Using Venue-Based Social Network Analysis to Improve Timely and Targeted Community-Based Intervention Delivery

Background and Purpose: Social work research often relies on intervention delivery in community settings; yet many researchers wait until an intervention is ready for testing to determine ideal community sites in which the intervention can be most impactful. This approach is contrary to a translational science model in which interventions move from development to testing as rapidly as possible. The present paper illustrates a methodology for identifying key community venues for intervention implementation using venue-based social network analyses that stem from the addition of a simple name-generator item in formative survey research. This approach adds little respondent burden yet can yield great benefits in informing targeted intervention delivery.

Methods: To illustrate this process, we used data from a longitudinal study of young men who have sex with men (YMSM) aimed to understand HIV risk behavior in this population to inform tailored prevention efforts. Participants were asked to nominate up to three favorite venues in which to socialize with other YMSM. Two-mode social network analysis was used to create a person-by-venue matrix with unique identifiers for study participants on the row (N=484) and nominated venues on the column (N=110). Matrix algebra in UCINet was used to generate a venue-by-venue matrix (110x110), where number of participants who nominated each venue were contained on the diagonal. This venue matrix was graphed using the social network visualization program Netdraw; increasing thresholds of person-sharing between venues was employed to identify the most popular venues. Aggregate HIV risk behavior information was calculated within venues to identify those where intervention delivery was most needed.

Results: Over 92% of study participants named at least one favorite place to socialize and nearly 62% nominated three venues. Visual inspection of the venue-based network, demonstrated a pronounced core-periphery structure, Specifically, several highly connected venues were located at the “core” of the network; these venues connected to a large number of peripheral venues that had fewer connections to each other than to the venues in the “core”. As the person-sharing threshold was increased progressively fewer venues remained connected. By the time the person-sharing threshold was set to 30 or more, only 6 venues remained in the network (5.4% of all venues nominated) yet together these venues captured 98% of all participants in the study. Venues were ranked in order of priority based on HIV-risk behavior occurring among men who nominated each venue.

Conclusions and Implications: The addition of a single item to the original study survey enabled us to quickly identify the most commonly accessed venues by YMSM and the venues with the highest rates of HIV-risk behaviors. In a climate of limited funding for social work research and intervention implementation, this approach represents a valuable tool for social work researchers who wish to quickly identify venues in which intervention programs may be most impactful. This methodology can be easily extended to other community-based research studies that seek to ultimately inform targeted intervention with diverse populations.