Bibliographic record
Abstract
Among the many challenges facing institutional review boards (IRBs) is to predict whether the activities and interventions proposed in a clinical trial protocol are likely to yield net harm or net benefit for trial participants. IRBs take these questions very seriously, and never more so than in the review of first-in-human (FIH) trials, where interpreting findings about risks to humans from animal data requires a leap of faith, regardless of the quality of the available data. In their paper published this week in PLoS Medicine [1], “Predicting harms and benefits in translational trials: Ethics, evidence, and uncertainty,” Jonathan Kimmelman and Alex London argue that decision-makers (which, from the context of their paper, I assume to mean IRB members) pay insufficient attention to threats to validity in preclinical studies and consult too narrow a set of evidence, thereby unnecessarily limiting predictions about risks and potential benefits for humans that they might otherwise be able to make. They advocate greater attention to the quality of preclinical evidence and to research on related agents. These strategies are meant to reduce what they call the “misestimation” of risks or anticipated benefits, which they argue “threatens the integrity of the scientific enterprise, because it frustrates prudent allocation of research resources”[1]. Kimmelman and London's proposal is likely to stimulate a great deal of constructive debate among clinical trialists, regulators, and other members of the research ethics community. In my brief comments here, I will attempt to open this debate by identifying a key aspect of their proposal that is likely to generate particular interest and perhaps even some controversy—that is, their framing of the problem in terms of how effectively decision-makers utilize evidence from preclinical or animal studies. Although IRB members often do not have deep grounding in the subtleties of research design and inferential statistics, it would be wrong to suggest that “misestimation” of risk and potential benefit arises solely from errors by IRB members (or other decision-makers).
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 | 0.001 | 0.005 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.000 | 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; a candidate call from one teacher head, not a consensus.
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".