[Skip to Content]
[Skip to Content Landing]
May 27, 1992

The Effect of Changing the Definition of AIDS on the Modeling of AIDS

Author Affiliations

The Multicenter AIDS Cohort Study Baltimore, Md
UCLA School of Public Health Los Angeles, Calif
Johns Hopkins School of Public Health Baltimore, Md
Northwestern University Medical School Chicago, Ill
University of Pittsburgh (Pa) School of Public Health

JAMA. 1992;267(20):2737-2738. doi:10.1001/jama.1992.03480200045011

To the Editor.  —Effective September 1987, the Centers for Disease Control added five diagnoses to the adult AIDS (acquired immunodeficiency syndrome) surveillance definition: (1) human immunodeficiency virus type 1 (HIV-1) encephalopathy; (2) disseminated Mycobacterium tuberculosis; (3) disseminated coccidioidomycosis; (4) wasting syndrome; and (5) recurrent Salmonella septicemia.1These additional diagnoses increased reported AIDS incidence and shortened time to AIDS in all risk groups, but particularly in those at risk through intravenous drug use and heterosexual contact.2 There were other 1987 AIDS definition changes that eliminated laboratory confirmation for some AIDS-defining diagnoses, but these changes probably did not affect time to AIDS or caseload, since misdiagnosis was unlikely.3The 1987 change in the AIDS definition could confound AIDS modeling with results based on the pre-1987 definition biased toward longer incubation times and fewer cases. When backcalculation is used to predict HIV-1/AIDS incidence based on historical AIDS diagnosis, adjustments for