Establishing a Control (C) Group
A randomized design will typically assert control over independent variables and minimize M-S-H biases by equally distributing, between the E and C group, participants with variable baseline characteristics. Each variable may have an independent or interactive effect, explaining part of an observed impact rate. The key question in establishing a C group is, Does it adequately control for independent predictors of change in the impact and outcome rates? Baseline equivalence of an E and C group needs to be documented: randomization does not always produce equivalent groups at baseline. Ideally, E and C baseline data confirm that they are almost identical for all dependent impact or outcome variables, and independent variables.
An issue, which at times has been a common barrier to C group establishment, is that program staff may not want to withhold the new treatment methods (X1 + X2 + X3 + X ... ) from participants. This may seem to be a serious barrier at first glance; it should not be in practice. Remember: an evaluation is being planned to confirm intervention efficacy or effectiveness. Valid impact data from a sufficient number of completed evaluations supporting the internal validity of an existing program or new program for your population or setting may not be available. Although the standard or minimum program (X1) should be delivered to all participants, the intensity and duration of methods and materials or the frequency of program procedures can be varied to document what is most effective. The C group, the standard program (X1) can be compared to a “best practice” program (X1 + X2 + X3 + X4) among an E group, where X2 is a brief, face-to-face reinforcement, X3 is follow-up telephone counseling, and X4 is systematic family reinforcement.
An evaluation should be designed to answer important practical questions about the feasibility of delivery, impact, cost, and cost-effectiveness of an HP-DP program for a public health agency or primary care organization. How effective is the existing program, and what new, best practices methods could be feasibly applied in all practice settings that might significantly increase the current level of behavior change and improve health? This type of evaluation may be described by some researchers as “Comparative Effectiveness Research” (IOM, 2009).