Return
Temperature–mortality associations and attributable burden in Vojvodina, Serbia, 2000–2020
I
DOI:10.1007/s00484-026-03295-1.png)
Abstract
En 中文
Ambient temperature is a well-established determinant of mortality, yet evidence from temperate continental regions of Central and Southeastern Europe remains limited. Public health attention has largely concentrated on extreme temperature events, notwithstanding growing evidence that non-extreme temperatures may account for a substantial share of temperature-related deaths. We examined the association between daily mean temperature and all-cause mortality in the Autonomous Province of Vojvodina, Serbia, over 2000–2020. Distributed lag non-linear models were used to capture non-linear and delayed temperature effects over a lag period of up to 21 days. Attributable mortality was estimated using a counterfactual framework applied across the full temperature distribution. The temperature–mortality relationship was strongly non-linear and asymmetric. Cold-related effects substantially exceeded those of heat, yielding a temperature-attributable fraction of 9.8% of all deaths. The burden was concentrated at moderately cold temperatures, particularly between 0 and 10 °C, whereas extreme cold and heat each contributed relatively small proportions. Lag–response patterns showed short-lived heat effects compatible with mortality displacement, contrasting with delayed and sustained cold effects persisting over several weeks. In this temperate continental setting, moderate cold is the primary contributor to temperature-related mortality. Public health strategies focused narrowly on rare temperature extremes are unlikely to capture most of this burden. Sustained wintertime risk management and improvements in housing and heating resilience should therefore be central to climate adaptation policy in comparable regions.
Keywords:
Temperature–mortality association
Attributable mortality
Moderate cold
Distributed lag non-linear models
Temperate continental climate
Vojvodina
Serbia
Journal
IF:
2.6
Papers:
3.5K
Citations:
9.7K
