SOLVE-It: An Interactive Nationwide Web-Deliverable Game for Changing Risky Sexual Behavior Among YMSM

Theoretical Background and research questions/hypothesis:  Socially optimized learning in a virtual environment (SOLVE), integrates diverse theoretical perspectives.  It places young men who have sex with men (YMSM) in an interactive virtual environment designed to simulate the emotional, interpersonal, and contextual narrative of an actual sexual encounter while challenging and changing YMSM’s more automatic patterns of risky responses.  SOLVE-IT uses SOLVE with intelligent agent/gaming technologies. YMSM (18-24), especially YMSM of color, are at especially high risk for contracting HIV, yet relatively few interventions are designed to reach them and take into account the affective and automatic factors (e.g., sexual arousal, shame/stigma) that may precipitate risk for them. A neuroscience-based model of decision-making (Bechara et al., 1994) suggests that affective/automatic as well as more traditional cognitive/deliberative factors affect risk-taking.  An NIAID-funded DVD interactive video intervention, adapting this model, produced effective risk reduction over 3-months for African American, Latino, and Caucasian YMSM. Limitations of that work (e.g., a limited number of risk challenges) are addressed in a readily updateable NIMH funded game with 3D animated intelligent agents. The game is designed for national web delivery. RQ₁: What are the key venues and contexts of risk for YMSM nationwide? RQ₂: What are the most promising modes of intervention that we might incorporate within our game, when, for whom? RQ₃: Do initial pre-post summative findings suggest that the intervention is changing what it should change?

Methods: To construct an effective game (i.e., one that will lower risk behaviors) we used survey research, literature reviews, and formative pilot research. To assess initial manipulation checks, we are conducting baseline (pre-post) assessments with over 100 YMSM from the targeted population randomly assigned to an experimental (game) vs. wait-list control condition.

Results: (1) Key venues for meeting MSM (e.g., house parties, bars/clubs, internet) were identified in pilot work conducted nationally. Specific contexts of risk (or that lead to increased risk) for YMSM (e.g., an attractive man who is interested in you who wants unprotected sex) were also identified.  These formed the basis of extensive alternative narratives within which to present MSM with risk challenges; (2) potential interventions that had empirical support for changing YMSM’s risky behavior were also identified (e.g., self-regulatory and sex positive interventions; social cognitive intervention components, etc.) and integrated into the game.  We will illustrate how social science theory/research (e.g., on narratives/challenges and intervention components) and virtual agent research for building the game are integrated. We will then report on preliminary findings from baseline results assessing whether the intervention initially changed key cognitive and affective variables that theory suggests should change risky sexual behavior.

Conclusions: SOLVE-IT suggests how to develop, integrate, and synthesize diverse research findings within a game (with intelligent agents) to create theory-based interventions for specific populations and behavior change efforts that can be delivered and tested nationwide over the web.

Implications for research and/or practice: Games provide potentially effective methods to go from theory to research test-bed to rapid dissemination over the web to targeted at-risk populations.