Background
Over the past two decades, Zanzibar substantially reduced malaria burden, with parasite prevalence decreasing from more than 30% in 2005 to 0.2% in 2017 [
1]. These gains followed the introduction and rapid scale-up of malaria rapid diagnostic tests (RDTs), artemisinin-based combination therapy (ACT), long-lasting insecticidal nets (LLINs) and indoor residual spraying (IRS) of households with insecticide [
2,
3]. The application of malaria control interventions in Zanzibar was aimed at achieving universal coverage and a recent impact evaluation comparing malaria incidence before and after inception of interventions suggested that this approach resulted in rapid declines of malaria [
4]. Building on these achievements, Zanzibar is now pursuing malaria elimination by 2023 by maintaining high population-level coverage of malaria interventions, as well as by reinforcing malaria surveillance to actively investigate and classify 100% of confirmed malaria cases [
5].
Numerous studies in Africa have examined the association of various demographic and behavioural factors with malaria outcomes. These studies have found that adolescent age, male gender, not sleeping under an LLIN the previous night, and the presence of fever in the 2 weeks preceding the study were positively associated with malaria infection [
6,
7]. Similarly, several studies have shown LLIN ownership and use, and/or the application of IRS to be associated with a significant protective effect against malaria infection [
6‐
9]. A recent literature review and meta-analysis of 22 studies investigating travel as a risk factor for malaria infection in sub-Saharan Africa showed that travel—within as well as to and from another country—was an important risk factor in many settings [
10].
As malaria decreases, sustainable improvements in prevention and control interventions may increasingly depend on accurate knowledge of malaria risk factors. For Zanzibar to accelerate towards malaria elimination, it is crucial to obtain detailed knowledge of risk factors associated with malaria infection. Such knowledge will help to confirm whether currently implemented preventive and control measures are effective, as well as indicate programmatic areas where additional efforts should be targeted or reinforced.
Discussion
This study investigated factors associated with malaria infection (as measured by RDT positivity) among the general population living in households where malaria cases had been detected passively when attending public or private health facilities in Zanzibar. The findings of this study suggest that age, history of fever during the last 2 weeks and history of travel outside Zanzibar during the last 30 days, and living on Unguja Island were independently associated with increased odds of RDT positivity. On the other hand, male gender, sleeping under an LLIN the previous night, higher household net access, living in a household that received IRS in the previous 12 months, as well as combining IRS in the last 12 months and sleeping under an LLIN the previous night were independently associated with reduced odds of RDT positivity.
Age was an important factor within the model, with children aged 5–14 years being at greatest risk of malaria infection. Malaria prevalence in children under 5 years of age has decreased in Tanzania in recent years [
15], and other studies in mainland Tanzania suggest the risk of malaria infection is now higher in school-aged children compared to children under 5 years of age [
6,
16]. The finding of this study might partially be explained by the association between age and those who slept under an LLIN the previous night, with a lower proportion of children aged 5–14 years (55.2%) and young adults aged 15–24 years (47.5%) having been observed to sleep under LLINs, compared to children < 5 years [
6,
9,
17]. Given their higher odds of having a positive RDT result, targeting interventions to improve LLIN access and use in children aged 5–14 years and young adults aged 15–24 years will thus be important in Zanzibar, since these age groups appear to be at greater risk of infection and are likely to contribute to on-going malaria transmission. Our analysis also showed that males had reduced odds of having a positive RDT result. Typically, an increased, gender-associated risk for a vector-borne disease such as malaria suggests gender-related differences in behaviour (e.g., night time work or leisure activities) [
18]. It is likely that the reactive case detection through HSaT was systematically missing males with increased risk of malaria because only residents present at time of DMSO visit are tested., Nonetheless, further studies are needed to investigate specific factors that could explain the higher odds of malaria RDT positivity among females observed in our study.
As reported previously [
8,
16,
19,
20], history of fever during the last 2 weeks was significantly associated with having a positive RDT result. While self-reported fever as a proxy for symptomatic malaria infection is prone to recall and/or reporting bias and thus lacks both the sensitivity and specificity, particularly in elimination settings [
20,
21], the finding suggests there is a subset of the population who do not present to health facilities when they experience malaria symptoms such as fever. Therefore, persistent messaging on recognizing malaria symptoms and prompting care-seeking behaviour at health facilities is needed even in a low transmission setting such as Zanzibar. The finding also reinforces the continued need for timely RCD, including the testing of all household members in index case households and possibly neighbouring households with or without malaria symptoms.
The results of this study confirm prior studies showing that individuals who travelled outside Zanzibar during the last 30 days had increased odds of a positive RDT result at the time of the household investigations [
10]. Malaria infection associated with travel (“malaria importation”) is frequently considered a concern in low transmission or elimination settings [
22‐
25]. There is evidence that the majority of people travelling outside Zanzibar head to malaria endemic areas on the Tanzania mainland [
23]. Therefore, implementing control measures to prevent infection among Zanzibaris visiting the mainland, as well as timely detection of travel-associated malaria cases and controlling onward transmission will be key to sustaining the low malaria burden and progressing towards malaria elimination in Zanzibar. Several approaches to limit the size and impact of travel-associated malaria could be considered, including targeting information and education messaging to potential travellers regarding the risks of malaria when travelling, malaria symptomatology, and the possible approaches for individual protection. Advice for travellers to malaria endemic locations could include recommendations to sleep under an LLIN every night, use mosquito repellents, and possibly take anti-malarial chemoprophylaxis depending on the travel destination and malaria risk. An approach to early detection for travel-associated malaria might involve screening every person arriving at key border entry points using ultra-sensitive diagnostic tests such loop-mediated isothermal amplification polymerase chain reaction, and when clinically appropriate treating or referring positive individuals. Such approach could also reinforce malaria surveillance and education on malaria prevention among travellers at these entry points.
The reduced odds of RDT positivity independently associated with LLIN access and use shown here adds to the vast body of evidence supporting the effectiveness of LLIN in their protection against malaria and other vector-borne diseases [
26]. This study also found that increased availability of LLINs in households was associated with a reduced odd of RDT positivity, with universal coverage (net access ≥ 80%) having the greatest protective effect [
27]. Nonetheless, having access to LLINs at household level is not sufficient, and approaches to encourage at-risk populations to consistently sleep under an LLIN, even in a low transmission setting like Zanzibar, are essential.
The protective effect of IRS reported in this study was within the range reported by other studies [
19,
28,
29]. The study findings suggest that IRS reduces the risk of malaria infection in the household; however, IRS operations are expensive and require considerable human resources. IRS operations in Zanzibar started in 2006, with different classes of insecticides used over the past 14 years [
30]. To mitigate programmatic costs and limit the expansion of insecticide resistance, Zanzibar has shifted from a blanket IRS approach used between 2006 and 2012 to a more targeted and focal approach from 2015 onwards, whereby only areas experiencing a high number of cases and malaria incidence are prioritized.
Combining IRS and LLINs has increasingly become a common approach to control malaria in sub-Saharan Africa. This study provides further evidence of the added benefit offered to individuals when combining IRS and LLINs [
31,
32]. While people who lived in households where IRS was applied in the last 12 months were less likely to be RDT positive compared to those who were not living in households where IRS has been applied, regardless of whether they slept under a LLIN or not, they were even less likely to test RDT positive if sleeping under an LLIN. This suggests that even in low transmission settings, the possibility of combining these two vector control interventions is effective.
This study has a number of potential limitations. Firstly, the analysis included members living in the same households as passively detected malaria cases (index cases); therefore, it is expected that this population may be at greater risk of malaria compared to the general population in Zanzibar. Secondly, only two thirds of the 30,647 passively detected cases were followed-up for HSaT; therefore, this may be a source of bias since most of the index cases lost-to-follow-up were detected during the peak malaria transmission season. Thirdly, RDTs’ have a limited sensitivity and are likely to miss a large proportion of infections, since in Zanzibar’s elimination setting most of these are sub-clinical, low-parasite density infections [
21]. The RDTs’ low sensitivity during RCD is a potential source of bias because they are designed to identify symptomatic malaria cases that are typically high-density infections. This low RDT sensitivity has broader programmatic implications, with Zanzibar possibly having to consider the use of more sensitive diagnostics (e.g., Polymerase chain reaction [PCR]), in addition to other strategies such as focal mass drug administration or seasonal malaria chemoprevention, to eliminate malaria [
3]. Finally, a fifth of the individuals with malaria test results were missing one or more explanatory variable; however, sensitivity analysis showed that this missingness was not a source of bias.
Conclusions
This study has identified risk factors associated with malaria infection (as measured by RDT positivity) that could be addressed through increased surveillance and targeted interventions. This study found that between 2012 and 2019 children aged 5–14 years, females, residents of Unguja and residents who travel outside of Zanzibar had greater odds of having a positive RDT result.
These findings suggest that even in a low transmission setting like Zanzibar, vector control remains an important malaria prevention intervention and underscores the need to maintain universal access to LLIN, persistent promotion of LLIN use, and IRS. In addition to continued effective coverage and use of vector control interventions, enhanced behavioural change and preventive strategies specifically targeting children aged 5–14 years and travellers could enhance Zanzibar’s progress towards malaria elimination.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.