Assessing program fidelity in substance abuse health services research.

This paper addresses how treatment fidelity and related constructs (e.g. program implementation) can be assessed in alcohol, drug abuse and mental health services research. First, it introduces definitions of fidelity and related concepts, and then describes various concepts and tools from program evaluation that have proven useful for assessing fidelity. Next, several of these are illustrated in detail through a case study of a multisite fidelity assessment in substance abuse services research: the process evaluation of the NIAAA Homeless Cooperative Agreement Program. This evaluation included analysis of implementation at the program- and participant-level, the development of scales from the individual services data to estimate intervention strength, fidelity, and "leakage" (i.e. the degree to which services intended exclusively for intervention groups were inadvertently delivered to comparison groups) and the methods with which these data were used to assess whether programs were implemented as planned.