Short-term association between daily mean temperature and acute respiratory infection counts in tropical Costa Rica: an ecological time-series study.

This study evaluated the association between daily mean temperature and reported non-COVID-19 acute respiratory infection (ARI) counts in tropical Costa Rica. Daily ARI counts from Liberia Canton (2009-2023) were analyzed using quasi-Poisson distributed lag non-linear models. Daily mean temperature was derived from ERA5-Land using polygon-based exposure assessment, with adjustment for dew-point temperature, season, and the COVID-19 period. Cumulative associations were estimated over lags 0-3 days, with exploratory windows of 0-7 and 0-21 days. Of 3,338 ARI records, 1,536 records explicitly coded as COVID-19 in the source cause field were excluded, leaving 1,802 non-COVID-19 cases; 73.9% of days had zero recorded cases. Lower daily mean temperatures were associated with higher ARI counts. At the 5th percentile of the temperature distribution (23.8 °C), the cumulative relative risk was 1.39 (95% CI: 1.20-1.61) compared with the reference temperature (25.5 °C) over lags 0-3 days. The direction was generally similar across non-temporal sensitivity analyses, but the estimate was attenuated under increasingly flexible calendar-time adjustment. These findings describe an association within a narrow tropical temperature range; residual temporal confounding cannot be excluded, and prospective evaluation is required before meteorological information can be considered for early-warning applications.