Qualitative Results from a Pilot Study of an Automated Directly Observed Therapy Intervention Using Artificial Intelligence with Conditional Economic Incentives among Young Adults with HIV

Scritto il 10/05/2025
da L Danielle Wagner

AIDS Behav. 2025 May 10. doi: 10.1007/s10461-025-04748-1. Online ahead of print.

ABSTRACT

Digitally monitoring and supporting daily antiretroviral therapy (ART) is a promising strategy for enhanced adherence among young adults with HIV (YWH). We implemented an innovative mobile app-based intervention that included automated directly observed therapy (aDOT) using artificial intelligence, medication reminders, and conditional economic incentives (CEIs). This intervention used facial recognition to track and record YWH's daily ART adherence. CEIs were given based on daily app use. A pilot study of the aDOT + CEI app intervention was conducted among YWH ages 18-29 for 3 months. We recruited virally unsuppressed YWH from AIDS Healthcare Foundation clinics in California and Florida and conducted qualitative interviews with YWH and providers regarding the intervention's motivators, deterrents, and helpfulness. Thirteen YWH and 5 providers were interviewed. YWH found the CEIs supported their adherence; however, most YWH did not find the CEI amount sufficiently motivating. The primary motivator for consistent app use was to improve health, and the app reminders were most helpful in improving adherence. YWH desired more reassurance and explanation about app privacy features. Providers suggested incorporating the intervention into counseling for newly diagnosed patients and recommended the intervention to help facilitate viral suppression with oral ART for those interested in transitioning to long-acting injectable ART. YWH and providers found the intervention helpful in reminders to take ART and in tracking adherence. Refinements to the CEI amount may further enhance the intervention. Additional research is needed to test the efficacy of a refined aDOT + CEI intervention to improve viral suppression in a larger sample.

PMID:40347396 | DOI:10.1007/s10461-025-04748-1