National Institute on Alcohol Abuse and Alcoholism.

Editorial Note

Editorial Note: Assessment of discontinuities associated with a crosswalk requires a study of the comparability of matched diagnoses under the two systems, usually based on dual coding of a set of records. No detailed comparability study of ICD-9 to ICDA-8 categories applied to mortality data has yet appeared in the literature. The crosswalk of alcohol-related ICD-9 diagnostic categories to those in ICDA-8 provides one of several possible models appropriate for application in studies of alcohol-related mortality. While adoption of a standard crosswalk would have great benefits, a certain level of discontinuity associated with the matches in some categories is evident. Each user should examine the classifications and the data in detail and apply this crosswalk, unless study requirements necessitate a modification. The details of any changes made for a particular study should be documented in the study report.

A comparability study of limited scope was conducted by the National Center for Health Statistics (1) using a 7% sample of death certificates from 1976. Comparability ratios were produced for 72 selected aggregate causes of death reported in many of the tables published in the Monthly Vital Statistics Report and in the annual mortality volumes of Vital Statistics of the United States. Chronic liver disease and cirrhosis, the only directly alcohol-related diagnosis included in the 72 causes, was represented in aggregate, without distinction between alcohol and nonalcohol varieties; ICD-9 code 571 was crosswalked to ICDA-8 code 571. A comparability ratio of 1.011 (i.e., 1.1% more deaths under ICD-9 than under ICDA-8) was reported. However, any deaths from portal hypertension (ICD-9 code 572.3) were excluded in the ICD-9 count for the category.

It should be noted that matches of diagnoses proposed in the present crosswalk may differ from matches used in other studies and that correspondences may vary according to whether mortality or morbidity data are involved. In addition, proposed matches apply to underlying cause of death and may not be appropriate for data on multiple cause of death.

For more detailed information, contact Henry Malin, National Institute on Alcohol Abuse and Alcoholism, Parklawn Building, Room 14C26, 5600 Fishers Lane, Rockville, Maryland 20857.

Reference

1. National Center for Health Statistics. Estimates of selected comparability ratios between dual coding of 1976 death certificates by the Eighth and Ninth Revisions of the International Classification of Diseases. Monthly Vital Statistics Report, Supplement, February 29, 1980;28(11).

Disclaimer   All MMWR HTML documents published before January 1993 are electronic conversions from ASCII text into HTML. This conversion may have resulted in character translation or format errors in the HTML version. Users should not rely on this HTML document, but are referred to the original MMWR paper copy for the official text, figures, and tables. An original paper copy of this issue can be obtained from the Superintendent of Documents, U.S. Government Printing Office (GPO), Washington, DC 20402-9371; telephone: (202) 512-1800. Contact GPO for current prices.

Page converted: 08/05/98

HOME  |  ABOUT MMWR  |  MMWR SEARCH  |  DOWNLOADS  |  RSS |  CONTACT POLICY  |  DISCLAIMER  |  ACCESSIBILITY

Morbidity and Mortality Weekly Report Centers for Disease Control and Prevention 1600 Clifton Rd, MailStop E-90, Atlanta, GA 30333, U.S.A Department of Health and Human Services

This page last reviewed 5/2/01