A plea for more careful scholarship in reviewing evidence: the case of mammographic screening
Bibliographic record
Abstract
Objectives: To identify issues of principle and practice giving rise to misunderstandings in reviewing evidence, to illustrate these by reference to the Nordic Cochrane Review (NCR) and its interpretation of two trials of mammographic screening, and to draw lessons for future reviewing of published results. Methods: A narrative review of the publications of the Nordic Cochrane Review of mammographic screening (NCR), the Swedish Two-County Trial (S2C) and the Canadian National Breast Screening Study 1 and 2 (CNBSS-1 and CNBSS-2). Results: The NCR concluded that the S2C was unreliable, despite the review's complaints being shown to be mistaken, by direct reference to the original primary publications of the S2C. Repeated concerns were expressed by others about potential subversion of randomisation in CNBSS-1 and CNBSS-2; however, the NCR continued to rely heavily on the results of these trials. Since 2022, however, eyewitness evidence of such subversion has been in the public domain. Conclusions: An over-reliance on nominal satisfaction of checklists of criteria in systematic reviewing can lead to erroneous conclusions. This occurred in the case of the NCR, which concluded that mammographic screening was ineffective or minimally effective. Broader and more even-handed reviews of the evidence show that screening confers a substantial reduction in breast cancer mortality. Advances in knowledge: Those carrying out systematic reviews should be aware of the dangers of over-reliance on checklists and guidelines. Readers of systematic reviews should be aware that a systematic review is just another study, with the capability that all studies have of coming to incorrect conclusions. When a review seems to overturn the current position, it is essential to revisit the publications of the primary research.
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 | 0.731 | 0.843 |
| Meta-epidemiology (narrow) | 0.004 | 0.006 |
| Meta-epidemiology (broad) | 0.013 | 0.010 |
| Bibliometrics | 0.029 | 0.014 |
| Science and technology studies | 0.018 | 0.135 |
| Scholarly communication | 0.043 | 0.124 |
| Open science | 0.018 | 0.037 |
| Research integrity | 0.068 | 0.088 |
| Insufficient payload (model declined to judge) | 0.005 | 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".