Framework for Exploration of Statistical Heterogeneity in Multi-Database Studies: A Case Study Using EXACOS-CV Studies
Bibliographic record
Abstract
Purpose: Multi-database studies may provide heterogeneous results despite using common protocols, leading to challenges in interpretation, but also providing an opportunity to gain insights on populations or healthcare systems. The objectives of these analyses were to develop a framework for exploring sources of statistical heterogeneity and apply it to the multi-database EXACOS-CV (EXAcerbations of COPD and their OutcomeS on CardioVascular diseases) program. Methods: A conceptual framework to systematically assess sources of statistical heterogeneity in multi-database studies was developed. This framework distinguishes between methodological diversity and true clinical variation. Methodological diversity includes differences in study design and database selection, while true variation considers population and healthcare differences. Possible sources of methodological diversity were identified via a novel checklist and explored. In turn, hypotheses were generated about true variation. The framework and checklist were applied to EXACOS-CV cohort studies in Germany, Canada, the Netherlands, and Spain which deviated least from the common protocol and so were included. Focus was on adjusted hazard ratios (aHR) for post-exacerbation associations with decompensated heart failure (HF) and all-cause death, for which results were most and least heterogeneous, respectively. Results: Across EXACOS-CV studies, the adjusted hazard ratios (aHR) for HF in the first 1-7 days post-exacerbation, compared to non-exacerbation periods, ranged from 2.6 (95% CI, 2.3, 2.9) in Germany to 72.3 (64.4, 81.2) in Canada, and the association with death, relative to non-exacerbation periods, ranged from 3.5 (2.4, 5.3) in the Netherlands to 22.1 (19.9, 24.4) in Spain. Completed methodological diversity checklists linked differences in aHRs to possible variation in ability to capture pre-existing cardiovascular comorbidities across studies, as well as differences in confounder measurement. Standardizing adjusted models across studies did not fully explain heterogeneity, suggesting other contributing factors. Heterogeneity may result from genuine variation in prevalence of CV disease. It was hypothesized that patients with pre-existing CV disease have more accurate diagnoses and management of post-exacerbation CV events, possibly leading to lower risks of such events. Conclusion: Multi-database studies can provide directional insights on the study research question while considering healthcare system and population differences. The developed framework aids assessment of heterogeneity sources.
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 | 0.602 | 0.657 |
| Meta-epidemiology (narrow) | 0.003 | 0.002 |
| Meta-epidemiology (broad) | 0.004 | 0.009 |
| Bibliometrics | 0.024 | 0.015 |
| Science and technology studies | 0.005 | 0.013 |
| Scholarly communication | 0.011 | 0.011 |
| Open science | 0.009 | 0.019 |
| Research integrity | 0.008 | 0.009 |
| Insufficient payload (model declined to judge) | 0.003 | 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, 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".