Abstract P274: A meta-analysis of meta-analyses: Is "reverse Epi" OK or not?
Bibliographic record
Abstract
Introduction: Applications of observational risk models assuming “all else equal” to estimate the effect on risk of interventions expected to change the level of one or more independent variables have been criticized as “reverse Epi” since clinical trials provide better estimates of treatment effects. Hypothesis: The purpose of this analysis is to assess how reasonable it is to use “reverse Epi” to project the number of events avoided under different scenarios in which different markers are used for determining to whom to prescribe and how to target preventative treatment. We pursued this aim by comparing two different meta-analyses which compared HRs for apoB and non-HDL-C. One meta-analysis included 12 published observational studies. In 11 of these studies all subjects were at primary CHD risk. The other meta-analysis included 7 placebo-controlled statin trials (3 primary and 4 secondary). Methods: We calculated the reciprocal of each observational study’s HR per 1-SD increase (36 mg/dl for non-HDL-C and 27 mg/dl of apoB) to compare with each trials’ HR per 1-SD decrease; calculated the percentage difference (95% CI) in standard HRs for each report assuming a correlation between the two markers of 0.93 (calculated from NHANES 2005-2010 to be representative of adult US residents); and compared the resulting statistics using standard subgroup meta-analysis. Results: The overall geometric mean point estimate (95% CI) of the difference favoring apoB over non-HDL-C for all observational studies was 11.1% (7.7%, 14.4%) versus 5.4% (2.9%, 9.2%) for all 7 trials (p = 0.007). However these same point estimates are 9.5% (7.3%, 14.0%) versus 10.6% (7.3%, 14.0%) (p=0.714) when only primary studies or trials were included. The geometric mean treatment effect HR among the 4 secondary trials was 0.732 (0.689, 0.792) versus 0.599 (0.504, 0.714) among the 3 primary trials (p = 0.035). Conclusion: Meaningful differences between primary and secondary observational and interventional studies warrant stratified analysis. Among subjects at primary CHD risk, “reverse Epi” application of primary observational results may provide projections similar to those from applying primary trial results.
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 | 0.135 | 0.373 |
| Meta-epidemiology (narrow) | 0.004 | 0.003 |
| Meta-epidemiology (broad) | 0.015 | 0.034 |
| Bibliometrics | 0.007 | 0.009 |
| Science and technology studies | 0.001 | 0.003 |
| Scholarly communication | 0.006 | 0.006 |
| Open science | 0.005 | 0.002 |
| Research integrity | 0.005 | 0.008 |
| Insufficient payload (model declined to judge) | 0.008 | 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; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
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".