Assessing the health and economic burden of sick leave associated with long-term fine particulate matter exposure: a 13-year analysis using German health claims data

Background Ambient fine particulate matter (PM 2.5 ) is a major environmental health risk, yet its contribution to sick leave and related productivity losses remains insufficiently quantified. We therefore examined the association between PM 2.5 exposure and all-cause sick leave days and estimated the corresponding model-based burden and productivity losses. Methods Using routine health insurance data from AOK Baden-Württemberg for 2010–2022, we fitted generalized additive models with penalised splines to estimate the exposure–response curve between monthly PM 2.5 exposure and all-cause sick leave days. Models were adjusted for long-term trends, seasonality, temperature, sex, the COVID-19 period, NO 2 , vacation days, and lockdown days, and accounted for spatial heterogeneity using a postal-code-based Markov random field. Model-based exposure contrasts were derived from predicted sick leave days at selected PM 2.5 concentrations. Productivity losses were estimated using the human capital approach. Results PM 2.5 showed a positive non-linear association with sick-leave days (p < 0.001). Compared with 5 µg/m 3 , adjusted predicted sick-leave days were 9.56% higher at 10 µg/m 3 (95% CI 9.09–10.00) and 15.00% higher at 20 µg/m 3 (95% CI 14.20–15.80), while little change was observed between 10 and 15 µg/m 3 (+0.04%, 95% CI − 0.30 to 0.38). The overall non-linear association was supported across additional analyses, although its magnitude and precise shape varied with temporal coverage and covariate adjustment. PM 2.5 concentrations above the WHO guideline of 5 µg/m 3 corresponded to 105 model-attributable excess sick-leave days per 1,000 working-age inhabitants per year, equivalent to approximately 5.31 million excess sick-leave days and €1.56 billion in annual productivity losses in Germany. Conclusion Higher PM 2.5 exposure was associated with a relevant model-attributable burden of sickness absence and productivity losses. These findings suggest that ambient PM 2.5 may be an important population-level correlate of work absence, although causal interpretation is limited by the ecological study design and residual confounding.