Predicting the Probability That Open-Access Clinical Literature Saves Lives
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
Whether open-access (OA) clinical literature directly saves lives is frequently debated, yet empirical documentation is scarce because clinical notes rarely record how evidence was accessed. This study synthesizes high-impact cases of OA-enabled clinical change—most notably the SARS-CoV-2 PCR diagnostic protocol and the RECOVERY dexamethasone findings—and develops an expanded Bayesian predictive model estimating the probability that a single clinician reading one OA article saves a life. We integrate three primary evidence bases: (1) clinician-reported rates of practice change following article consultation, (2) the proportion of clinical decisions that influence short- or long-term mortality, and (3) empirically observed mortality reductions following OA-mediated dissemination of life-saving therapeutic evidence. We then extend this model by incorporating additional determinants of diagnostic and therapeutic accuracy, including medical error rates, years of clinical experience, multimorbidity-dependent diagnostic entropy, cognitive load, structural barriers, team-based reliability, guideline adherence, and electronic health record (EHR)–related error susceptibility, formalized in a multilevel Bayesian framework. The core model yields a probability range of p ≈ 0.003–0.02 that a clinician–article encounter prevents one death, corresponding to a Number Needed to Treat (NNT) analog of approximately 50–330 clinician–article encounters. After accounting for heterogeneity in clinical acuity, multimorbidity, and the extended set of clinician and system parameters, hierarchical Bayesian extensions adjust the predictive interval to p ≈ 0.002–0.03 and NNT ≈ 30–500. The integrated analysis demonstrates that OA literature meaningfully increases the probability of life-saving clinical decisions, especially in high-acuity environments where marginal improvements in evidence latency and accuracy have large mortality consequences.
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 | 0.020 | 0.007 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.000 |
| Bibliometrics | 0.000 | 0.002 |
| Science and technology studies | 0.005 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.006 | 0.006 |
| Research integrity | 0.002 | 0.008 |
| 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