Association of climate variables with pneumonia patterns in a tropical coastal province of thailand: a bayesian structural time series analysis

Pneumonia is a significant global public health problem, and climatic factors play a crucial role in disease epidemic dynamics. However, there are limited studies on this relationship in tropical monsoon coastal climate regions. This study aimed to analyze the impact of climatic factors on pneumonia pattern in Satun Province, a southern coastal province of Thailand, using Bayesian Structural Time Series (BSTS) modeling. This ecological study utilized monthly time series data spanning 10 years (January 2015 - December 2024). Pneumonia case data (ICD-10: J12-J18) were collected from the National Disease Surveillance System, and monthly climatic data were obtained from Satun Province meteorological stations. The BSTS model was employed to decompose long-term trend components, seasonal patterns, and the simultaneous effects of multiple climatic variables, using spike-and-slab priors technique for Bayesian variable selection. The median number of pneumonia cases was 63.0 cases per month (IQR: 34.0-102.0). BSTS model analysis revealed that relative humidity was the most influential climatic predictor of pneumonia cases, demonstrating a 2-month lag effect (Posterior Inclusion Probability [PIP] = 0.284). A one standard deviation increase in relative humidity was associated with approximately 6.4% increase in cases. Other climatic variables such as temperature, precipitation, and wind speed were not identified as robust predictors in the final model. Additionally, distinct seasonal epidemic patterns were observed, with case peaking during October-November. Relative humidity is a factor associated with pneumonia epidemics in Satun Province, challenging traditional concepts that typically emphasize temperature as the primary factor in temperate regions. The discovery of this 2-month lag relationship is significant for developing public health early warning systems, enabling effective preparation of resources and proactive preventive measure planning.