Background
Although Mali remains a highly malaria-endemic country, with an estimated 7.2 million cases in 2017 and 16.9 million people living in high-risk communities [
1], the National Malaria Control Programme (NMCP) has recently demonstrated success in reducing morbidity and mortality: between 2013 and 2017 malaria deaths fell by an estimated 38% [
1,
2] and nationwide malaria prevalence in children under 5 years old (u5) dropped from 52 to 32% [
3,
4]. Helping to drive these reductions is the multifaceted National Malaria Control Strategic Plan, implemented by the NMCP with support from several partner organizations [
5]. One of the key elements of the strategic plan is malaria vector control [
6], which includes the goal of universal coverage of the population at risk (18.5 million; 100% of Mali’s population) with access to a long-lasting insecticidal net (LLIN), complimented by indoor residual spraying (IRS) in certain high-risk districts across the country. The LLIN strategy employs rolling mass LLIN distribution campaigns, organized every 3 years at a regional level, as well as routine distribution of LLINs to pregnant women and children visiting public health clinics for antenatal care and routine childhood immunization visits. The IRS strategy has been mostly implemented through a close collaboration with the US President’s Malaria Initiative (PMI) Africa Indoor Residual Spraying (AIRS) project (now the VectorLink programme).
Until 2010, the PMI AIRS project in Mali focused on spraying pyrethroid insecticides in several high-burden districts of Koulikoro and Ségou Regions. In 2011, concerns about the emergence and spread of pyrethroid resistance in the main malaria vectors of Mali (the
Anopheles gambiae species complex) prompted a switch in IRS active ingredients from pyrethroids to the carbamate insecticide bendiocarb [
7]. Beginning in 2014, the programme transitioned from using bendiocarb to a third-generation indoor residual spray product (3GIRS; insecticide formulations that are effective at controlling pyrethroid-resistant mosquitoes and have a target residual efficacy of 6 months)—a microencapsulated formulation of the organophosphate insecticide pirimiphos-methyl (PM) (Actellic
® 300CS; Syngenta A.G., Basel, Switzerland) [
8]. Recent studies using routine surveillance data from Ségou Region have shown that the IRS campaigns from 2012 to 2015, which utilized bendiocarb and PM to spray more than 200,000 structures each year, were good public health investments—protecting more than 500,000 people a year for around US$7.00 a person while simultaneously reducing indoor resting densities of
An. gambiae sensu lato (
s.l.), by around 80%, and passively reported confirmed malaria cases, by around 33% [
9,
10].
By 2015, though, national survey data showed that Mopti Region had the highest u5 malaria prevalence at 53.4%—nearly twice the national average and significantly greater than Ségou (21.9%)—despite having high access to LLINs and seasonal malaria chemoprevention (SMC) [
4]. Additionally, migration of displaced people to Mopti from northern Mali, where malaria transmission is substantially lower and acquired immunity is thought to be low, was further complicating the malaria control situation in the region [
11]. Accordingly, in 2016 a decision to shift IRS activities from Barouéli District in Ségou Region to the districts of Mopti, Bandiagara, Bankass, and Djenné in Mopti Region was made by the NMCP and other stakeholders [
12].
Here, a retrospective, observational time series analysis of monthly malaria incidence rates reported by health facilities in Ségou and Mopti Regions from January 2016 to February 2018 is used to estimate the epidemiological impact of (1) introducing IRS into parts of four previously unsprayed districts in Mopti and (2) suspending IRS operations in the Barouéli District of Ségou.
Discussion
Vector control, primarily the widespread use of LLINs and targeted use of IRS, is widely recognized as one of the primary drivers of the substantial reduction in global malaria burden that has been observed since 2000 [
1,
22] and is widely recognized as a critical component of global efforts to further control, and eventually eliminate, malaria [
23]. Nonetheless, the emergence, intensification, and rapid spread of insecticide resistance in malaria vectors (in combination with many other ecological and systemic factors) is increasing the complexity of the global malaria vector control landscape and driving the need for new tools and innovative approaches [
23,
24]. One set of questions that frequently arises in this environment asks how, when, and where to most effectively add IRS to a malaria control programme that already uses strategies to support universal coverage of populations at risk with LLINs.
The present study investigated the impact of IRS with Actellic
®300CS, a 3GIRS product, in communities of central Mali where levels of pyrethroid-only LLIN ownership (90–95% of households own at least one LLIN) and use (more than 80% of children u5 are reported to have slept under an LLIN) were reported to be high [
16], SMC coverage in the u5 population was consistent in study districts across both study years (Table
1), and the primary vector mosquitoes were resistant to pyrethroid insecticides [
9,
17]. Results show that the IRS campaigns of 2016 and 2017 were associated with reduced malaria transmission in the u5 population, clearly indicating a positive incremental impact of IRS with a non-pyrethroid insecticide used in combination with standard, pyrethroid-only LLINs. The general analysis of monthly incidence trends comparing rates of confirmed cases passively reported in the DHIS2 from IRS and non-IRS communities is compelling, suggesting protective effects ranging from 16 to 31% depending on the region sprayed and the year (Fig.
2).
The positive impact of IRS is in these communities is emphasized when looking at differences in how peak transmission rates changed from year to year as the location of IRS operations shifted from 2016 to 2017. In HFCAs of Mopti where IRS was newly introduced in 2017, malaria incidence fell almost twice as much as in comparable HFCAs that did not receive IRS in either year. Conversely, in HFCAs of Ségou where IRS was suspended prior to the 2017 spray season, malaria incidence increased substantially compared to similar HFCAs that did not receive IRS in either year, where a general trend of modestly decreasing malaria case incidence was observed (Fig.
3). Of note, these seasonal difference-in-differences results from Ségou were observed 3 years after the previous LLIN universal coverage campaign at a time when those nets were likely near the end of their target residual effectiveness. In Mopti, however, the introduction of IRS occurred in the same year as a successful universal LLIN coverage campaign. Though it is not possible to estimate the overall impact of LLINs (or how this impact may have waned over time as the nets aged) with the data analysed here, the positive impact of IRS evident in both regions suggests an incremental impact of IRS in addition to whatever baseline level of control was achieved by the pyrethroid-only LLINs also present.
The ecological studies presented here take advantage of some natural comparisons that arose as the result of normal operational decisions made by the NMCP in Mali. As such, there are many important limitations to the study, including its observational, non-randomized design. To help further ensure good comparability between intervention (IRS) and non-intervention (non-IRS) HFCAs evaluated here, IRS facilities from Ségou were compared only to non-IRS facilities also from Ségou. Similarly, intervention facilities from Mopti were compared only to non-intervention facilities also from Mopti. This step was important not only as an attempt to enhance the similarity between study arms in terms of population, socio-economic, cultural, and ecological factors, but also because direct comparisons of passive surveillance data across the two regions are challenging. Ségou and Mopti Regions have dramatically different health service utilization rates: based on outpatient visits per capita reported in the DHIS2 during this study period, a resident of Mopti was around 30% less likely to visit a health centre during the year than a resident of Ségou (2460 visits per 10,000 per year vs. 3500 visits per 10,000 per year).
Another limitation of the study is that the influx of migrants from more Northern Districts into the underlying study population of Mopti complicates the estimation of case incidence rates in those districts. Though this influx began before the study period and was not known to be skewed between IRS and non-IRS districts, an overall increase in the baseline population of Mopti from the estimates based on the 2012 census could have led to an over-estimation of crude incidence rates across the province in both years. Also, safety and accessibility were used to help prioritize which HFCAs received IRS in Mopti in 2017. It is possible that these same factors might have influenced differences in healthcare-seeking behaviours across IRS and non-IRS HFCAs. If residents of more insecure, less accessible non-IRS communities were also less likely to engage in healthcare-seeking then the case incidence rates from those HFCAs could be under-estimated, and the overall impact of IRS estimated from the routine surveillance data analysed here could also be under-estimated.
These limitations make direct comparisons of health facility case numbers between regions very difficult to interpret, and therefore comparisons of the cost-effectiveness of IRS across the regions potentially misleading. Though it is tempting to ask whether the net benefit, in terms of health system cases averted, of introducing IRS into new districts of Mopti was counterbalanced by the increase in cases observed after suspending IRS in Ségou, given these limitations this question is beyond the scope of the present analysis. What is clearly evident, however, is that IRS continues to make demonstrable positive contributions to malaria control throughout central Mali, where malaria vectors are resistant to pyrethroids, LLIN access and use are high, and access to seasonal malaria chemoprevention has rapidly expanded in recent years. Further work is ongoing to describe the impact of these IRS campaigns on key entomological indicators from the same study areas [
25], which should lead to a broader understanding of the overall effect of IRS on the reduction of malaria transmission in central Mali.
Conclusions
Annual IRS campaigns continue to make dramatic contributions to the seasonal reduction of malaria transmission in communities across central Mali, where LLIN access and use are high and SMC is also successfully co-implemented. While attempting to apply these lessons to other transmission settings, it is important to note that the successful campaigns evaluated here (1) were timed in advance of peak seasonal transmission and (2) utilized a microencapsulated product with an active ingredient that was of a different class than the one found on the LLINs used throughout the region, and to which local malaria vectors were shown to be susceptible.
Though decisions about precisely where to implement IRS as part of an integrated malaria control programme remain complex, the results presented here add to an expanding evidence base that supports the value of IRS with non-pyrethroid insecticides in communities with moderate to high transmission, abundant pyrethroid-only LLINs, and pyrethroid-resistant vectors. As such, it is easy to envision a continued role for IRS across much of Africa moving forward, though the annual sustainability of such efforts must be considered. Also, if suspension of IRS activities in a particular region is considered, strategies to help mitigate the resurgence of malaria cases that can be expected should be prioritized. In addition, exactly where, when, and how to maximize the impact and cost-effectiveness of next-generation IRS as part of integrated malaria control programmes will require continual re-evaluation, especially as the LLIN landscape rapidly changes to incorporate the use of novel, next-generation bed nets.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.