Annual health data from 807 Indonesian coal miners followed from 2013-2021 showed 13.9% cumulative black-lung incidence. A complete-blood-count neural network and Cox model both identified eosinophils and monocytes as influential predictors.
Key findings
- Cumulative incidence was 13.9%. Eosinophils had aHR 1.286 (95% CI 1.254-1.319) and monocytes aHR 1.136 (95% CI 1.034-1.247), consistently influential across models. Hazard ratios require the full paper's unit scale and adjustment set for interpretation.
Why this matters globally
CBC is inexpensive and routinely available. If externally validated, the model could prioritize miners for imaging or exposure assessment, but it cannot replace dust control, standard surveillance, or specialist diagnosis.
Thai researcher contribution
Researchers from Chulalongkorn University and Chiang Mai University contributed to a regional analysis of Indonesian occupational-health data; the miner cohort was not Thai.
Limitations to consider
One industry and country limit transportability. CBC counts are affected by allergy, infection, smoking, medication, and many inflammatory conditions. The abstract omits outcome definition, dust exposure, missingness, external validation, and numerical support for high performance.