Do pragmatic trials trade-off internal validity for external validity?
Bibliographic record
Abstract
Trials can be described as being on a design spectrum between highly explanatory (roughly, ‘Can the intervention work?’) to highly pragmatic (‘Does the intervention work in routine care?’). A criticism levelled at trials that take a pragmatic approach is that they sacrifice internal validity for external validity, i.e. there is a trade-off to be made. Proponents of pragmatic trials argue that there is no trade-off. However, both critics and defenders of the pragmatic approach have made their arguments in the absence of empirical evidence one way or the other. As part of our work developing the PRECIS-2 trial design tool, we looked for evidence of the validity trade-off in two ways. Firstly, a sample of 14 cardiovascular explanatory trials was matched to trials of the same intervention but which took a more pragmatic design approach. Secondly, 23 trials included in a Cochrane review of first-line treatment for hypertension were compared. The Cochrane Risk of Bias tool was used to assess internal validity; PRECIS-2 was used to assess design approach. We found no clear difference in Cochrane Risk of Bias assessments between trials taking explanatory and pragmatic approaches. Work with a larger sample of trials is required before we can be completely confident in this result but this represents the first evidence that the suggested trade-off between internal and external validity in trials may be false. This work is part of the Trial Forge initiative to improve trial efficiency.
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 | Theoretical or conceptual | low |
| gpt | Metaresearch Domain: Methods · Genre: Methods About the Canadian research system: no · About a Canadian topic: no | Theoretical or conceptual | low |
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 | 0.848 | 0.939 |
| Meta-epidemiology (narrow) | 0.004 | 0.006 |
| Meta-epidemiology (broad) | 0.017 | 0.015 |
| Bibliometrics | 0.014 | 0.013 |
| Science and technology studies | 0.004 | 0.045 |
| Scholarly communication | 0.024 | 0.049 |
| Open science | 0.008 | 0.017 |
| Research integrity | 0.022 | 0.020 |
| Insufficient payload (model declined to judge) | 0.013 | 0.004 |
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.
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".