Retrieval barriers in individual participant data reviews with network meta-analysis
Bibliographic record
Abstract
OBJECTIVES: Individual participant data (IPD) from randomised controlled trials (RCTs) can be used in network meta-analysis (NMA) to underpin patient care and are the best analyses to support the development of guidelines about the use of healthcare interventions for a specific condition. However, barriers to IPD retrieval pose a major threat. The aim of this study was to present barriers we encountered during retrieval of IPD from RCTs in two published systematic reviews with IPD-NMA. METHODS: We evaluated retrieval of IPD from RCTs for IPD-NMA in Alzheimer's dementia and type 1 diabetes. We requested IPD from authors, industry sponsors and data repositories, and recorded IPD retrieval, reasons for IPD unavailability, and retrieval challenges. RESULTS: In total, we identified 108 RCTs: 78 industry sponsored, 11 publicly sponsored and 19 with no funding information. After failing to obtain IPD from any trial authors, we requested it from industry sponsors. Seven of the 17 industry sponsors shared IPD for 12 950 participants (59%) through proprietary-specific data sharing platforms from 26 RCTs (33%). We found that lack of RCT identifiers (eg, National Clinical Trial number) and unclear data ownership were major challenges in IPD retrieval. Incomplete information in retrieved datasets was another important problem that led to exclusion of RCTs from the NMA. There were also practical challenges in obtaining IPD from or analysing it within platforms, and additional costs were incurred in accessing IPD this way. CONCLUSIONS: We found no clear evidence of retrieval bias (where IPD availability was linked to trial findings) in either IPD-NMA, but because retrieval bias could impact NMA findings, subsequent decision-making and guideline development, this should be considered when assessing risk of bias in IPD syntheses.
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: Empirical About the Canadian research system: no · About a Canadian topic: no | Observational | low |
| gpt | Metaresearch Domain: Methods · Genre: Empirical About the Canadian research system: no · About a Canadian topic: no | Other design | medium |
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 | 0.539 | 0.198 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.015 | 0.003 |
| Bibliometrics | 0.001 | 0.020 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.008 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.142 | 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, 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".