The Epidemiology of Cholera in Zanzibar: Implications for the Zanzibar Comprehensive Cholera Elimination Plan

Abstract

Background: Cholera poses a public health and economic threat to Zanzibar. Detailed epidemiologic analyses are needed to inform a multisectoral cholera elimination plan currently under development.

Methods: We collated passive surveillance data from 1997 to 2017 and calculated the outbreak-specific and cumulative incidence of suspected cholera per shehia (neighborhood). We explored the variability in shehia-specific relative cholera risk and explored the predictive power of targeting intervention at shehias based on historical incidence. Using flexible regression models, we estimated cholera's seasonality and the relationship between rainfall and cholera transmission.

Results: From 1997 and 2017, 11921 suspected cholera cases were reported across 87% of Zanzibar's shehias, representing an average incidence rate of 4.4 per 10000/year. The geographic distribution of cases across outbreaks was variable, although a number of high-burden areas were identified. Outbreaks were highly seasonal with 2 high-risk periods corresponding to the annual rainy seasons.

Conclusions: Shehia-targeted interventions should be complemented with island-wide cholera prevention activities given the spatial variability in cholera risk from outbreak to outbreak. In-depth risk factor analyses should be conducted in the high-burden shehias. The seasonal nature of cholera provides annual windows of opportunity for cholera preparedness activities.

Figure 1.

Maps with annual incidence of suspected cholera by shehia over study period. A and B include data from the full study period from 1997 to 2017. C and D illustrate data from 2012 to 2017. Shehias with annual incidence of 20 per 10000 and above have their names shown on the maps.

Figure 2.

Time series of suspected cholera cases and rainfall (weekly [A]) and estimated seasonality of cholera reports (B). Gray bars in A indicate period where only biweekly data from part of Zanzibar were available (part of vaccine effectiveness study [14]). In B, the black line represents estimate of risk of reporting cholera cases by week compared with the average risk across the year, with gray bands representing 95% confidence intervals. Blue bars at the bottom of plots represent the weekly rainfall, with darker shades indicating higher values.

Figure 3.

Stability and predictability of shehia-level cholera risk. A and B show relationship between relative cholera risk in a single outbreak in each shehia and the mean relative cholera risk in that shehia in all other years. Each dot represents 1 shehia in 1 outbreak with the red dots representing the priority shehias (ie, those that contributed the first 50% of total cases since 1997 [per island] when ordered by cumulative incidence). C and D illustrate the potential efficiency of cholera intervention targeting strategies where shehias are prioritized/ranked based on historic cumulative incidence data. The black line represents the full data ordered by overall cumulative incidence of each shehia (1997–2017). The gray lines illustrate the efficiency of targeting or prioritizing shehias in each outbreak, based on historic cumulative incidence, leaving out data from that specific outbreak. Each horizontal segment represents the addition of a single shehia. Red labels indicate priority shehias on each island. (NOTE: Only some priority areas are labeled due to space.)

Figure 4.

Distribution of mean outbreak attack rate in each shehia by number of outbreaks when cases were reported in each shehia. Priority shehias are shown in red as they are in Figure 3. Triangles represent shehias that have recently reported cases (since 2015), whereas circles represent areas that have not recently reported cases. All high-risk shehias in Pemba are labeled, whereas only the top 10 priority shehias in Unguja are labeled.