Research Article: Increase in Reported Cholera Cases in Haiti Following Hurricane Matthew: An Interrupted Time Series Model

Date Published: February 26, 2019

Publisher: The American Society of Tropical Medicine and Hygiene

Author(s): Erin Hulland, Saleena Subaiya, Katilla Pierre, Nickolson Barthelemy, Jean Samuel Pierre, Amber Dismer, Stanley Juin, David Fitter, Joan Brunkard.

http://doi.org/10.4269/ajtmh.17-0964

Abstract

Matthew, a category 4 hurricane, struck Haiti on October 4, 2016, causing widespread flooding and damage to buildings and crops, and resulted in many deaths. The damage caused by Matthew raised concerns of increased cholera transmission particularly in Sud and Grand’Anse departments, regions which were hit most heavily by the storm. To evaluate the change in reported cholera cases following Hurricane Matthew on reported cholera cases, we used interrupted time series regression models of daily reported cholera cases, controlling for the impact of both rainfall, following a 4-week lag, and seasonality, from 2013 through 2016. Our results indicate a significant increase in reported cholera cases after Matthew, suggesting that the storm resulted in an immediate surge in suspect cases, and a decline in reported cholera cases in the 46-day post-storm period, after controlling for rainfall and seasonality. Regression models stratified by the department indicate that the impact of the hurricane was regional, with larger surges in the two most highly storm-affected departments: Sud and Grand’Anse. These models were able to provide input to the Ministry of Health in Haiti on the national and regional impact of Hurricane Matthew and, with further development, could provide the flexibility of use in other emergency situations. This article highlights the need for continued cholera prevention and control efforts, particularly in the wake of natural disasters such as hurricanes, and the continued need for intensive cholera surveillance nationally.

Partial Text

In October 2010, 10 months after the devastating earthquake in metropolitan Port-au-Prince in January, the first cholera cases, later confirmed as Vibrio cholera O1 (serotype Ogawa, biotype El Tor), were reported in Haiti in the Artibonite and Centre administrative departments.1 In the next 2 years, > 600,000 cholera cases (suspect and laboratory-confirmed) were reported through the Ministry of Public Health and Population’s (French Acronym MSPP) National Cholera Surveillance System (NCSS).1–3 In 2016, 41,421 cholera cases were reported to NCSS.3 Although the MSPP’s Directorate of Epidemiology, Laboratory, and Research, which is responsible for NCSS, has reported a reduction in the number of cases in recent years, cholera continues to be reported daily from almost all of Haiti’s 10 departments and transmission has recently been considered endemic in Haiti.4–6

The impact of rainfall and cholera incidence rate in the 46-day period following Hurricane Matthew can be seen in Figure 2, with the highest levels of rainfall in the Sud-Est, Nippes, and Grand’Anse departments and highest cholera incidence rates occurring in the Sud and Grand’Anse. A total of 133,092 suspect cholera cases in persons aged 5 years and older were reported between January 1, 2013, to November 19, 2016, of which 6,365 (5%) occurred in the 46-day period following Hurricane Matthew. Average daily cholera case counts in the pre- and post-hurricane period can be seen in Table 1, as well as average daily rainfall and average daily cholera mortality.

These models demonstrate that Hurricane Matthew had a significant, immediate impact on reported suspect cholera cases among persons aged 5 years and older; suspect cholera cases increased both at the national level and in the most highly affected departments. These models also suggest that in most of the affected departments, and on a national scale, the suspect cholera cases decreased significantly over the 46-day period following Hurricane Matthew after the observed immediate significant increase, demonstrated by the negative estimates for the level + slope change interaction term. Although not exclusive to Sud and Grand’Anse, the finding that the significant increases in suspected cholera case counts following Hurricane Matthew were not uniform in all departments provides evidence that Hurricane Matthew did have an impact on those highly affected departments or those that have a longstanding history of cholera treatment and transmission, including Centre, a department known for having an established referral hospital and experience with cholera treatment, notable increases in case counts during the rainy season, and high attack rates compared with the rest of the country.22,23 Possible reasons for the significant decreasing slopes following the immediate increase in case counts include general seasonal trends and decreased rainfall, heightened reporting in the immediate post-storm period that gradually leveled off, or quick response to the emergency preventing further cases and treating those identified.

The national ITS regression model developed in this study suggests an immediate increase in reported cholera cases following the hurricane followed by a significant decline in the post-hurricane period; departmental models further demonstrated that this trend was not uniform for the country and that those more highly impacted departments had significantly higher surges in reported cases following the hurricane. These findings demonstrate that there is a need for continued intensive surveillance following a hurricane as increased cholera transmission is likely. Further research could be beneficial to develop similar models for other infectious or vector-borne diseases following a hurricane or modified for other emergencies. These results of this study revealed that ITS regression models can be developed and implemented relatively quickly in the wake of a natural disaster such as a hurricane and should be considered as part of disaster response to identify where disease transmission was significantly increased to help prevent lives lost and reduce the burden of communicable diseases.

 

Source:

http://doi.org/10.4269/ajtmh.17-0964

 

0 0 vote
Article Rating
Subscribe
Notify of
guest
0 Comments
Inline Feedbacks
View all comments