Objective: We evaluated whether heavy alcohol use, illicit drug use or high levels of anxiety, and depression symptoms were modifiers of the retention through enhanced personal contact intervention. The intervention had previously demonstrated overall efficacy in the parent study. Design: Randomized trial. Methods: A total of 1838 patients from six US HIV clinics were enrolled into a randomized trial in which intervention patients received an 'enhanced contact' protocol for 12 months. All participants completed an audio computer-assisted self-interview that measured depression and anxiety symptoms from the Brief Symptom Inventory, alcohol use from the Alcohol Use Disorders Identification Test-Consumption instrument, and drug use from the WHO (Alcohol, Smoking and Substance Involvement Screening Test) questions. The 12-month binary outcome was completing an HIV primary care visit in three consecutive 4-month intervals. The outcome was compared between intervention and standard of care patients within subgroups on the effect modifier variables using log-binomial regression models. Results: Persons with high levels of anxiety or depression symptoms and those reporting illicit drug use, or heavy alcohol consumption had no response to the intervention. Patients without these 'higher risk' characteristics responded significantly to the intervention. Further analysis revealed higher risk patients were less likely to have successfully received the telephone contact component of the intervention. Among higher risk patients who did successfully receive this component, the intervention effect was significant. Conclusion: Our findings suggest that clinic-based retention-in-care interventions are able to have significant effects on HIV patients with common behavioral health issues, but the design of those interventions should assure successful delivery of intervention components to increase effectiveness.