An application of semi-bayes modeling to a study of the occupational etiology of lung cancer /
Bibliographic record
Abstract
The occupational environment has been a fruitful source of research on causes of cancer. Analyses in studies of occupational risk factors for cancer can experience problems if an attempt is made to model large numbers of exposures, some of which may be highly correlated. Typical analyses of such studies focus on one chemical at a time, but this may not adequately deal with mutual confounding. Based on a large study in Montreal, the objective of this thesis was twofold: to assess several occupational chemicals for their etiologic role in lung cancer, and to explore the use of semi-Bayes modeling to simultaneously estimate the effects of many chemicals at a time. Methods. Data came from a multiple-cancer case-control study of exposures in the work place. The study was comprised of 857 cases of lung cancer and 2172 controls consisting of patients with other types of cancer diagnosed from 1979 to 1985. Detailed occupational histories were collected and occupational hygienists translated these into exposure histories for 231 chemicals. All chemicals were analysed with conventional modeling strategies of both single and multiple parameter models. Of the 231 chemicals, 184 were singled out for analysis in a single large semi-Bayes model, which is a variant of classical empirical Bayes. This analysis is a fairly novel method suited to estimating large numbers of parameters in the face of sparse data. For the Bayesian portion of this model, chemicals were grouped by shared chemical and physical properties, based on the belief that these shared properties would imply similar effects on the risk of lung cancer. Results. Estimates for all 231 chemicals were derived under the various modeling strategies. For most chemicals, estimates changed little across these analytic approaches, though some differences were apparent. Of the 231 chemicals assessed, 53 were earmarked as requiring further evaluation and underwent additional analyses. Discussion. While semi-Bayes models have been shown previously to offer improved estimation over conventional analyses, the gains in using semi-Bayes models in the present study were less clear. Effort put into some portions of the Bayesian modeling did not materially influence the results. A number of chemicals were earmarked as potential lung carcinogens.
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 | 0.018 | 0.034 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.002 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 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; 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".