Spatiotemporal dynamics of malaria in Zanzibar, 2015-2020

Abstract

Background: Despite high coverage of malaria interventions, malaria elimination in Zanzibar remains elusive, with the annual number of cases increasing gradually over the last 3 years.

Objective: The aims of the study were to (1) assess the spatiotemporal dynamics of malaria in Zanzibar between 2015 and 2020 and (2) identify malaria hotspots that would allow Zanzibar to develop an epidemiological stratification for more effective and granular intervention targeting.

Methods: In this study, we analysed data routinely collected by Zanzibar's Malaria Case Notification (MCN) system. The system collects sociodemographic and epidemiological data from all malaria cases. Cases are passively detected at health facilities (ie, primary index cases) and through case follow-up and reactive case detection (ie, secondary cases). Analyses were performed to identify the spatial heterogeneity of case reporting at shehia (ward) level during transmission seasons.

Results: From 1 January 2015 to 30 April 2020, the MCN system reported 22 686 index cases. Number of cases reported showed a declining trends from 2015 to 2016, followed by an increase from 2017 to 2020. More than 40% of cases had a travel history outside Zanzibar in the month prior to testing positive for malaria. The proportion of followed up index cases was approximately 70% for all years. Out of 387 shehias, 79 (20.4%) were identified as malaria hotspots in any given year; these hotspots reported 52% of all index cases during the study period. Of the 79 hotspot shehias, 12 were hotspots in more than 4 years, that is, considered temporally stable, reporting 14.5% of all index cases.

Conclusions: Our findings confirm that the scale-up of malaria interventions has greatly reduced malaria transmission in Zanzibar since 2006. Analyses identified hotspots, some of which were stable across multiple years. Malaria efforts should progress from a universal intervention coverage approach to an approach that is more tailored to a select number of hotspot shehias.

Keywords: Epidemiology; Malaria; Public Health.

Figure 1

Geographical location of Zanzibar Archipelago, including the main islands Pemba (northern island) and Unguja (southern island).

Figure 2

Number of confirmed malaria cases per shehia from 1 January 2015 to 30 April 2020. Each row of the plot represents the time series of reported index cases per week per shehia. The shehias were grouped by district ordered by latitude (from North to South). The image shows the cumulative number of index cases per year split by travel history (top; horizontal) and cumulative number of cases per shehia during the 2015–2020 study period (right; vertical). The numbers on the bottom x-axis represent months.

Figure 3

Weekly rainfall and reported malaria index cases in Zanzibar from 1 January 2015 to 30 April 2020. The figure also shows the number and fraction of index cases with travel history outside Zanzibar (yellow bars). The grey boxes indicate the ten high transmission seasons of the study period; transmission seasons were classified using Roman numerals. The numbers on the bottom x-axis represent months.

Figure 4

Cross-correlation of weekly reported index cases and weekly precipitation. The black lines passing the red dotted lines are significant correlations, p<0.05.

Figure 5

Results of Kendall’s agreement test comparing hot-spot spatial pattern of all index cases, index cases with no travel history outside and inside Zanzibar (Znz), and index cases with travel history outside Znz per transmission season. Orange dots represent masika transmission seasons while blue dots represent vuli transmission seasons.

Figure 6

Number of years in which each shehia was identified as hotspot of all reported index cases during peak transmission seasons following the masika (A) and vuli (B) rainy seasons from 1 January 2015 to 30 April 2020. Non-highlighted shehias were not identified as hotspot during the study period as per G_i* local spatial clustering test. The maps show Zanzibar’s major islands: Pemba (northern island) and Unguja (southern island).

Figure 7

Number of years in which each shehia was identified hotspot of reported index cases with no travel history outside or inside Zanzibar peak transmission seasons following the masika (A) and vuli (B) rainy seasons from 1 January 2015 to 30 April 2020. Non-highlighted shehias were not identified as hotspot during the study period as G_i* local spatial clustering test. The maps show Zanzibar’s major islands: Pemba (northern island) and Unguja (southern island).

Figure 8

Number of years in which each shehia was identified as hotspot of reported index cases with travel history outside Zanzibar during peak transmission seasons following the masika (A) and vuli (B) rainy seasons from 1 January 2015 to 30 April 2020. Non-highlighted shehias were not identified as hotspot during the study period as per G_i* local spatial clustering test. The maps show Zanzibar’s major islands: Pemba (northern island) and Unguja (southern island).