BMJ Open (Aug 2024)

Spatial-temporal analysis of climate and socioeconomic conditions on cholera incidence in Mozambique from 2000 to 2018: an ecological longitudinal retrospective study

  • Joacim Rocklöv,
  • Aditi Bunker,
  • Maquins Odhiambo Sewe,
  • Mohsin Sidat,
  • Chaibo Jose Armando,
  • Yesim Tozan,
  • Alberto Francisco Mavume

DOI
https://doi.org/10.1136/bmjopen-2023-082503
Journal volume & issue
Vol. 14, no. 8

Abstract

Read online

Objectives This study aims to assess both socioeconomic and climatic factors of cholera morbidity in Mozambique considering both spatial and temporal dimensions.Design An ecological longitudinal retrospective study using monthly provincial cholera cases from Mozambican Ministry of Health between 2000 and 2018. The cholera cases were linked to socioeconomic data from Mozambique Demographic and Health Surveys conducted in the period 2000–2018 and climatic data; relative humidity (RH), mean temperature, precipitation and Normalised Difference Vegetation Index (NDVI). A negative binomial regression model in a Bayesian framework was used to model cholera incidence while adjusting for the spatiotemporal covariance, lagged effect of environmental factors and the socioeconomic indicators.Setting Eleven provinces in Mozambique.Results Over the 19-year period, a total of 153 941 cholera cases were notified to the surveillance system in Mozambique. Risk of cholera increased with higher monthly mean temperatures above 24°C in comparison to the reference mean temperature of 23°C. At mean temperature of 19°C, cholera risk was higher at a lag of 5–6 months. At a shorter lag of 1 month, precipitation of 223.3 mm resulted in an 57% increase in cholera risk (relative risk, RR 1.57 (95% CI 1.06 to 2.31)). Cholera risk was greatest at 3 lag months with monthly NDVI of 0.137 (RR 1.220 (95% CI 1.042 to 1.430)), compared with the reference value of 0.2. At an RH of 54%, cholera RR was increased by 62% (RR 1.620 (95% CI 1.124 to 2.342)) at a lag of 4 months. We found that ownership of radio RR 0.29, (95% CI 0.109 to 0.776) and mobile phones RR 0.262 (95% CI 0.097 to 0.711) were significantly associated with low cholera risk.Conclusion The derived lagged patterns can provide appropriate lead times in a climate-driven cholera early warning system that could contribute to the prevention and management of outbreaks.