Study designs: Observational epidemiology (cross-sectional, cohort, case–control), exposure assessment studies, case reports/series (if informative), intervention/evaluation studies, and qualitative studies providing contextual information. Experimental laboratory studies not directly assessing occupational conditions will be excluded unless they evaluate exposure–health relationships directly relevant to workplace exposures.
Language and date limits: Studies published in English or Indonesian between 2000-01-01 and 2025-11-09 (search may be extended on request).
Searches and sources: Systematic searches will be conducted in MEDLINE/PubMed, Scopus, ProQuest, SpringerLink, Google Scholar, and Web of Science. Reference lists of included studies and citation-tracking of key papers will be performed. Search strategies will combine controlled vocabulary (MeSH/EMTREE) and free-text terms for charcoal, occupational exposure and relevant hazards/outcomes. All search strings and dates will be recorded.
Screening and selection: Records will be de-duplicated and screened in two stages (title/abstract, then full text) by two independent reviewers. Discrepancies will be resolved by discussion or third-party adjudication. Reasons for exclusion at full text will be recorded.
Data extraction: A piloted extraction form will capture: study identifiers, country/setting, population characteristics (age, sex, worker type), study design, sample size, exposure assessment methods and results (sampling method, metric, concentrations), outcome definitions and measures, effect estimates (with CIs), control measures/PPE details, confounders considered, and authors’ conclusions. Two reviewers will extract data independently; disagreements resolved by consensus.
Risk of bias/quality assessment: Appropriate tools will be used by study design: Newcastle–Ottawa Scale or ROBINS-I for cohort/case–control/non-randomised studies, AXIS or adapted JBI checklist for cross-sectional and prevalence/exposure studies, Cochrane RoB2 for any randomized evaluations, and a structured checklist for exposure-assessment methodological quality (sampling strategy, instrument calibration, personal vs area sampling, duration). Two reviewers will assess risk of bias independently.
Data synthesis: A narrative synthesis will describe exposures, methods and outcomes across studies. Where ≥2 studies report comparable quantitative outcomes (e.g., mean PM2.5 concentrations or pooled effect estimates for a specific health outcome), meta-analysis will be considered using a random-effects model. Statistical heterogeneity will be assessed with I² and Cochran’s Q. If meta-analysis is not appropriate, results will be tabulated and synthesised narratively. Publication bias will be examined with funnel plots and Egger’s test where ≥10 studies are combined.
Planned subgroup and sensitivity analyses: Where data allow: subgroup analyses by region/country income level, worker subgroup (kiln operator, transport, packing), study design, exposure metric (personal vs area sampling), and by informal vs formal production. Sensitivity analyses will exclude studies at high risk of bias.
Assessment of evidence certainty: The GRADE approach (or an adapted GRADE for observational evidence) will be used to rate the certainty of evidence for primary outcomes.
Reporting and dissemination: Findings will be reported according to PRISMA guidelines, including a flow diagram, summary tables of exposures and outcomes, risk-of-bias assessments, and gaps/recommendations. Results will be submitted for peer-reviewed publication and shared with occupational health stakeholders and NGOs working in biomass/charcoal sectors.