Linking interventions to outcomes: Designing a multisite data collection protocol

Issues: Cross-site studies in which sites differ with respect to participants and intervention types present specific data collection challenges. One challenge in garnering the benefits of cross-site studies is measuring the characteristics of participants, both before and after an intervention period, and the characteristics of interventions through a study design that allows the evaluator to statistically control differences among sites (Sinacore and Turpin, 1991). This requires common protocols for measuring participant, organization and service attributes that capture the important characteristics of participants and interventions. Description: This paper describes the steps in designing the multisite data collection protocol for the Substance Abuse and HIV Prevention Program funded by SAMHSA/CSAP as part of the broader Minority AIDS Initiative. The National Cross-Site Team, housed within CSAP's Data Analysis Coordination and Consolidation Center revised a pre-existing data collection system to allow for linkage between characteristics of interventions and intervention outcomes. This provided the evaluation team with the means for constructing a multilevel data structure and conducting hierarchical analyses. Lessons Learned/Recommendations: This paper highlights the challenges encountered during implementation and how these challenges have been addressed.