Adaptive randomized trial
Can prespecified interim information change a randomized trial while preserving valid inference?
When to use this project
Use planned adaptations such as sample-size reassessment, arm dropping or allocation changes when their operating characteristics and operational safeguards are established in advance.
How it differs
An adaptive design prespecifies decisions based on accumulating information. It is not an unrestricted protocol amendment; it can overlap with a platform or dose-finding design.
Decisions before you start
- Specify information times, available data, decision boundaries and authorized actions.
- Evaluate error control, power, estimation bias and interval coverage under the complete adaptation process.
- Protect interim information and use final inference compatible with selection, stopping and allocation changes.
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. Adaptive trial overview
Overview · Question through adaptive trial reporting
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. Recruitment and allocation
Selection · Track enrollment and changing assignments
Not every adaptive design changes allocation probabilities. If allocation can change, prespecify how, when and for whom, and retain the probability and rule version for each epoch.
3. Endpoints and estimands
Design · Prespecify objectives outcomes and effects
An estimand specifies the population, treatment conditions, outcome variable, handling of intercurrent events and population-level summary. An intercurrent event changes interpretation or existence of an outcome; it is not synonymous with a missing measurement.
4. Adaptation and interim decisions
Statistics · Specify triggers actions and boundaries
Prespecify information times and authorized actions. Distinguish planned adaptations from unplanned amendments and explain deviations. Interim confidentiality protects ongoing conduct and interpretation.
5. Design simulations and sample size
Statistics · Evaluate power error and feasibility
Simulate the complete stopping and adaptation process under null, alternative and operational scenarios. Report Monte Carlo uncertainty along with error rates, power, bias and coverage.
6. Follow-up safety and data
Conduct · Protect monitoring integrity and participants
Arm closure or early stopping does not automatically end participant follow-up. Keep interim access, data cutoffs, safety review and final data release responsibilities explicit.
7. Adaptive final analysis plan
Statistics · Account for adaptations in inference
Selection and stopping can affect estimation as well as testing. Prespecify compatible estimators and intervals, and report the rules actually applied rather than analyzing as though the design were fixed.
Equations and statistical interpretation
These explanations describe notation and assumptions. Flow does not calculate these values.
Monte Carlo standard error of a simulated proportion
B is the number of independent simulated trial replicates and q̂ the fraction with a specified outcome, such as rejecting a null hypothesis.
This describes simulation precision, not uncertainty in a clinical effect. Include the complete adaptive procedure in each replicate; a small Monte Carlo error does not establish that the simulated assumptions are realistic.
Published example
Park et al. (2016), Adaptive Randomization of Neratinib in Early Breast Cancer
Inspect the I-SPY 2 adaptive allocation and decision framework. This is also a platform setting, illustrating that these design categories can overlap.
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 team wants to compare interventions while allowing a prespecified interim decision to stop a poorly performing arm.
- Adaptive trial overview: Define the adaptation objective and estimands before simulation and prospective registration.
- Recruitment and allocation: Specify initial allocation and any permitted epoch-specific changes.
- Endpoints and estimands: Define endpoints and intercurrent-event strategies independently of interim convenience.
- Adaptation and interim decisions: Prespecify information triggers, decision rules and independent review authority.
- Design simulations and sample size: Simulate null, alternative, delayed-outcome and operational scenarios with Monte Carlo uncertainty.
- Follow-up safety and data: Maintain follow-up and control access to interim data and decision cutoffs.
- Adaptive final analysis plan: Use adaptation-aware inference and report implemented and unimplemented rules with reasons.
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
- Adaptive designs CONSORT Extension (ACE, 2020)
- FDA: Adaptive Designs for Clinical Trials of Drugs and Biologics (2019)
- 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