Diffusion of a Peer-Led Suicide Prevention Program through Social Networks

Introduction: Peer leader programs draw on a diffusion of innovations model emphasizing how ‘key opinion leaders’ persuade close affiliates to adopt new behaviors. This model suggests that peer leader program diffusion will be maximized when more members of a target population have close ties to peer leaders. In this paper, we examined the reach of peer leaders trained in a suicide prevention curriculum (Sources of Strength) to other students in their schools. We then tested, among the non-peer leaders, how closeness to Peer Leaders predicted intervention exposure after a school year.

Method: Students in 17 high schools enrolled in an RCT of Sources of Strength (RO1MH091452) completed baseline assessments (M=82% participation) that included naming up to 7 closest friends at school. Peer Leaders were selected, trained and conducted suicide prevention activities over 4-5 months. School sizes ranged from 60-1,132 students (M=483) and proportion of Peer Leaders ranged from 1% - 31% (M=9%) of the school population. Whereas 12.5% of students were Peer Leaders in schools with < 500 students, only 3.2% were Peer Leaders in schools > 500 students. Social networks were constructed using baseline nominations to calculate how many steps each non-Peer Leader was to a Peer Leader. Students completed a follow-up survey 6 months later asking about two types of exposure to suicide prevention programming, direct messaging from a peer and participation in an activity.

Results: An average of 38.3% of students (SD=23.4%) were directly connected to a Peer Leader, 75.0% (SD=20.5%) were connected within two steps, and 92.8% (SD=9.8%) were within three steps. Students were more removed from Peer Leaders in larger schools (b=-0.17, p<.001) and in schools that trained fewer Peer Leaders (b=-0.02, p<.001). The proportion of students within two ties to Peer Leaders was 88% for schools < 500 students and 56% for schools >500 students. Students closer to Peer Leaders were most likely to have received direct suicide prevention messaging (OR=1.40) and to have participated in a prevention activity such as naming a trusted adult (OR=1.35). Intervention exposure was highest for those within one step of a Peer Leader, remained high for those two steps away, and dropped and leveled off for those three steps or more removed.

Conclusion: These preliminary findings support a diffusion of innovations model and highlight the importance of selecting Peer Leaders connected to other students within one or two steps. In large schools, maximizing Sources of Strength diffusion will require greater numbers of Peer Leaders than are currently being selected. We will also examine school-level network characteristics that enhance diffusion above and beyond ties to peer leaders.