Why clinical trials are terminated
Bibliographic record
Abstract
Abstract Background: Evidence-based clinical practice relies on unbiased reporting of negative results. Meta-analysis of drug safety and efficacy across many clinical trials is difficult given the unconstrained nature of reasons that are provided to ClinicalTrials.gov to explain clinical trial terminations. Methods and Findings: We scanned all trials in ClinicalTrials.gov marked with the “terminated” status (N=3122), meaning the trial had been stopped before the scheduled end date. Under the current reporting framework, any number of reasons may be given for termination, and these need not conform to a controlled vocabulary. Here we develop a controlled vocabulary for trial termination, and map each terminated trial to as many as three vocabulary terms. Mapping to this “ontology of termination” allows further analysis and conclusions. First, we identify the subset of terminated trials that ended citing safety concerns (6.2%) or failure to establish efficacy (10.8%), and were further able to stratify these rates across trials of different phases. Second, we examine termination reasons where a stricter data model could have preserved more evidentiary value, either because the data model was misused (7.6%) or because the given reason left unclear whether the decision to terminate was based on analysis of the data (74.9%, with 20.4% mentioning a decision-maker that may have had access to the data). Third, we show that imposing a controlled vocabulary of reasons for termination would avoid ambiguity and improve the evidentiary value of clinical trials. Conclusions: We encourage wider use of an “ontology of termination” and propose four questions that should be posed on trial termination. These simple steps would promote transparency and enable ready access to negative trial results for meta-analysis.
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 | 0.594 | 0.832 |
| Meta-epidemiology (narrow) | 0.001 | 0.002 |
| Meta-epidemiology (broad) | 0.004 | 0.007 |
| Bibliometrics | 0.011 | 0.010 |
| Science and technology studies | 0.004 | 0.020 |
| Scholarly communication | 0.017 | 0.021 |
| Open science | 0.006 | 0.009 |
| Research integrity | 0.012 | 0.014 |
| Insufficient payload (model declined to judge) | 0.007 | 0.003 |
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; the direct Gemma label and the distilled Codex classifier 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".