Skip to main content
Publication

Evaluating Complex Health Interventions: A Guide to Rigorous Research Designs

This guide provides a framework to guide decision-making around appropriate designs to evaluate public health and other service interventions.

Evaluating Complex Health Interventions

Evaluating Complex Health Interventions: A Guide to Rigorous Research Designs

This guide provides a range of approaches that could be used to enhance the rigor of evaluations thus improving the quality of the evidence upon which decisions are made and ultimately improving the public’s health.

Download Publication

As the number, variety and complexity of innovations increase and the need to understand which ones are working, for whom, and under what circumstances grows across sectors, it is clear there is no single “correct” evaluation design. This guide is aimed at program managers and other stakeholders implementing innovations in public health and community settings who are involved in evaluation but may not themselves be evaluators.

The evaluation designs presented in this guide represent a mix of experimental, quasi-experimental and observational designs—randomized controlled trial, cluster randomized stepped wedge design, interrupted time series design, controlled before and after design, regression discontinuity design and natural experiment. Each design includes a general description with a diagram to illustrate the design; two examples from the peer-reviewed literature of how the design was used to evaluate a specific health or social service; key strengths and weaknesses of the study design; timeline and budget considerations; as well as policy implications and considerations for future use. The guide aims to help the reader make informed decisions by providing information on the various tradeoffs involved in the selection of an evaluation design. A flow chart to inform the selection of evaluation designs is featured in the guide. Resources for more detailed information on a broader range of designs are also listed.