Effect of standardized training on the reliability of the Cochrane risk of bias assessment tool: a prospective study
Bibliographic record
Abstract
BACKGROUND: The Cochrane risk of bias tool is commonly criticized for having a low reliability. We aimed to investigate whether training of raters, with objective and standardized instructions on how to assess risk of bias, can improve the reliability of the Cochrane risk of bias tool. METHODS: In this pilot study, four raters inexperienced in risk of bias assessment were randomly allocated to minimal or intensive standardized training for risk of bias assessment of randomized trials of physical therapy treatments for patients with knee osteoarthritis pain. Two raters were experienced risk of bias assessors who served as reference. The primary outcome of our study was between-group reliability, defined as the agreement of the risk of bias assessments of inexperienced raters with the reference assessments of experienced raters. Consensus-based assessments were used for this purpose. The secondary outcome was within-group reliability, defined as the agreement of assessments within pairs of inexperienced raters. We calculated the chance-corrected weighted Kappa to quantify agreement within and between groups of raters for each of the domains of the risk of bias tool. RESULTS: A total of 56 trials were included in our analysis. The Kappa for the agreement of inexperienced raters with reference across items of the risk of bias tool ranged from 0.10 to 0.81 for the minimal training group and from 0.41 to 0.90 for the standardized training group. The Kappa values for the agreement within pairs of inexperienced raters across the items of the risk of bias tool ranged from 0 to 0.38 for the minimal training group and from 0.93 to 1 for the standardized training group. Between-group differences in Kappa for the agreement of inexperienced raters with reference always favored the standardized training group and was most pronounced for incomplete outcome data (difference in Kappa 0.52, p < 0.001) and allocation concealment (difference in Kappa 0.30, p = 0.004). CONCLUSIONS: Intensive, standardized training on risk of bias assessment may significantly improve the reliability of the Cochrane risk of bias tool.
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 | 0.433 | 0.726 |
| Meta-epidemiology (narrow) | 0.002 | 0.003 |
| Meta-epidemiology (broad) | 0.005 | 0.009 |
| Bibliometrics | 0.005 | 0.007 |
| Science and technology studies | 0.001 | 0.005 |
| Scholarly communication | 0.003 | 0.006 |
| Open science | 0.003 | 0.004 |
| Research integrity | 0.003 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
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".