Co-occurrence of AIDS and tuberculosis: Results of a database ''match'' and investigation

Purpose: To conduct a match between the AIDS and Tuberculosis (TB) Registries in Philadelphia. Methods: Database extracts for the year 1993 were prepared manually and matched by name, birth date, and social security number. Reported cases not matching with those on the primary registry were investigated. Proportion levels of comorbidity were calculated. Predictors of comorbidity were evaluated separately for the TB sample and for the AIDS sample. Results: The proportion of comorbid patients with AIDS alive at midyear was 4.7%; 17.1% of TB cases were also HN-positive. Twenty-three percent of AIDS cases were falsely reported as having active TB; the false-positive rate in TB Control for HIV seropositivity was 4.2%. Having public or no health insurance, injection drug use (IDU) or heterosexual risk background (HET), and being nonwhite and female were significant predictors of active TB in persons with AIDS. Conclusions: A registries' data match can provide useful information and result in improved validity for both registries. Although women with AIDS initially appeared to have a higher risk of having active TB, additional parallel analyses suggested that this effect was primarily the result of the 1993 expansion of the definition of AIDS.