Variations in self-rated health among patients with HIV infection

To assess how patients with HIV who are enrolled in a clinical trials cohort rate their health and to compare their ratings with those of patients with HIV from 2 other cohorts: the HIV Cost and Services Utilization Study (HCSUS), and Adult AIDS Clinical Trials Group protocol 320 (ACTG 320).

We analyzed baseline information for the 1649 subjects enrolled in the Adult AIDS Clinical Trials Group Longitudinal Linked Randomized Trials (ALLRT) study prior to March 2002 who had self-rated health data available. We compared those results with results from 2 other groups: HCSUS, the only nationally representative sample of people in care for HIV in the U.S., which conducted baseline interviews in 1996 and 1997, and ACTG 320, a randomized, double-blinded, placebo-controlled trial comparing a 3-drug antiretroviral regimen with a 2-drug combination, which enrolled subjects in the same general time frame as HCSUS. We used t tests, Pearson correlations, and linear regression to determine factors associated with self-rated health and z scores to compare results between cohorts.

The mean (SD) rating scale value on a 0-100 scale for ALLRT participants was 79.8 (16.8). Values were significantly lower for subjects who were older, had a history of injection drug use, had lower CD4 cell counts, or were beginning salvage antiretroviral therapy. Subjects in ALLRT reported significantly better self-rated health at baseline than those in HCSUS or ACTG 320 (11-12% higher rating scale values in ALLRT; p<0.05). When cohort differences were accounted for through regression and stratification, the differences in scores between subjects in ALLRT and HCSUS increased and the differences in scores between subjects in ALLRT and ACTG 320 diminished.

Self-rated health varied significantly by age, CD4 count, injection drug use history, and salvage therapy status. Differences in self-rated health for clinical trials and non-clinical trials samples appear to be substantial and should be considered when applying trial results to clinical populations.