The trustworthiness and impact of trial preprints for COVID-19 decision-making: A methodological study
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
Purpose: To assess the trustworthiness and impact of preprint trial reports during the COVID-19 pandemic. Data sources: WHO COVID-19 database and the L-OVE COVID-19 platform by the Epistemonikos Foundation (up to August 3rd, 2021) Design: We compare the characteristics of COVID-19 trials with and without preprints, estimate time to publication of COVID-19 preprint reports, describe discrepancies in key methods and results between preprint and published trial reports, report the number of retracted preprints and publications, and assess whether including versus excluding preprint reports affects meta-analytic estimates and the certainty of evidence. For the effects of eight therapies on mortality and mechanical ventilation, we performed meta-analyses including preprints and excluding preprints at 1 month, 3 months, and 6 months after the first trial addressing the therapy became available either as a preprint or publication (120 meta-analyses in total). Results: We included 356 trials, 101 of which are only available as preprints, 181 as journal publications, and 74 as preprints first and subsequently published in journals. Half of all preprints remain unpublished at six months and a third at one year. There were few important differences in key methods and results between trial preprints and their subsequent published reports. We identified four retracted trials, three of which were published in peer-reviewed journals. With two exceptions (2/60; 3.3%), point estimates were consistent between meta-analyses including versus excluding preprints as to whether they indicated benefit, no appreciable effect, or harm. There were nine comparisons (9/60; 15%) for which the rating of the certainty of evidence differed when preprints were included versus excluded, for four of these comparisons the certainty of evidence including preprints was higher and for five of these comparisons the certainty of evidence including preprints was lower. Limitations: The generalizability of our results is limited to COVID-19. Preprints that are subsequently published in journals may be the most rigorous and may not represent all trial preprints. Conclusion: We found no compelling evidence that preprints provide less trustworthy results than published papers. We show that preprints remain the only source of findings of many trials for several months, for a length of time that is unacceptable in a health emergency. We show that including preprints may affect the results of meta-analyses and the certainty of evidence. We encourage evidence users to consider data from preprints in contexts in which decisions are being made rapidly and evidence is being produced faster than can be peer-reviewed.
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 | 0.104 | 0.406 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.002 | 0.000 |
| Open science | 0.008 | 0.010 |
| Research integrity | 0.000 | 0.002 |
| 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 it