Pilot Testing an Artificial Intelligence Algorithm That Selects Homeless Youth Peer Leaders Who Promote HIV Testing
📝 Abstract
Objective. To pilot test an artificial intelligence (AI) algorithm that selects peer change agents (PCA) to disseminate HIV testing messaging in a population of homeless youth. Methods. We recruited and assessed 62 youth at baseline, 1 month (n = 48), and 3 months (n = 38). A Facebook app collected preliminary social network data. Eleven PCAs selected by AI attended a 1-day training and 7 weekly booster sessions. Mixed-effects models with random effects were used to assess change over time. Results. Significant change over time was observed in past 6-month HIV testing (57.9%, 82.4%, 76.3%; p < .05) but not condom use (63.9%, 65.7%, 65.8%). Most youth reported speaking to a PCA about HIV prevention (72.0% at 1 month, 61.5% at 3 months). Conclusions. AI is a promising avenue for implementing PCA models for homeless youth. Increasing rates of regular HIV testing is critical to HIV prevention and linking homeless youth to treatment.
💡 Analysis
Objective. To pilot test an artificial intelligence (AI) algorithm that selects peer change agents (PCA) to disseminate HIV testing messaging in a population of homeless youth. Methods. We recruited and assessed 62 youth at baseline, 1 month (n = 48), and 3 months (n = 38). A Facebook app collected preliminary social network data. Eleven PCAs selected by AI attended a 1-day training and 7 weekly booster sessions. Mixed-effects models with random effects were used to assess change over time. Results. Significant change over time was observed in past 6-month HIV testing (57.9%, 82.4%, 76.3%; p < .05) but not condom use (63.9%, 65.7%, 65.8%). Most youth reported speaking to a PCA about HIV prevention (72.0% at 1 month, 61.5% at 3 months). Conclusions. AI is a promising avenue for implementing PCA models for homeless youth. Increasing rates of regular HIV testing is critical to HIV prevention and linking homeless youth to treatment.
📄 Content
1
Pilot Testing an Artificial Intelligence Algorithm That Selects Homeless Youth Peer Leaders Who Promote HIV Testing
Eric Rice, PhD a Robin Petering, MSW a Jaih Craddock, MSW a Amanda Yoshioka-Maxwell, MSW a Amulya Yadav, MS b Milind Tambe, Phd b
Unpublished manuscript: August 18, 2016
a School of Social Work, University of Southern California b Viterbi School of Engineering, University of Southern California
Acknowledgements: Funding for this study was provided by the USC School of Social Work. The authors would like to thank the staff at Safe Place for Youth and the young people who participated in the project.
2
Abstract Objective. To pilot test an artificial intelligence (AI) algorithm that selects peer change agents (PCA) to disseminate HIV testing messaging in a population of homeless youth. Methods. We recruited and assessed 62 youth at baseline, 1 month (n = 48), and 3 months (n = 38). A Facebook app collected preliminary social network data. Eleven PCAs selected by AI attended a 1-day training and 7 weekly booster sessions. Mixed- effects models with random effects were used to assess change over time. Results. Significant change over time was observed in past 6-month HIV testing (57.9%, 82.4%, 76.3%; p < .05) but not condom use (63.9%, 65.7%, 65.8%). Most youth reported speaking to a PCA about HIV prevention (72.0% at 1 month, 61.5% at 3 months). Conclusions. AI is a promising avenue for implementing PCA models for homeless youth. Increasing rates of regular HIV testing is critical to HIV prevention and linking homeless youth to treatment.
3
Implications and Contribution Homeless youth are in great need of linkage to HIV testing and treatment. Artificial intelligence can be used to augment intervention delivery of peer-led dissemination models. A pilot test with a pre-test post-test design resulted in a nearly 20% increase in the number of youth reporting recent HIV testing.
4
Despite the need for HIV prevention for homeless youth (HY), few evidence- based interventions exist for HY.1 Given the important role peers play in the HIV risk and protective behaviors of HY,2,3 it has been suggested that peer change agent (PCA) models for HIV prevention be developed for HY.1-3 PCA models have been effective for the prevention of HIV in many contexts,4 but there have been some notable failures5 that may be due to how the PCAs were selected to participate in the intervention.6-8 Change agents can often be as important that the messages they convey. Rarely have network methods that select PCAs based on structural position been attempted.6-8 Selecting PCAs based on structural position requires: (a) the ability to “map” the network space of the target population and (b) a viable structural solution. Prior methods of collecting whole networks of homeless youth accessing drop-in centers required resources prohibitive to future community-based implementation.4 Thus, an integrative Facebook app was developed to collect this information. Computer scientist partners developed an artificial intelligence (AI) algorithm to select PCAs that outperforms other structural network PCA selection rules suggested by Schneider,6 such as degree or betweenness centrality.7-8 This paper presents results of a pilot study of an AI-enhanced PCA prevention program for homeless youth. In accordance with the field’s push to engage underserved populations in the HIV continuum of care, PCA training and peer messaging focused on increasing regular HIV testing (every 3 to 6 months). METHODS Recruitment 5
All study procedures were approved by the [blinded for review] institutional review board. Sixty-two youth (aged 16–24) seeking drop-in homelessness support services (e.g., food, clothing, case management, mobile HIV testing site) in Los Angeles were recruited into the study. All youth receiving services were eligible to participate and were informed of the study as they entered the drop-in center. Participants were required to have a Facebook profile, although there were no requirements regarding how often they use it and if a participant did not have a Facebook account they could create one (n = 5). Assessments Participants completed a computer-based self-administered survey at baseline (n = 62), 1 month (n = 48, 77.4%), and 3 months (n = 38, 61.3%) and received $20, $25, and $30 for each respective assessment. Network Data A Facebook app collected network data regarding which participants were connected to one another, i.e., friends. No information about individuals who were not study participants was collected by the app, which did not appear on their Facebook profiles in any way. These data were augmented by field observations collected by the research team during the 2 weeks of recruitment, based on which participants regularly interacted with one another. AI-Based Peer Selection Papers detailing the development and com
This content is AI-processed based on ArXiv data.