Using Artificial Intelligence to Deliver an HIV Testing Promotion App to Homeless Youth in LA

Research Question: Can disparities in HIV testing and linkage to care among homeless youth (HY) be eliminated via a peer-led, Facebook-delivered HIV testing promotion intervention?

Specific Aims: Evaluate: 1) if the pro-social peer leader method (leaders selected based on positive behaviors), or 2) the PSINET leader selection, an artificial intelligence, promote more HIV testing than the general health control arm; 3) whether the PSINET leader selection HIV testing arm promotes more HIV testing than pro-social peer leader selection.

Background: In LAC, nightly more than 4000 youth aged 13 to 24 spend the night on the streets; 66% of are racial/ethnic minorities and 40% are LGBT. Many HY are unaware of their HIV status; many racial/ethnic minority HY do not get tested for HIV regularly. A Peer Change Agent (PCA) model may be particularly effective for HY, given their distrust of adults and their important interconnections with one another.

Approach: CBPR will be used to refine the intervention protocol. A RCT of 900 HY with three study arms will occur: (1) pro-social peers promoting HIV testing, (2) structurally most efficient peers promoting HIV testing, selected via PSINET, and (3) general health promotion control.

Expected Results: HIV testing rates will increase in the two HIV testing promotion arms.