Adapting DBT for Persons Living with HIV/AIDS Using an Artificial Intelligence-Powered Conversational Agent

One in four older persons living with HIV/AIDS (PLWHA) report at least one suicide attempt in their lifetime, and the risk of death by suicide is 100 times higher in PLWHA than in the general population. In the United States, there are approximately 1.2 million PLWHA, and 35% are adults aged 55 and over. Older PLWHA experience unique combinations of stressors including externalized and internalized stigma from advanced age and disease status. Despite the increased risk and unique risk factors older PLWHA have for suicide mortality, there have been no interventions developed specifically for this population to address suicide risk. Dialectical Behavioral Therapy (DBT) shows significant impact at reducing suicidal behaviors by improving one’s self- efficacy to manage negative emotions. To our knowledge, DBT has never been evaluated with PLWHA for suicide prevention, representing a critical gap in the science. Innovative solutions to mental health care, such as AI-powered chatbots, have also not been studied extensively in older PLWHA. The objective of this R34 Intervention Development and Pilot grant proposal is to adapt DBT Skills and Coaching for suicide prevention in older PLWHA to be delivered via brief videos and Angel. Angel is an artificial intelligence (AI)-powered conversational agent developed by CareAngel, a company that leverages cutting edge AI and natural language processing technology. We plan to adapt existing evidence-based DBT skills videos and add coaching support from Angel in a 6-week intervention with older PLWHA. This proposal has the following specific aims: Phase 1: Adaptation: Aim 1) To adapt DBT Skills and Coaching for PLWHA seeking care in infectious disease practices to be delivered via Angel. Phase 2: Pilot feasibility and preliminary efficacy RCT with n=50 PLWHA: Aim 2) To establish the feasibility, acceptability, and usability of the adapted DBT Skills and Coaching intervention through key metrics of use, survey item responses, and qualitative interviews surrounding barriers and facilitators to use with patients enrolled in the intervention. Aim 3) To explore the preliminary efficacy, relative to usual care control, of DBT skills videos and Angel-delivered DBT Coaching on the target mechanism, self- efficacy to manage negative emotions, as well as exploratory clinical outcomes including suicide outcomes, mood, and medical treatment adherence. This proposed work is of high scientific and public health significance for numerous reasons: 1.) older PLWHA have elevated suicide risk. 2.) there is a paucity of suicide prevention intervention research for PLWHA. 3.) while Dialectical Behavioral Therapy has been studied in over 30 randomized controlled trials with demonstrated efficacy to improve psychological outcomes in diverse populations, to our knowledge, this intervention has never been piloted with older PLWHA. 4.) Innovative solutions to mental health care, such as AI-powered chatbots + videos, have not been studied in older PLWHA, thus this proposal represents an exciting opportunity to fill critical gaps in the literature on potentially scalable, implementable, and effective approaches to suicide prevention in PLWHA.