Evaluation of systems-oriented public health interventions: alternative research designs

Abstract

The need to provide sound evidence of the costs and benefits of real-world public health interventions has driven advances in the development and analysis of designs other than the controlled trial in which individuals are randomized to an experimental condition. Attention to methodological quality is of critical importance to ensure that any evaluation can accurately answer three fundamental questions: (a) Has a change occurred, (b) did the change occur as a result of the intervention, and (c) is the degree of change significant? A range of alternatives to the individual randomized controlled trial (RCT) can be used for evaluating such interventions, including the cluster RCT, stepped wedge design, interrupted time series, multiple baseline, and controlled prepost designs. The key features and complexities associated with each of these designs are explored.