Improving Bayesian adaptive clinical trials with covariate information
Bibliographic record
Abstract
Bayesian adaptive clinical trials allow for predetermined changes to the trial design based on evidence provided by the accumulating data. Early phase trials assess the safety and efficacy profiles of first-in-human doses of an experimental intervention, whereas late phase trials seek to confirm the efficacy of an experimental intervention. It is common in both phases of trials to collect information on many covariates, though less common to utilize this information within the trials. Through three manuscripts, this thesis investigates the benefit of using this additional information in both early phase dose-finding trials for combination therapies and late phase confirmatory trials. In the first manuscript, I focus on late phase confirmatory trials, where I propose adjusting for covariates known to be at least moderately associated with the outcome. Adjusted and unadjusted analyses are described within the context of collapsibility and non-collapsibility of commonly used estimands in clinical trials, and a tutorial for obtaining marginal treatment effect estimates from adjusted analyses is provided. It is shown that covariate adjustment increases the power and probability of stopping trials early, and decreases expected sample sizes, demonstrating covariate adjustment leads to trials which stop earlier and more often. This is followed by a second manuscript which proposes using Bayesian optimization methods for early phase personalized dose-finding trials with combination therapies, where individual patient characteristics are used to recommend patient-specific optimal dose combinations. Focus is placed on a minimal toxicity setting where formal dose escalation rules are not required. The approach demonstrates the feasibility of a personalized approach for dose-finding trials despite the limited sample sizes which are common in early phase trials. The third and final manuscript generalizes this approach to the setting of higher-grade toxicities, where a formal dose escalation scheme is introduced and where toxicities are incorporated into the personalized dose-finding strategy. In summary, this thesis showcases the benefit of utilizing additional covariate information in the context of Bayesian adaptive clinical trials and seeks to promote more widespread use of the methods proposed herein
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.087 | 0.284 |
| Meta-epidemiology (narrow) | 0.002 | 0.003 |
| Meta-epidemiology (broad) | 0.006 | 0.003 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.003 | 0.004 |
| Open science | 0.004 | 0.004 |
| Research integrity | 0.004 | 0.008 |
| Insufficient payload (model declined to judge) | 0.008 | 0.002 |
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; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
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".