A Spatio-temporal analysis to assess the influence of climate change on respiratory diseases in Kampala city, Uganda

Abstract

Background: Climate change poses serious public health risks, especially the urban areas experiencing rapid growth and environmental stress. Kampala city is vulnerable to climate variability such as rising temperatures, fluctuating humidity, and irregular precipitation, which have amplified air pollution levels, exacerbating respiratory health problems. These conditions, alongside urbanisation, have increased respiratory illnesses, including asthma, chronic obstructive pulmonary disease, and pneumonia. However, little is known about how these environmental exposures interact to influence the spatial and temporal distribution of respiratory diseases in Kampala. Objectives: The study aimed to assess the impacts of climate factors on the morbidities and spatial-temporal distribution of respiratory diseases in Kampala over the last four years (2021-2024). This study provides insights into the relationship between microclimate variability, air quality, and respiratory health patterns across Kampala City, Uganda.Methods: A retrospective ecological study design was employed, using monthly aggregated data on respiratory disease illnesses obtained from the Ministry of Health (DHIS2). Climate variables were accessed and extracted through the DHIS2 climate system, an integrated tool maintained by HISP Uganda, and air quality was obtained from Air Qo devices through KCCA. All data cleaning, transformation, and statistical analyses were conducted in R. Generalized Additive Models (GAMs) were used to examine nonlinear relationships between environmental variables and disease outcome. To capture spatial and temporal dynamics, Spatial-Temporal General Additive Models (SP-GAMs) were applied. Residual checks and model diagnostics ensured the validity of results using randomised quantile residuals and smoothing terms. Results: Between 2021-2024, a total of 89253 respiratory illnesses were recorded, comprising pneumonia (65%), asthma (23%), and COPD (12%). In the adjusted GAM models, particulate matter at 2.5(pm2.5) was identified as the most statistically significant and consistent predictor of respiratory disease across all models, showing nonlinear associations with COPD (edf = 2.59, p < 0.001), asthma (edf = 3.25, p = 0.016), and pneumonia (edf = 3.62, p = 0.006). COPD also increased with rising temperatures (edf = 1.01, p < 0.001) and significantly with increased precipitation (edf = 1.00, p < 0.001). ST-GAM results showed strong spatial clustering (edf ≈ 4, p <0.001) and temporal variation. Spatiotemporal analysis using ST-GAMs showed significant clustering of respiratory diseases across Kampala (edf ≈ 4, p < 0.001), with Central, Makindye, and Kawempe divisions, as confirmed by LISA high–high clusters (p < 0.05). Temporally, Pneumonia cases peaked between March and May and again from September to November, while asthma and COPD were more prevalent between January and February and from June to August. Conclusion: The study provides evidence that air pollution, particularly PM2.5, is a key modifiable risk factor for respiratory illness in Kampala. The interaction between environmental exposures and disease outcomes is complex and varies by disease type, location, and season. These findings call for urgent intervention in air quality monitoring, climate-sensitive public health planning, and targeted interventions for vulnerable populations in high-risk divisions. Long-term surveillance and integrated data systems are essential to strengthen Kampala’s urban health resilience in the face of ongoing environmental change.