Time-series analysis of ambient temperature and respiratory hospitalizations in Gansu Province, China: a suburban farming population study.

This study investigated the association between temperature and hospitalizations for respiratory diseases (RD) among suburban farmers in Zhangye, Wuwei, Dingxi, and Tianshui in Gansu province. We collected the daily hospital admission data for RD in four cities from the local public hospitals, covering the period from January 1, 2018 to December 31, 2019. The association was estimated using a quasi-Poisson generalized additive model (GAM) and distributed lag nonlinear model (DLNM) to account for lagged and non-linear effects, and the association varies geographically. Our study found that both low and high temperatures were associated with RD morbidity, and had significant lag effects in four cities, the risk of temperature on RD morbidity increased significantly in Zhangye (low temperature: RR = 2.107, 95%CI: 1.749, 2.540; high temperature: RR = 2.407, 95%CI: 1.932, 2.998), Wuwei (low temperature: RR = 1.758, 95%CI: 1.134, 2.726; high temperature: RR = 1.936, 95%CI: 1.541, 2.431), Dingxi (low temperature: RR = 1.876, 95%CI: 1.593, 2.208; high temperature: RR = 2.432, 95%CI: 1.932, 3.061) and Tianshui (low temperature: RR = 1.083, 95%CI: 1.021, 1.150; high temperature: RR = 1.630, 95%CI: 1.191, 2.229). Susceptible demographics linked to RD morbidity differ by gender and age group in four cities. For Wuwei, Dingxi, and Tianshui, females exhibited higher adverse effects when exposed to both low and high temperatures than males. By contrast, males in Zhangye showed higher relative risks (RR) than females. Additionally, in Zhangye, Wuwei, and Tianshui, low temperature had a greater impact on patients aged < 65 years than on those aged ≥ 65 years. For high temperature, patients aged < 65 years in Zhangye, Wuwei, and Dingxi were more susceptible. These findings emphasize the need for region-tailored early warning systems and targeted preventive measures for vulnerable groups. The application of distributed lag non-linear modeling in a suburban agricultural population offers novel insights into environmental epidemiology in resource-constrained settings. Future research should prioritize refining temperature-health threshold definitions and leveraging micro-level exposure data to inform adaptive public health strategies.