Post-hoc power of clinical trials supporting anticancer drug approval by FDA.
Bibliographic record
Abstract
2018 Background: Regulatory approval of drugs is based typically on randomized control trials (RCTs) observing statistically significant superiority of an experimental agent over a prior standard. Statistical significance can result from large effect size and/or over-sampling (as a result of large sample size or long follow-up). Here we explore the source(s) of statistically significant results in trials supporting anti-cancer drug approval by the FDA. Methods: We searched Drugs@FDA to identify anti-cancer drug approvals for solid tumors (excluding lymphoma) from 2015-2019. We retrieved corresponding manuscripts and associated appendices and extracted data on study characteristics, statistical plan, primary outcomes and accrual and follow-up times. Post-hoc power was calculated based on observed results and was compared to expected effect size and power in the statistical plan. We explored associations with higher than expected power resulting from over-sampling using binary logistic regression. Results: We identified 75 unique drug-approvals reporting 94 endpoints. The most common tumour types were lung, breast, melanoma, and renal cell carcinoma. The most common endpoints were progression free survival and overall survival (OS). In 74 endpoints (79%), observed power was greater than expected power. The magnitude of higher than expected power ranged from 0.1 to > 20%. Of these, 59 (80%) had an effect size greater than predicted in the statistical plan. In 44/74 over-powered endpoints (60%), post-trial power was 100%. When post-hoc power was calculated based on expected effect size rather than observed effect size, 50 endpoints (85%) remained over-powered. Higher than expected power resulting from over-sampling was associated with OS compared to other endpoints (OR 3.03), with targeted agents compared to immunotherapy (OR 1.63) and inversely associated with year of approval (OR 0.57). Conclusions: Most cancer drug approvals result from statistically significant studies which are over-powered due to greater than anticipated effect size. Approximately 1 in 5 studies are over-powered likely due to over-sampling. In this setting, benefit observed in RCTs may not translate to the real-world setting.
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 | 0.045 | 0.068 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.004 | 0.002 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.001 | 0.003 |
| Insufficient payload (model declined to judge) | 0.001 | 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".