Stepped-wedge cluster trial
Can randomized rollout timing evaluate an intervention introduced across clusters?
When to use this project
Randomize clusters to intervention transition sequences, with repeated observations over calendar periods, when a staged rollout is scientifically and operationally justified.
How it differs
Clusters transition from control to intervention at different times. Unlike a parallel cluster trial, intervention exposure is tied to calendar time, requiring explicit separation of time trends and intervention effects.
Decisions before you start
- Choose closed cohort, open cohort or repeated cross-sectional sampling and define period denominators.
- Prespecify transition periods, exposure-duration effects and calendar-time adjustment.
- Plan correlation across and within cluster-periods, repeated people when present and sensitivity to time-model assumptions.
How the flows work together
Each flow below links to its diagram, steps and tables. Flow order organizes the project; it does not imply that all work is strictly sequential. Data assessment, analysis planning and protocol refinement may inform one another.
1. Stepped-wedge trial overview
Overview · Question through phased rollout evaluation
Use this overview to state the purpose and link to detailed decisions. Keep a versioned protocol and decision register, and add evidence references for the rationale.
2. Cluster and participant selection
Selection · Define clusters and observation populations
Closed cohorts, open cohorts and repeated cross-sections have different observation processes. Keep unique people, repeated observations and cluster-period denominators distinct.
3. Rollout and transition schedule
Conduct · Assign sequences periods and transitions
Randomize transition sequences and record assigned and actual transition dates. A delayed implementation is a deviation to analyze, not a reason to silently rewrite the randomized sequence.
4. Endpoints and estimands
Design · Define effects and exposure duration
The intervention effect may depend on time since adoption. Define the target weighting over periods and exposure durations and the treatment of transition or learning periods.
5. Design scenarios and power
Statistics · Account for time and clustering
Simulate or calculate using the actual rollout schedule and plausible correlation within and across periods. A simple parallel-cluster design effect does not describe stepped-wedge information.
6. Period data and follow-up
Data · Track repeated observations and missingness
Preserve cluster-period keys and repeated-person identifiers when relevant. Record calendar time and time since transition separately; changing participant composition can affect interpretation.
7. Time-adjusted analysis plan
Statistics · Separate intervention effects from trends
Calendar time and intervention exposure are associated by design. Model time explicitly, justify dependence assumptions and assess sensitivity to exposure-duration and time-trend specifications.
Equations and statistical interpretation
These explanations describe notation and assumptions. Flow does not calculate these values.
Illustrative continuous-outcome model
Yᵢⱼₜ is the outcome for person i in cluster j and period t; γₜ represents calendar-period effects, Aⱼₜ intervention status and uⱼ a cluster effect.
This illustration assumes a constant intervention effect and a particular dependence structure. Repeated participants, changing cluster effects or exposure-duration effects need extensions; binary and count outcomes need appropriate models.
Published example
Trace the randomized transition sequences and observation periods. Consider why a simple before–after comparison would discard important features of the design.
This publication illustrates a study design, not an endorsement of every method or a reproduction of its results. Follow its protocol, supplement and update history when studying the example.
Hypothetical scenario
A hospital network will introduce a discharge coordination service in stages and can randomize the rollout sequence.
- Stepped-wedge trial overview: Define why staged introduction is suitable and register the randomized rollout protocol.
- Cluster and participant selection: Specify eligible hospitals and patients in each period and retain changing denominators.
- Rollout and transition schedule: Randomize hospitals to transition sequences and document implementation delays.
- Endpoints and estimands: Define whether the target effect is immediate, delayed or changes with time since adoption.
- Design scenarios and power: Evaluate power using the rollout schedule, time trends and multilevel correlation.
- Period data and follow-up: Build cluster-period records and link repeated patients when applicable.
- Time-adjusted analysis plan: Estimate the planned intervention effect with calendar-time adjustment and examine alternative exposure-duration assumptions.
Deliver a populated protocol, decision log, data specification, and planned outputs. Enter numerical findings only after analysis and verification. Keep decisions made before seeing results distinguishable from later amendments.
Sources
- CONSORT 2025: reporting randomized trials
- SPIRIT 2025: randomized trial protocols
- CONSORT extension for stepped-wedge cluster trials (2018)
- Hemming et al. (2015): stepped-wedge design, analysis and reporting
- ICH E9(R1): estimands and sensitivity analysis (2019)
- WHO: clinical trials and prospective registration
- ISRCTN: what the clinical study registry records
- ClinicalTrials.gov: about the registry
Source check: 2026-09-17. Record the version of each reporting or appraisal tool used in your actual protocol.
Compare all Clinical trials project types · Open editor in a new tab