To ascertain the prevalence and predictors of psychiatric distress in an inner-city HIV primary care clinic. Cross-sectional study. Inner-city adult HIV clinic. A series of 222 HIV-infected patients newly presenting to the medical clinic for evaluation over a 1-year period. A screening method, based on the General Health Questionnaire (GHQ) and the Beck Depression Inventory (BDI), whose sensitivity, specificity and positive predictive value for psychiatric diagnosis were previously established in this clinic. Fifty-two per cent of participants scored above the screening threshold (i.e., scored > 14 on the BDI or > 6 on the GHQ). A comorbid substance use condition was the most powerful and consistent predictor of psychiatric distress (P < 0.05). Limited education and current unemployment contributed to higher scores on the BDI or the GHQ (P < 0.05). However, HIV illness variables and psychiatric personal or family histories were not significant predictors of psychiatric distress (P > 0.05 in all cases). Rates of psychiatric distress in inner-city adult HIV clinics are much higher than in the general population or than in other outpatient medical clinics. They are also not associated with what most clinicians perceive as traditional risk groups such as psychiatric histories and social disadvantage. These findings support the position that easy access to psychiatric care is essential to HIV clinics.