Characterizing Grantee Prevention Infrastructure in a National Cross-Site Evaluation

Introduction: Since 2004, the Substance Abuse and Mental Health Services Administration has funded Strategic Prevention Framework State Incentive Grants (SPF SIG) to prevent the onset/progression of substance abuse, reduce substance abuse-related problems, and build prevention capacity/infrastructure. SPF SIG grantees are a heterogeneous group of states, tribal organizations, and jurisdictions with varying levels of prevention infrastructure. These differences have the potential to dampen findings related to program impact and moderate associations between process and outcome variables. We examined whether SPF SIG Cohort III-V grantees (N=50) can be grouped into clusters based on patterns of baseline infrastructure and whether these clusters differ on implementation characteristics.

Method: Grantees completed the SPF SIG Grantee-level Instrument (GLI) near the beginning of their grant. Six infrastructure scales included: organizational structure, workforce development, strategic planning, evaluation/monitoring, data systems, and evidence-based programs/policies/practices. We conducted a two-stage cluster analysis of these scales: hierarchical cluster analysis to determine the optimal number of clusters and k-means clustering to assign grantees to clusters (Burns & Burns, 2008). We used ANOVAs to compare clusters on implementation characteristics reported in the GLI.

Results: A three-cluster solution indicated that grantees can be placed into one of three baseline infrastructure categories: low (n=18), mixed (n=18), and high (n=14). The low and high clusters are characterized by grantees with uniformly low and high levels of infrastructure, whereas the mixed cluster is characterized by more variability across scales. Grantees in the low cluster reported significantly fewer resources than grantees in other clusters (ps < .001). High cluster grantees completed significantly more capacity-building activities (ps < .05), especially those relating to evaluation, sustainability, and improved/standardized service delivery. Although the number of barriers did not differ between low and high clusters, availability of staff time was a more common barrier for high cluster grantees and lack of prevention infrastructure was a more common barrier for low cluster grantees. 

Conclusions: Grantees with the highest levels of prevention infrastructure implemented more, and more resource-intensive types of capacity building activities.  This suggests potential accrual of advantage for these grantees over time. Grantees with varying levels of infrastructure face similar numbers of barriers but the types of barriers faced are qualitatively different. These findings support the validity of using infrastructure clusters in future moderation analyses of SPF SIG outcomes.