Who benefits? Uncovering hidden heterogeneity of treatment effects in adaptive trials using Bayesian methods: a systematic review
Bibliographic record
Abstract
BACKGROUND: Adaptive clinical trials increasingly aim to detect heterogeneity of treatment effect (HTE) to guide personalized care. However, most adaptive designs rely on predefined subgroups and are limited in their ability to uncover unknown or complex sources of HTE. Bayesian statistical methods offer a flexible alternative, enabling real-time learning and adaptation within trials. This review evaluates Bayesian methods used to detect hidden HTE in adaptive clinical trials, with attention to their methodological innovations, operating characteristics, and consideration of equity and inclusion in trial design. METHODS: We conducted a systematic search of MEDLINE, Embase, and other databases to identify original studies that developed Bayesian methods for detecting unknown HTE within adaptive clinical trial designs. Eligible studies were reviewed and synthesized based on design features, statistical methodology, operating characteristics, reproducibility, and whether equity-related factors were explicitly considered. Equity considerations included whether studies incorporated variables related to underrepresented populations-such as age, sex, race/ethnicity, or geography-examined intersectional subgroup effects, or explicitly framed their methods as tools to address health disparities. RESULTS: Of 2826 screened records, seven studies met inclusion criteria. Bayesian methods included random partition models, spatial models, logistic regression with dimension reduction, adaptive randomization using machine learning classifiers, and adaptive enrichment or platform designs incorporating model averaging or latent subgroup estimation. In simulation studies, these methods often showed improvements in subgroup detection, efficiency, or power relative to non-Bayesian comparators. None were tested using real-world trial data. Reproducibility was limited overall, with analytic code only available for the three most recent studies. Notably, none explicitly framed their methods as tools to address inequities in treatment outcomes across population subgroups. CONCLUSIONS: The small number of simulation-based studies illustrates preliminary but promising directions for applying Bayesian methods to detect HTE in adaptive clinical trials. While these approaches demonstrate potential to enhance trial adaptability, scalability, and inclusiveness, current evidence remains limited and largely conceptual. Incorporating an equity lens into future methodological development, alongside greater emphasis on empirical validation and open science practices, will be essential to determine their practical value in advancing equitable clinical research.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.105 | 0.693 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.014 | 0.001 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; both teacher heads agree on what is shown here.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".