Background: In Uganda, a previous study reported high HIV prevalence in persons with severe mental illness (SMI) compared to the general population, suggesting that persons with SMI might constitute a high-risk group for HIV. However, the study included first-time psychiatric admissions only, a group whose HIV prevalence may not reflect the prevalence in persons with SMI in general. We determined prevalence and correlates of HIV in both first-time and previous psychiatric admissions, in a psychiatric hospital in Uganda.

Methods: Cross-sectional study of HIV status in persons consecutively discharged from psychiatric admission wards in Butabika hospital, Uganda. Inclusion criteria: age 18–49 years; schizophrenia, bipolar disorder, depression, or other non-substance-use-related psychosis; Luganda or English proficiency. Exclusion criterion: Mental incapacity to give informed consent. Participants were HIV-tested, and interviewed using a structured questionnaire. Data were analysed using logistic regression.

Results: HIV prevalence was 11.3% (CI 8.8-13.8) overall, 7.3% (CI 4.1-10.5) in men and 14.3% (CI 10.6-18.0) in women. Females had higher risk of HIV infection than males (OR 2.10; CI 1.20-3.67), after adjustment for age. Older patients had higher risk of HIV infection than younger patients (40–49 vs. 18–29 years: OR 2.34; CI 1.27-4.32), after adjustment for sex. Place of residence, marital status, income, education, occupation, psychiatric diagnosis and history of previous admission were not associated with HIV infection, after adjustment for sex and age. The above associations did not significantly differ between men and women.

Conclusions: Persons admitted for SMI in Uganda have higher HIV prevalence than persons in the general population, irrespective of previous admissions. The excess HIV prevalence is mainly confined to women. The findings call for the integration of HIV prevention, testing and care with mental health services in settings with generalized HIV epidemics. Moreover, further research is needed to clarify the mechanisms underlying the increased HIV prevalence in women with SMI in Uganda, and to identify effective community-based interventions for this vulnerable group.