Power calculation in stepped-wedge cluster randomized trial with reduced intervention sustainability effect

ABSTRACT

The stepped-wedge design for pragmatic clinical trials has received increased attention in health service-related research seeking to evaluate the effect of interventions. Compared with the parallel design, the stepped-wedge design is preferred when there is prior knowledge supporting the effectiveness and harmlessness of the intervention, and/or when practical or financial constraints exist such that the intervention can only be implemented sequentially on a fraction of clusters. In some health service studies, the study period may consist of two parts: an active implementation followed by a sustainability phase, where the intervention effect is possibly reduced. There is a gap in current literature of the stepped-wedge design for cluster randomization trials for dealing with this specific scenario. We aim to provide an analytical formula for power analysis under this situation to aid the stepped-wedge design of an ongoing PREVENT trial.

KEYWORDS:

• Stepped-wedge design
• power analysis
• cluster randomized trials
• pragmatic clinical trials

Acknowledgments

This work was supported by the Department of Veterans Affairs (VA), Health Services Research & Development Service (HSRD) (QUE 15-280). The authors are grateful to the comments from the two anonymous reviewers that greatly help us improve the presentation of this manuscript from an early version.

Additional information

Funding

This work was supported by the U.S. Department of Veterans Affairs, Health Services Research & Development Serivice, Quality Enhancement Research Initiative (QUERI) [QUE 15-280].