Verifying Death Reports in the Platelet Inhibition and Patient Outcomes (PLATO) Trial
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
BACKGROUND: Excess vascular deaths in the PLATO trial comparing ticagrelor to clopidogrel have been repeatedly challenged by the Food and Drug Administration (FDA) reviewers and academia. Based on the Freedom of Information Act, BuzzFeed won a court order and shared with us the complete list of reported deaths for the ticagrelor FDA New Drug Application (NDA) 22-433. This dataset was matched against local patient-level records from PLATO sites monitored by the sponsor. STUDY QUESTION: Whether FDA death data in the PLATO trial matched the local site records. STUDY DESIGN: The NDA spreadsheet contains 938 precisely detailed PLATO deaths. We obtained and validated local evidence for 52 deaths among 861 PLATO patients from 14 enrolling sites in 8 countries and matched those with the official NDA dataset submitted to the FDA. MEASURES AND OUTCOMES: Existence, precise time, and primary cause of deaths in PLATO. RESULTS: Discrepant to the NDA document, sites confirmed 2 extra unreported deaths (Poland and Korea) and failed to confirm 4 deaths (Malaysia). Of the remaining 46 deaths, dates were reported correctly for 42 patients, earlier (2 clopidogrel), or later (2 ticagrelor) than the actual occurrence of death. In 12 clopidogrel patients, cause of death was changed to "vascular," whereas 6 NDA ticagrelor "nonvascular" or "unknown" deaths were site-reported as of "vascular" origin. Sudden death was incorrectly reported in 4 clopidogrel patients, but omitted in 4 ticagrelor patients directly affecting the primary efficacy PLATO endpoint. CONCLUSIONS: Many deaths were inaccurately reported in PLATO favoring ticagrelor. The full extent of mortality misreporting is currently unclear, while especially worrisome is a mismatch in identifying primary death cause. Because all PLATO events are kept in the cloud electronic Medidata Rave capture system, securing the database content, examining the dataset changes or/and repeated entries, identifying potential interference origin, and assessing full magnitude of the problem are warranted.
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.
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 | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| 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 it