Cholera forecast for Dhaka, Bangladesh, with the 2016 El Ni\~no

TL;DR

This study develops and compares models to forecast cholera outbreaks in Dhaka influenced by the 2016 El Niño, demonstrating high prediction accuracy and providing early warning for public health preparedness.

Contribution

It introduces a process-based cholera transmission model incorporating SST anomalies and evaluates its forecasting performance for the first time in this context.

Findings

High probability of increased cholera cases in Dhaka during Aug-Dec 2016

Forecast models accurately predicted low cholera risk in early 2016

El Niño significantly influences cholera dynamics in Bangladesh

Abstract

A substantial body of work supports a teleconnection between the El Ni\~no Southern Oscillation (ENSO) and cholera incidence in Bangladesh. In particular, high positive anomalies during the winter (Dec-Feb) in Sea Surface Temperatures (SST) in the Tropical Pacific have been shown to exacerbate the seasonal outbreak of cholera following the monsoons from Aug to Nov, and climate studies have indicated a role of regional precipitation over Bangladesh in mediating this long-distance effect. Thus, the current strong El Ni\~no has the potential to significantly increase cholera risk this year in Dhaka, Bangladesh, where the last five years have experienced low seasons of the disease. To examine this possibility and produce a forecast for the city, we considered two models for the transmission dynamics of cholera: a statistical model previously developed for the disease in this region, and a process-based model presented here that includes the effect of SST anomalies in the force of infection and is fitted to extensive cholera surveillance record between 1995 and 2010. Prediction accuracy was evaluated with 'out-of-fit' data from the same surveillance efforts, by comparing the total number of cholera cases observed for the season to those predicted by model simulations 8 to 12 months ahead, starting in January each year. Encouraged by accurate forecasts for the low risk of cholera for this period, we then generated a prediction for this coming season. An increase above the third quantile in cholera cases is expected for the period of Aug - Dec 2016 with 92% and 87% probability respectively for the two models. This alert warrants the preparedness of the public health system. We discuss the possible limitations of our approach, including variations in the impact of El Ni\~no events, and the importance of this large, warm event for further informing an early-warning system for cholera in Dhaka