Vol. 17, No. 2	July 1996

REVISIONS OF THE INTERNATIONAL CLASSIFICATION OF DISEASES (ICD-9 AND ICD-10): IMPACT ON HEALTH STATISTICS

Introduction
Main Diferences between the Ninth and Tenth Revisions
Bridge Studies
Conclusion

Introduction

One of the continuing concerns about changing from one revision of the International Classification of Diseases (ICD) to the next is the potential impact of the changes on health statistics. To what extent do the differences between revisions affect the continuity of historical series and facilitate, hamper, or impede comparison of the data?

This concern is even more valid for the current implementation of the Tenth Revision, because the new revision contains the most sweeping changes since the Sixth Revision, which was introduced in 1949.

An article in an earlier issue of the PAHO Epidemiological Bulletin (Vol. 16, No. 1, March 1995) outlined the broad features of the Tenth Revision of the International Statistical Classification of Diseases and Related Health Problems. The main differences between the Ninth and the Tenth Revisions that could have an impact on health statistics are listed below.

Return to top

Main Differences between the Ninth and Tenth Revisions

• The exclusion notes at the beginning of each chapter have been expanded to explain the relative hierarchy of chapters, and to make it clear that the "special groups" chapters have priority of assignment over the organ or system chapters and that among the special group chapters, those on Pregnancy, childbirth, and the puerperium and on Certain conditions originating in the perinatal period have priority over the others.

• Categories have been created at the end of certain chapters for postprocedural disorders. These categories include important conditions that constitute medical problems in their own right--for example, endocrine and metabolic diseases following ablation of an organ and other specific conditions, such as postgastrectomy dumping syndrome. Postprocedural conditions that are not specific to a particular organ or body system, including immediate complications such as air embolism and postoperative shock, continue to be classified under the chapter "Injury, poisoning and certain other consequences of external causes".

• For tumors (neoplasms), a category (C97) was created for Malignant neoplasm of independent (primary) multiple sites. As a result, causes previously coded to one of the sites mentioned, are now assigned to this new category. Another important change is that in the Tenth Revision the existing codes for "secondary malignant neoplasm" are acceptable as underlying cause of death when there is no information as to the primary site and the morphologic type does not identify it. In the Ninth Revision this was not acceptable, and in such cases "unspecified site" was to be coded with respect to the morphologic type (199.1 for carcinoma, 171.9 for sarcoma, and 172.9 for melanoma).

• The Ninth Revision rules, and related notes, for selection and modification of the underlying cause of death were revised and a number of changes were introduced-- in the rules and, particularly, the notes. The modification rules were simplified and went from 9 in the Ninth Revision to 6, identified by the letters A-F.

• The notes for use in coding the underlying cause of death underwent extensive modifications, which are found in Volume 2 (pp. 50-62).

• Some changes in the acceptance of causality (sequence) can also create difficulties for the comparison of data, for example:
  1. Pneumonia and bronchopneumonia can be accepted as terminal complications, if they are registered as due to, or with mention of malignant neoplasms, malnutrition, paralyzing diseases, communicable diseases, or serious injuries. This could lead to an apparent reduction in pneumonia and bronchopneumonia, with a consequent increase in the aforementioned disorders.

  2. Chickenpox and herpes zoster can be accepted as consequences of diabetes mellitus, tuberculosis, or lymphoproliferative neoplasms, with consequences identical to those mentioned above.

  3. Malignant neoplasms and infectious and parasitic diseases can be accepted as a consequence of HIV infection. They can also be accepted as a consequence of immunosuppression by chemotherapy and radiation or of tumors affecting the immune system.

The table that follows summarize some of the most significant differences between the two revisions that could make it difficult to compare historical series of health statistics.

Table 1.

Comparison between the Ninth and Tenth Revisions of the International Classification of Diseases

Ninth Revision

Tenth Revision

- International Statistical Classification of Diseases, Injuries and Causes of Death - 17 Sections - Two supplementary classifications: (a) External Causes of Injury and Poisoning (E800 - E999) (b) Factors Influencing Health Status and Contact with Health Services (V01 - V82) - Section III, Endocrine, Nutritional, and Metabolic Diseases, and Immunity Disorders (240 - 279) - Section VI, Diseases of the Nervous System and Sense Organs (320 - 389) - Classification: Base: 909 categories (183 "X" categories --that is, categories that do not have fourth digit) 5,161 subcategories Suppl. "E:" 192 categories 1,001 subcategories Suppl. "V:" 77 categories 537 subcategories Total: 6,882 codes - Groups: 141 - "*": 4 categories and 67 subcategories dispersed among different categories, with a total of 98 "*" codes." - Fifth digit: Tuberculosis: confirmation Diabetes: adult/juvenile Complications of childbirth: method of delivery Musculoskeletal: anatomical site no: subcategory level no: subcategory level Accidents: place of occurrence . no used - Some infectious diseases are coded in diferent sections. 037 Tetanus 634-639 with fourth digit. 0--tetanus complicating abortion 670 Obstetrical tetanus 771.3 Tetanus neonatorum - AIDS is coded in 279.5 and 279.6 (free subcategories of category 279, Disorders involving the immunie mechanism) - Complications affecting specified body systems are coded in 997.-, under section XVII (Injury and poisoning). Example: 997.0 Central nervous system complications

- International Statistical Classification of Diseases and Related Health Problems - 21 Chapters - Are now in the core classification: (a) Chapter XX, External causes of morbidity and mortality (V01 - Y98) (b) Chapter XXI, Factors influencing health status and contact with health services - Chapter III, Diseases of the blood and blood-forming organs and certain disorders involving the immune mechanism (D50 - D89) - Chapter VI, Diseases of the nervous system (G00 - G99) - Chapter VII, Diseases of the eye and adnexa (H00 - H59) - Chapter VIII, Diseases of the ear and mastoid process (H60 - H95) - Classification: 2,036 categories (261 "X" categories) 12,159 subcategories Both are Chapters of the main classification Total: 12,420 codes - Groups: 261 - "*": 83 complete categories, with a total of 298 codes - Fifth character: . no: category level . not used . no: category level anatomical site Fractures: open/closed, internal injuries: with or without open wound no: subcategory level Accidents: activity involved - Moved to Chapter I (Certain infectious and parasitic diseases) A35 Other tetanus A34 Obstetrical tetanus A33 Tetanus neonatorum - Code in B20-B24, under Chapter I (Certain infectious and parasitic diseases) - Coded within the chapters corresponding to the body systems. (E89.-, G97.-, H59.-, H95.-, I97.-,J95.-, K01.-, M96.-, and N99.-) G97. - Pstprocedural disorders of the nervous system, not elsewhere classified.

Return to top

Bridge Studies

The term "bridge studies" is customarily used to describe the procedures for coding medical information (death certificates or morbidity registries) with two successive revisions of the ICD in order to study the impact of the change in Classification. These studies are generally conducted by the WHO Collaborating Centers in Classification of Diseases during the preparation of a new revision.

An example of such a study is the one conducted in 1987 by CEVECE (Venezuelan Center for the Classification of Diseases), whose findings were presented at an expert meeting sponsored by PAHO in Buenos Aires, Argentina.

Although the study was conducted using draft versions of the ICD-10 chapters, the adjustments resulting from the final version do not alter the results.

A sample of 3,838 death certificates was used, with selection of the underlying cause of death based on the Ninth and Tenth Revisions. The results are presented in table 2 below.

TABLE 2.
Distribution of Death Certificates by Cause,
ICD-9 and ICD-10, Venezuela, 1995

ICD - 9	(1)	ICD-10	(2)	(2)/(1)
I - 001-139 II - 140-239 III - 240-279 IV - 280-289 V - 290-319 VI - 320-389 VII - 390-459 VIII - 460-519 IX - 520-579 X - 580-629 XI - 630-676 XII - 680-709 XIII - 710-739 XIV - 740-759 XV - 760-779 XVI - 780-799 EXT - E800-E999 Total	325 506 173 29 12 88 1,136 325 157 43 11 9 23 90 302 138 471 3,838	A00-B99 C00-D48 E00-E90 D50-D89 F00-F99 G00-H95 I00-I99 J00-J99 K00-K93 N00-N99 O00-O99 L00-L99 M00-M99 Q00-Q99 P00-P96 R00-R99 V01-Y98 Total	330 507 159 30 12 92 1,136 350 158 43 11 9 22 91 297 119 472 3,838	1.0154 1.0020 0.9191 1.0345 1.0000 1.0455 1.0000 1.0769 1.0064 1.0000 1.0000 1.0000 0.9565 1.0111 0.9834 0.8623 1.0021 1.0000

As the data indicates, 39 death certificates (1.02%) were coded to different chapters in the ICD-10 compared to the ICD-9. The most relevant differences are found in sections III (Nutritional and Metabolic Disorders), VIII (Diseases of the Respiratory System) and XVI (Signs, Symptoms and Ill-defined Conditions). Such differences are attributable to the criteria for coding terms such as "dehydration," "insufficiency," and "respiratory failure," and not to real changes in the Classification. The increase in Section I (Infectious and Parasitic Diseases) corresponds to cases of neonatal tetanus, previously coded to Section XV (Certain Conditions Originating in the Perinatal Period).

Naturally, if chapter subdivisions or isolated categories are compared, differences will be found to be greater.

At the Meeting of Directors of WHO Collaborating Centers for Classification of Diseases held in Canberra, Australia (October 10-16, 1995), preliminary data was presented from another bridge study conducted by the National Health Council of Denmark, which compared ICD-8 and ICD-10 coding of 5,256 death certificates (Denmark has never used the ICD-9).

The study, which is in press, found that 95% of the certificates were coded to the same chapter of both revisions, determining that in most cases, the differences lie in the criteria utilized and not the Classification. The distribution of the causes of death in the 49 categories of the "Danish List" ("DK-listen"), indicated that 92% of certificates were coded to the same category.

Return to top

Conclusion

As was verified, there is no equivalence between codes of the successive revisions. The Ninth and the Tenth Revisions have perhaps a 60% 1 x 1 equivalence--that is, a code (category or subcategory) in the Ninth corresponds to a single code in the Tenth, where the same diagnostic terms would be coded. As for the rest, in general there are more codes in the Tenth, but the opposite is also true. Moreover, there are codes that exclude only some of the terms of the previous Revision, or that exclude certain age groups.

When trends are compared over time and the statistics are interpreted, it is important to bear in mind that the presumptions can change from one ICD revision to the next. For example, before the Eighth Revision, an unspecified aneurism of the aorta was presumed to be due to syphilis, which is no longer the case.

The last sentence of Volume 1, page 26, paragraph 6 of the ICD-10 in English ("A key for conversion from the Ninth to the Tenth Revision, and the reverse, should be available before the implementation of the Tenth Revision.") should not be interpreted to mean a kind of "software" capable of assigning Tenth Revision codes on the basis of data coded under the Ninth Revision and vice versa. What is actually being developed is a kind of "Index" where, using a code (for example, from the ICD-9), it is possible to obtain one or, frequently, many codes corresponding to the ICD-10, in addition to a "multirevision index," in which the code corresponding to a term in more than one revision can be obtained.

What makes it possible to the maintain the historical series and comparability of the data coded under different revisions is the preparation and management of the lists for the presentation of the data. Normally, the data are not analyzed or broken down into all the ICD codes but into defined lists (e.g., Eighth Revision, Lists A, B, C; Ninth Revision, Basic Tabulation List, PAHO 6/61, ICD-BR2, ARG-3; Tenth Revision, Mortality tabulation lists 1, 2, 3; etc.). Lists can be prepared for a particular revision that retain the same structure and the same basic categories as the lists used with other revisions of the ICD. The "conversion key" should therefore be individually adapted to each list utilized.

Source: Division of Health and Human Development, Program on Health Situation Analysis, SHA, PAHO.

Return to top
Return to Index

---

Index of Online Bulletins