Nonlinear and lagged effects of climate variability on dengue incidence in an urban megacity: a distributed lag non-linear model (DLNM) based study in Bangkok, Thailand.

BACKGROUND

Dengue fever represents a significant and escalating public health challenge in tropical megacities such as Bangkok, Thailand. This study aims to examine the associations between temperature, humidity, rainfall, and wind speed with dengue incidence across multiple lag periods and to explore spatial heterogeneity in these climate-dengue relationships across Bangkok’s urban zones. METHODOLOGY: Analysis of 105,890 confirmed dengue cases reported in Bangkok between 2015 and 2024, stratified across six urban zones. Distributed Lag Non-Linear Models (DLNM) were employed to quantify the effects of four key meteorological variables: temperature, humidity, rainfall, and wind speed.

RESULTS

Significant non-linear relationships and heterogeneity in climate-driven dengue risk were found across both space and time. Spatially, the effects were most pronounced for humidity and wind speed. For instance, over a 0–2 months lag, higher humidity was associated with a substantial increase in risk in North Krung Thon (RR=1.477, 95% CI: 1.290–1.690) but a significant protective effect in South and Middle Bangkok. Similarly, wind speed was associated with a significant risk increase in South Bangkok (RR=1.405, 95% CI: 1.293–1.526) but a protective effect in East Bangkok (RR=0.749, 95% CI: 0.710–0.789). Elevated minimum temperature also exhibited a spatially varied impact, peaking in Middle Bangkok (RR=1.350, 95% CI: 1.288–1.414). The analysis confirmed distinct patterns over time, with climate impacts manifesting as immediate risks in some zones and as pronounced delayed risks (8–12 months) in others.

CONCLUSION

The associations between climatic variables and dengue incidence in Bangkok are highly complex, non-linear, and characterized by significant spatial and temporal heterogeneity. The varied timing of climate impacts, from immediate to delayed, suggests that public health responses must be adapted to the unique temporal risk structures of each urban zone, providing a framework for more precise interventions in complex urban environments.