P-value and Bayesian analysis in randomized-controlled trials in child health research published in 2007 and 2017: a methodological review
Bibliographic record
Abstract
Abstract Background Reliance on P-value of significance in clinical trials is a source of debate because of misconceptions and misinterpretations associated with it. Bayesian methods are suggested as an alternative approach. As randomized-controlled trials (RCTs) are essential in generating research evidence, we investigated the change in the use of P-values and Bayesian analysis and the clustering of P-values at key significance levels in child health RCTs published in 2007 and 2017. Methods We searched Cochrane Central Register of Controlled Trials to identify random samples of child health RCTs published in 2007 (n = 300) and 2017 (n = 300). Data on trial characteristics and analytic approaches were extracted. We analyzed the 600 RCTs using the frequentist and Bayesian methods. The change in the proportion of trials reporting P-values and Bayesian analyses was assessed using Pearson/Fisher Exact tests and non-informative Dirichlet priors. Results Of 600 RCTs, 535 (89%) used frequentist methods only versus 65 (11%) that included some Bayesian methods. Only 2 of the 65 trials used Bayesian inferential statistics. The use of frequentist methods decreased from (273, 91% to 262, 87%) while the inclusion of Bayesian analysis slightly increased from (27, 9% to 38, 13%) between 2007 and 2017. Although most RCTs were from Europe (172, 29%) and North America (133, 22%), the increase in proportion of trials by continent was most in Asia (mean difference (MD) = 0.14, 95% credible interval (CI) 0.08–0.20) with posterior probability (PP) of 1.00. Parallel (487, 81.2%) and cluster (58, 9.7%) RCTs were the most common RCT types but the increase in cluster RCT (0.06, 95%CI 0.02–0.11, PP = 0.99) was more than any other RCT types over 10 years. We found clustering of P-values at the significance level of 0.05 (437, 72.8%), which increased between 2007 (209, 69.7%) and 2017 (228, 76%). The smallest P-value reported in this review was 0.0001 (1, 0.2%). Conclusions The statistical framework in child health RCTs has not changed from the frequentist methods that is based on P-values with an unexplained clustering at the significance level of 0.05. Bayesian methods may increase the confidence in interpretation of results of RCTs
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
Direct model labels (unvalidated)
Per-model category and study-design labels from the labeling rounds. They are machine output, unvalidated, and the disagreement between models ships as data. No study design here is MEDLINE-validated yet.
| Model arm | Categories | Study design | Confidence |
|---|---|---|---|
| gemma | Metaresearch Domain: Methods · Genre: Review About the Canadian research system: no · About a Canadian topic: no | Observational | low |
| gpt | Metaresearch Domain: Methods · Genre: Review About the Canadian research system: no · About a Canadian topic: no | Other design | medium |
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 | 0.414 | 0.806 |
| Meta-epidemiology (narrow) | 0.002 | 0.003 |
| Meta-epidemiology (broad) | 0.016 | 0.018 |
| Bibliometrics | 0.026 | 0.029 |
| Science and technology studies | 0.002 | 0.009 |
| Scholarly communication | 0.010 | 0.012 |
| Open science | 0.008 | 0.006 |
| Research integrity | 0.008 | 0.009 |
| Insufficient payload (model declined to judge) | 0.004 | 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, unvalidatedLabeled directly by 2 models reading the full record.
The models disagree on parts of this classification; every voice is preserved in the section at the end of the page.
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".