Integrated malaria prevention in low- and middle-income countries: a systematic review

Abstract

Background

As many countries aim to eliminate malaria, use of comprehensive approaches targeting the mosquito vector and environment are needed. Integrated malaria prevention advocates the use of several malaria prevention measures holistically at households and in the community. The aim of this systematic review was to collate and summarize the impact of integrated malaria prevention in low- and middle-income countries on malaria burden.

Methods

Literature on integrated malaria prevention, defined as the use of two or more malaria prevention methods holistically, was searched from 1st January 2001 to 31st July 2021. The primary outcome variables were malaria incidence and prevalence, while the secondary outcome measures were human biting and entomological inoculation rates, and mosquito mortality.

Results

A total of 10,931 studies were identified by the search strategy. After screening, 57 articles were included in the review. Studies included cluster randomized controlled trials, longitudinal studies, programme evaluations, experimental hut/houses, and field trials. Various interventions were used, mainly combinations of two or three malaria prevention methods including insecticide-treated nets (ITNs), indoor residual spraying (IRS), topical repellents, insecticide sprays, microbial larvicides, and house improvements including screening, insecticide-treated wall hangings, and screening of eaves. The most common methods used in integrated malaria prevention were ITNs and IRS, followed by ITNs and topical repellents. There was reduced incidence and prevalence of malaria when multiple malaria prevention methods were used compared to single methods. Mosquito human biting and entomological inoculation rates were significantly reduced, and mosquito mortality increased in use of multiple methods compared to single interventions. However, a few studies showed mixed results or no benefits of using multiple methods to prevent malaria.

Conclusion

Use of multiple malaria prevention methods was effective in reducing malaria infection and mosquito density in comparison with single methods. Results from this systematic review can be used to inform future research, practice, policy and programming for malaria control in endemic countries.

Background

In 2020, there was an estimated 241 million cases of malaria, and 627,000 deaths from the disease globally [1]. Africa is the continent with the highest burden of malaria, accounting for 95% of all malaria cases and deaths, with children under 5 years of age and pregnant women being most vulnerable [1]. Estimates of the true burden of malaria in low-and middle-income countries (LMICs) including in Africa have been difficult to obtain due to underreporting of malaria cases and deaths [2]. The current malaria burden could thus be much higher than the estimates suggest. In addition to its impact on morbidity and mortality, the occurrence of malaria results in vast social and economic consequences. The economic costs are directly related to seeking treatment or preventive measures, or indirectly related to low productivity due to absenteeism from school or work, and time lost caring for the sick [3, 4].

Globally, malaria prevention has mainly relied on mosquito vector control by using insecticide-treated nets (ITNs), particularly long-lasting insecticidal nets (LLINs), and indoor residual spraying (IRS) [1]. In many malaria endemic countries, a vast effort has been made by ministries of health and their partners to increase coverage and utilization of ITNs and IRS. These efforts include the provision of LLINs and IRS to most at risk populations, although universal coverage for these interventions has not been achieved [5]. However, global malaria burden has not decreased significantly in recent years despite the efforts of increasing ITN and IRS coverage [1, 6]. Countries that have recorded significant gains in malaria control, such as El Savador which was certified by the World Health Organization (WHO) in 2021 as malaria free have used comprehensive approaches [1]. Although global and national malaria vector control efforts have predominantly focused on ITNs and IRS, several other control strategies can be implemented at household level to reduce mosquito density. These control measures include improving housing quality to limit mosquito entry, larval source management, and minimizing the presence of mosquitoes in houses for example by using insecticide sprays [7].

The use of appropriate combinations of non-chemical and chemical methods of malaria vector control in the context of integrated vector management has been recommended by the WHO [8]. Indeed, a combination of malaria prevention strategies has been shown to have greater impact than single methods [9,10,11]. Integrated malaria prevention therefore is an innovative approach that advocates the use of several malaria prevention measures in a holistic manner at household and community levels [12, 13]. These measures include proven malaria control methods and other approaches known to reduce mosquito populations. The specific methods advocated in the integrated approach are: (1). sleeping under LLINs; (2). installing screening in windows, vents and open eaves to prevent mosquito entry into houses; (3). IRS; (4). improving housing structure to limit mosquito entry; (5). larval source management; 6). closing windows and doors at sunset to reduce mosquito entry into houses; (7). larviciding in large water pools of stagnant water; (8). topical and spatial mosquito repellents; (9). mosquito coils; (10). insecticide sprays [12]. Although these various methods to reduce mosquito populations and prevent malaria exist, it is not expected that all of them will be used in a household due to several reasons including high cost, labour intensity, and side effects related to those that are insecticide based.

There is increasing evidence on the benefits of using multiple malaria prevention methods in households and communities particularly ITNs and IRS in comparison with single methods [14]. However, several studies have employed other malaria prevention methods in the integrated approach such as improving housing quality [15] and larviciding in recent years [16] Despite this available literature, there is limited evidence synthesizing findings from studies that have used two or more malaria prevention methods holistically. In addition, it is important to establish which other malaria prevention measures beyond ITNs and IRS have been used, and their contribution to controlling the disease. The aim of this systematic review was therefore to collate and summarize the impact of integrated malaria prevention in low- and middle-income countries on malaria burden. The systematic review adds to the evidence on use of two or more malaria prevention methods beyond the commonly used ITNs and IRS in endemic countries.

Methods

PubMed, CINAHL, Web of Science, Embase, Cochrane library, Scopus, and The Malaria in Pregnancy Consortium Library, thesis online, Google Scholar, OpenGrey, ProQuest, ClinicalTrials.Gov, PACTR registry, and World Health Organization International Clinical trials registry platform were searched for literature from 1st January 2001 to 31st July 2021. Integrated malaria prevention was defined as the use of two or more malaria prevention methods holistically. The primary outcome variables were malaria incidence and prevalence, while the secondary outcome measures were human biting and entomological inoculation rates, and mosquito mortality. This systematic review was conducted and reported in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analysis Protocols Guidelines (PRISMA-P) [17]. The review protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO), registration number CRD42021277364.

Eligibility criteria

The inclusion criteria was developed using the Population, Exposure, Comparator, Outcomes, Study characteristics framework, and studies were included if they met the following criteria:

• Study population: Individuals of all ages.

• Type of exposure: The intervention was use of integrated malaria prevention defined as the use of two or more malaria prevention methods holistically at a household or in the community [12].

• Comparator: The comparator group were households, individuals or communities not using integrated malaria prevention (using a single or no malaria prevention method at all).

• Outcomes: The primary outcome was occurrence of malaria (incidence or prevalence). Secondary outcomes were related to presence of mosquitoes in houses (including human biting rates, entomological inoculation rates, mosquito deterrence (preventing mosquito entry into houses) and mosquito densities). Definition of outcomes were as provided/defined by authors of included articles.

• Study designs: All study designs were considered.

• Context: Only studies conducted in LMICs, as defined by the World Bank Gross National Income per capita, calculated using the World Bank Atlas method as of June 2021 [18] were considered. The review included only literature published in English for a period of 20 years (January 2001 to June 2021). This period was expected to enable access to sufficient and relevant literature on integrated malaria prevention in LMICs.

• Exclusion: Duplicate publications, systematic or narrative reviews, reviews, abstracts, letters to the editor, comments, case reports, conference presentations, and study protocols were excluded.

Search strategy and information sources

Two reviewers (EA and CN) independently conducted searches in PubMed, CINAHL, Web of Science, Embase, Cochrane library, Scopus, and The Malaria in Pregnancy Consortium Library. Other sources included: thesis online, Google Scholar, OpenGrey, ProQuest, ClinicalTrials.Gov, PACTR registry, and WHO International Clinical trials registry platform. A comprehensive search strategy with key terms based on the study population, exposure, and outcomes of interest was developed in PubMed (Table 1) and adjusted to suit other databases. Controlled descriptors were used (such as MeSH terms and Boolean operators) to ensure a robust search strategy (Table 1).

Table 1 Search Strategy

The database searches were supplemented by screening the bibliographies of relevant original research articles and systematic reviews. The references of all publications identified in the primary search were inspected, and where necessary, authors of individual studies were contacted for more information or clarity.

Study screening and data extraction

The screening process was conducted at the title, abstract and full text levels by two reviewers (EA and CN) independently using defined criteria, and any discrepancies were resolved by consensus. Where necessary, the third reviewer (RN or DM) made the final decision, and all reasons for any exclusion of specific studies were documented. The study inclusion process was presented using the PRISMA flow chart (Fig. 1). Endnote reference management was used to store, organize, cite and manage all the included articles.

Fig. 1

PRISMA flow diagram showing the selection process

Data extraction was done independently by two members of the research team (EA and CN), extracting the relevant information from included full text articles onto an Excel spreadsheet, and later comparing and resolving any discrepancies. The specific data extracted included: country, design, participants/population, aim, intervention characteristics such as malaria prevention methods used, comparator group(s), duration, outcome measures, and main findings summary (Table 2). To avoid double counting, the results of studies presented in multiple papers for the same population were included once in the review.

Table 2 Characteristics of the included 57 studies

Quality or risk of bias assessment of individual studies

Two reviewers (EA and CN) independently assessed and scored the quality of selected studies. Observational studies were assessed according to the Newcastle–Ottawa scale (NOS) [19] where studies are scored between zero to nine stars for nine questions that cover three areas: selection, comparability and outcome. For randomized controlled trials (RCTs), Cochrane Collaboration’s tool (RoB tool) for assessing the risk of bias in randomized trials [20] was used. This tool provides judgment whether the study is having high, moderate, low or unclear risk of bias. Inconsistencies in the findings were resolved by discussion (EA and CN reaching consensus or by involving a third reviewer (RN or DM) where necessary).

Data synthesis and analysis

Review data was synthesized narratively while answering the review question (does use of two or more malaria prevention methods holistically at households or in the community lead to reduced occurrence of the disease (primary outcome) or presence of mosquitoes in houses (secondary outcome). Findings are descriptively presented and discussed while elaborating integrated malaria interventions and the related primary and secondary outcomes. Data are presented in tabular form for comparison, highlighting country, year of study, study objective, intervention, context, population and outcomes among others.

Results

In total, 10,931 records were identified by the search strategy from databases (n = 6652 studies) and grey literature (n = 4279). After screening titles (removing duplicates and irrelevant information), 137 studies were retained and their abstracts screened. A total of 74 articles were accessed and screened at full text level. Of these, 17 were excluded for different reasons such as not addressing the review objective, as well as being mathematical models or cost-effectiveness analyses, leading to 57 articles which were used in the review. A flow chart with details of the article screening process is shown in Fig. 1.

Characteristics of included studies

Study setting

Of the included 57 studies, majority (n = 10) were conducted in Tanzania [21,22,23,24,25,26,27,28,29,30]; followed by Kenya (n = 7) [10, 31,32,33,34,35,36]; Ethiopia (n = 5) [11, 32, 37,38,39] including one study in both Kenya and Ethiopia [32]; and Uganda (n = 5) [40,41,42,43,44]. Mozambique [45,46,47], Benin [48,49,50], Ghana [51,52,53], and India [54,55,56] had 3 studies each. Two studies were from Nigeria [57, 58], Malawi [59, 60]; and Cote d’Ivoire [61, 62]. Other articles were from Namibia [63]; Colombia [64]; Lao People’s Democratic Republic [65]; The Gambia [66]; Cambodia [67]; Brazil [68]; Bolivia [69]; Mali [70]; Island of Príncipe [71]; Cameroon [72]; Burundi [73]; Suriname [74]; and Equatorial Guinea [75].

Risk of bias

Overall, the quality of studies included in the review was generally fairly good. Most RCTs (n = 11) were of moderate risk [21, 22, 24, 27, 32, 35, 38, 49, 60, 65]; while three were of low risk [39, 66, 69] and two of high risk [47, 67]. Although several pre-post evaluation studies were scored as good quality [10, 33, 34, 41, 42, 45, 59, 68, 70,71,72,73,74], others had fair scores [11, 29, 40, 44, 48, 54,55,56, 75]. Only one quasi-experimental study was rated good [25], while others fair [23, 26, 30, 31, 50, 61, 62]. The cross-sectional surveys included [36, 37, 43, 46, 51,52,53, 57, 63] were generally rated good quality and a few (n = 2) [36, 37] were rated fair. Given the nature of the interventions, blinding of participants and evaluators was generally absent in most studies. Some information including details of intervention concealment, response rate, and use of validated measures were not clearly provided which limited objective judgement of the quality of some studies.

Study population

The studies involved different community-based interventions on malaria control that targeted general populations [10, 11, 22, 23, 25,26,27,28,29,30,31,32,33,34,35,36,37, 39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64, 66,67,68, 70, 71, 73,74,75], including those that focused on residents in controlled intervention households [21, 24, 38, 65, 69, 72], pregnant women and children [51, 58], and children care takers [57]. A few studies involved individuals of all ages [65, 72], household members older than 6 months [21, 24], while several studies reported on malaria outcomes in children of different ages including: below 6 years [34, 46, 47, 49]; 0.5–14 years [27, 28, 66]; 0.5–10 years [41, 42, 70], 2–10 years [63], 0.5–13 years [33], under 10 years [59], 2–14 years [75], 1–15 years [58], and primary school children [29, 32]. One study [62] was among the French Military troops in Côte d’Ivoire. The studies using experimental designs generally used adult volunteers.

Malaria prevention interventions

Various interventions were used, which were mostly combinations of two or three malaria prevention methods including ITN, IRS, topical repellents, insecticide sprays, microbial larvicides, and house improvements including screening, wall hangings and eaves screening. Of the intervention studies, majority (n = 22) assessed ‘ITN plus IRS’ [9, 10, 23, 27, 28, 34, 35, 39, 41, 44, 47, 48, 50, 66, 70,71,72,73,74,75,76,77], followed by the ‘ITN plus topical or spatial repellents’ (n = 9) [21, 24, 38, 54, 55, 59, 65, 67, 69]. Larval source management was another intervention used in different combinations including with: (a) ITNs [33]; (b) larviciding, LLINs, IRS, and mass drug administration [50]; (c) IRS, LLINs, intermittent preventive treatment for pregnant women, as well as early diagnosis and prompt treatment with artemisinin-based combination therapy, and larviciding [71]; (d) chemical larviciding, IRS, space spraying of insecticides at ultra-low volume, ITNs, and environmental management [29]; (e) LLINs, screening in windows and ventilators, removing mosquito breeding, and early closing of doors [40]; (f) LLINs plus larviciding with Bacillus thuringiensis var. israelensis (Bti) and community education and mobilization [32]; and (g) housing improvement [60]. Besides intervention studies, the review also included some cross-sectional surveys that analysed data regarding different malaria control combinations such as ITNs and IRS [37, 43, 46, 51, 53, 63]; ITNs and household insecticide products such as spatial repellent, smoke, and insecticide canister sprays [36, 52]; and ITNs, screening of windows, regular environmental sanitation, and insecticide sprays [57].

Study designs

The studies included cluster RCTs (n = 16) [21, 22, 24, 27, 28, 32, 35, 38, 39, 47, 49, 60, 65,66,67, 69], and evaluation of programmes/interventions (n = 17) [11, 29, 34, 40, 44, 45, 48, 54,55,56, 59, 70,71,72,73,74,75]. Other studies included national Demographic and Health Survey [63]; panel data from administrative districts [52]; hospital data [43, 51]; and other surveys [36, 37, 46, 53, 57, 58]. Experimental hut/houses and field trials [23, 25, 26, 30, 31, 50, 61, 62, 68] were also included.

Primary outcomes

Malaria incidence

A total of 24 studies reported malaria incidence alone [10, 21, 33, 35, 39, 41,42,43, 45, 47,48,49, 54, 63,64,65,66, 68,69,70,71,72,73,74], with most (14) indicating significantly reduced incidence in integrated interventions compared to single ones [10, 33, 41, 43, 47, 48, 54, 63, 64, 68,69,70, 73, 74]. Most studies used IRS plus ITN combinations [10, 41, 43, 47, 48, 63, 70], some of which were regarded as cost-effective in urban areas [48], while others significantly reduced malaria in all-age groups (61%) [10] or among children aged 2–10 years in high transmission areas [47, 63, 70]. Two studies using IRS plus ITNs showed that non-pyrethroid IRS, such as pirimiphos-methyl was more effective than pyrethroids for IRS in areas with widespread pyrethroid resistance (incidence rate of 2.7 per 100 person-months in the intervention area compared with 6.8 per 100 person-month in the control area) [70] and in highly endemic areas (incidence rate of 3532 per 10,000 children-month in intervention group versus incidence rate of 4297 per 10,000 children-month in the control group [47]. A combination of several interventions including filling ditches, larval source management, education, as well as strengthened diagnosis and treatment mechanisms reduced cerebral malaria (from 90 to 110 to zero cases annually); and malaria incidence by 45% [64]. A combination of nets with a plant-based insect topical repellent significantly reduced episodes of fever including 80% reduction in Plasmodium vivax episodes in comparison to nets alone [69], and microbial larvicides plus ITNs reduced the risk of acquiring new infection in 1.5–13 year olds in comparison to ITNs alone [33].

However, no significant incidence reductions in the integrated approach were reported in 10 studies [21, 35, 39, 42, 45, 49, 65, 66, 71, 72] including nets plus topical repellent [21, 65]; LLINs plus IRS or carbamate-treated plastic sheeting plus LLINs combination [49]. Weekly larviciding of stagnant water bodies combined with nets, IRS, and mass drug administration showed modest reduction in human infections [35]. In another study, a significant reduction in all-age incidence was recorded, and over 76.7% of expected cases were averted. However, effects on malaria prevalence were varied and not associated with expected level or trend changes in an area with significant pyrethroid resistance [45].

Malaria prevalence

Of 22 studies [11, 27,28,29, 32, 34, 36,37,38, 40, 46, 51,52,53, 56,57,58,59,60, 67, 73, 75, 76] reporting prevalence alone, most (15) showed more benefits in the integrated approach [11, 28, 29, 36,37,38, 40, 46, 51,52,53, 56, 57, 59, 73]. These included combination of daily application of topical repellents during the evening plus use of LLINs at community level [38]; as well as spatial repellent devices plus LLINs in endemic villages [59]. Compared to single or no intervention at all, IRS plus LLIN combinations [28, 37, 46, 73, 75] specifically reduced malaria in all ages [37] and children between 0.5 and 14 years [28, 46, 73, 75] including across a range of transmission intensities and net utilization levels (76). Some large-scale multiple intervention combinations of IRS plus ITNs plus behaviour change/communication strategies reduced prevalence with greater impact in high-risk areas [73], as well as among females [52] and children under 5 years [51,52,53]. Prevalence reductions in children under 5 years were also related to LLINs mass campaigns alongside other anti-malarial interventions which reduced malaria cases (by 50%) and deaths (by 65%) [51]. Other combinations of multiple interventions including community-based education promoting integrated vector management plus environmental management, larviciding plus LLINs and IRS [11]; early detection and prompt treatment plus larvivorous fishes [56]; and vector control, rapid diagnosis and treatment, and community health education [29] also significantly reduced prevalence in comparison with single methods.

However, mixed (n = 1) [75] and non-significant effects (n = 6) [27, 32, 34, 58, 60, 67] of integrated approaches on malaria prevalence were also reported. For instance, adding ITNs to IRS had a significant impact at baseline and immediately after the first round of IRS, but showed limited effects after the second round when the IRS impact was strongest [34]. Similar mixed results were noted among primary school children where combinations of three interventions (LLINs plus Bti plus community education and mobilization arm, compared to the LLINs only arm, LLINs plus Bti arm, and LLINs plus community education and mobilization arm) were used in a low prevalence setting, but not at sites with relatively higher prevalence [32]. In another study, there was no significant contribution of community-based house improvement and/or larval source management to reductions in malaria prevalence beyond the reductions provided by the mass ITN distribution, community engagement programme and other national malaria control interventions [60]. Furthermore, a topical mosquito repellent (picaridin) plus LLINs versus LLINs group alone showed no significant differences in Plasmodium prevalence [67], similar to results of insecticide sprays plus ITNs in another study [58]. One study [75] showed mixed effects of LLINs and IRS combination, with a rapid decline of IRS insecticidal effectiveness 3 months following spraying posing considerable operational concerns given that malaria transmission occurred throughout the year.

Secondary outcomes

Human biting and entomological inoculation rates

Human biting rates (HBRs) [30, 41, 44, 68], and entomological inoculation rates (EIRs) [22, 44, 60, 68], were significantly reduced in the integrated approach compared to single interventions. For example, ITNs plus IRS combinations were associated with a significant decrease in: Anopheles darlingi HBRs trend; EIR from total anophelines; Anopheles darlingi and Anopheles albitarsis [68]; HBR of female Anopheles mosquitoes [41, 44]; and annual EIRs decline [44]. Similarly, a study of carbamate IRS plus ITNs produced major reduction in EIRs compared to ITNs alone in an area of moderate coverage of LLINs and high pyrethroid resistance in Anopheles gambiae sensu stricto [22]. Compared to LLINs with untreated baskets [30], transfluthrin-treated baskets combined with LLINs reduced the proportion of An. arabiensis mosquito bites by more than three quarters, and Anopheles funestus sensu lato (s.l.) mosquitoes bites by nearly half [30]. However in another study, reductions in EIR over the full trial period did not significantly differ between the four trial arms (control, house improvement, larval source management, nor house improvement and larval source management) [60].

Mosquito deterrence and mortality

Mosquito deterrence (related to preventing mosquito entry into houses) [23, 30, 50] and mosquito mortality [23, 25, 30, 50, 61] were also evaluated in some studies, indicating benefits of using a combination of methods compared to single ones. For example, sisal decorative baskets treated with transfluthrin repellents induced a tenfold increase in 24-h mortality of Anopheles arabiensis mosquitoes, providing additional household and personal protection against indoor biting malaria and nuisance mosquitoes in the early evening [30]. This combination intervention also deterred three-quarters of An. arabiensis mosquitoes from entering huts in comparison with untreated nets and IRS alone [30]. In one study, indoor mosquito traps in combination with LLINs enhanced mortality of pyrethroid-resistant An. gambiae compared to single interventions [61]. In another study, use of chlorfenapyr and alpha-cypermethrin together as a mixture on nets or a combined chlorfenapyr IRS and pyrethroid LLIN intervention provided better deterrence of An. gambiae s.l. and induced significantly higher levels of mortality of pyrethroid-resistant malaria vectors [50].

Mosquito densities

Mosquito densities [22, 24, 32, 35, 37, 55, 60, 66] generally reduced significantly in the combined approach than the reference groups [22, 24, 26, 37, 55]. The ITN plus IRS arm in one study was associated with significant reduction in overall mosquito density [37], and mean An. gambiae s.l. density [22] compared to control groups. In another study, LLINs in combination with topical repellents where everybody received 15% N,N-Diethyl-meta-toluamide (DEET) had resting mosquito densities fewer than half that of households in the placebo scenario [24]. Similarly, the total anopheline density, including Anopheles dirus, Anopheles minimus and Anopheles philippinensis, in houses in mosquito nets plus topical repellent arm declined in comparison with other arms (ITNs, topical repellent, and no treatment) [55]. A series of preliminary experiments evaluating eave tubes indicated that installing them plus screening following introduction of LLINs in a model village reduced larval density greatly compared to pre intervention values, and virtually eliminated indoor host-seeking mosquitoes [26].

On the other hand, there were no significant differences in mosquito density in some studies [32, 35, 60, 66]. Specifically, 3 years of sampling in trial arms (house improvement, larval source management, house improvement and larval source management) [60], and combining LLINs with larviciding with Bti plus community education and mobilization showed no significant differences in reduction of adult anopheline density in each of the groups [32]. In addition, despite the high coverage of interventions targeting malaria hotspots, no statistically significant difference was observed in the mosquito densities of female anophelines between the intervention (larviciding, LLINs plus IRS plus focal mass drug administration) arms and control clusters (standard national programme) [35].

Discussion

This systematic review synthesized evidence on integrated malaria prevention and its effectiveness in controlling the disease in LMICs. The study found various combinations of prevention methods used in malaria control across countries, with ITNs and IRS the most utilized, followed by various combinations including topical repellents, environmental management, and larviciding. The use of several malaria prevention measures in combination was effective in reducing both malaria incidence and prevalence in several studies in comparison with single methods. The review also indicated a reduction in human biting and entomological inoculation rates and mosquito densities, as well as an increase in mosquito deterrence and mortality following implementation of a combination of malaria prevention methods. The improved malaria outcomes related to occurrence of the disease and mosquito abundance in the studies can be related to the synergistic effect of combining several methods to prevent the disease. However, some studies found no additional benefits of using combinations of several malaria prevention measures in comparison with single interventions. Overall, these findings from the systematic review show promise in the use of multiple malaria prevention methods holistically to complement existing strategies in endemic countries striving for malaria elimination [1]. Such evidence is needed to contribute to the WHO global vector control response particularly for malaria, a leading cause of morbidity and mortality in many LMICs [77].

The review found that except in a few studies where the effect was modest or mixed [34, 49], ITNs and IRS combinations reported reduced malaria incidence and prevalence [10, 28, 37, 41, 43, 46,47,48, 63, 70, 73, 75]. Indeed, a systematic review on the effect of adding IRS to communities using ITNs established reduced prevalence of malaria [78]. When ITNs and IRS are used in combination, their synergistic effect is likely to be greater as ITNs would be most effective against vectors that primarily feed late at night while IRS would be most effective against vectors that spend much of their adult lives resting inside houses [79]. Indeed, these two methods are most effective in areas where predominant mosquito vectors are both strongly endophagic and endophilic [34]. Other research has demonstrated that the combined effect of ITNs and IRS could depend on the levels of malaria transmission in the area. A nationally representative survey from 17 sub-Saharan African countries using the two methods indicated that intervention effects varied across malaria transmission levels. Indeed, ITNs were associated with a significant reduction in malaria morbidity in high and medium transmission settings, while IRS appeared to be most effective in medium and low transmission areas [9]. In a related review, integrated malaria prevention using ITNs and IRS contributed to a reduction in malaria incidence but had little impact on prevalence dependent on transmission levels [80]. In this study, the use of both interventions together showed more protection than each intervention on its own especially in medium transmission settings. Therefore, whereas existing evidence predominantly demonstrates additional benefits of combining ITNs and IRS, malaria transmission levels in the target areas need to be considered while planning such interventions. In addition, optimal coverage of either ITNs or IRS should be prioritized before introduction of another malaria prevention intervention. Indeed, the WHO has discouraged the use of the second intervention to compensate for deficiencies in the first malaria control method [81].

Beyond use of ITNs and IRS, the systematic review shows reduced incidence and prevalence of malaria while using a combination of other methods. There was reduced malaria incidence/prevalence while using the following combinations: environmental management, health education and case management [64]; ITNs and repellents (topical and spatial) [38, 59, 69]; larviciding and ITNs [33]; health education and environmental management [11]; and case management and health education [29]. However, many of these methods such as larviciding and environmental management, that have been known to reduce breeding of mosquitoes for many years, have been largely ignored in many endemic communities [82]. Indeed, the use of other malaria prevention methods in many LMICs beyond ITNs and IRS is minimal despite the WHO recommending the use of a mix of chemical and non-chemical measures to prevent the occurrence of the disease [83]. The non-core malaria prevention methods should be explored to complement existing strategies. Such approaches could offer significant benefits despite related challenges such as high cost of repellents [84]; environmental management being cumbersome [85]; and larviciding being labour intensive and expensive [86]. In promoting integrated malaria prevention beyond ITNs and IRS, barriers and facilitators of use of the different measures in LMICs should be considered. In addition, the current evidence for each of the malaria prevention methods should inform their use in particular settings. For example, larviciding has been recommended to supplement ITNs and IRS in areas where aquatic habitats are few, fixed and findable, whereas more robust research is needed on environmental management as a strategy for malaria control [81].

The review findings show that use of some multiple malaria prevention methods led to reduced human biting [30, 41, 44, 68] and entomological inoculation rates [22, 44, 60, 68], as well as an increase in mosquito deterrence [23, 30, 50] and mosquito mortality [23, 25, 30, 50, 61] in comparison with single interventions. These findings could partly explain the mechanism of reduced malaria occurrence observed in this systematic review. Indeed, reduced presence and biting of mosquitoes indoors for example due to improved housing is directly related to low malaria transmission [87]. In addition, some of the methods predominantly used in the studies included in the review such as ITNs provide protection against biting from mosquitoes [81]. Whereas some malaria prevention methods such as ITNs and IRS are used indoors, many of the other interventions that were being used in the integrated approach such as environmental management, larviciding and housing improvement target mosquito populations before entering houses. Therefore, these methods directly contribute to lower numbers of mosquitoes indoors which could result in reduction in the occurrence of malaria. Although ITNs protect users while sleeping, evidence suggests that mosquitoes may bite hence transmit malaria before one goes to bed or outdoors in many endemic countries [88, 89]. This occurrence emphasizes the need to explore interventions that not only target mosquitoes indoors but also those that reduce mosquito breeding and prevent their entry into houses as advocated in integrated malaria prevention. It is also important to note that effectiveness of ITNs and IRS is dependent on the ability of mosquito vectors coming into contact with the insecticides, as well as susceptibility to the insecticides used. Overall, the varied effects in different studies observed in the review demonstrate that appropriate combinations that are not “one size fits all” should be recommended in consideration of the complex and dynamic nature of mosquito populations, insecticide resistance patterns, local epidemiology, and the operational effectiveness of malaria control interventions [9, 41].

Some studies in the systematic review showed mixed results or no benefits of using multiple approaches to prevent malaria at households and in communities [21, 27, 32, 34, 35, 39, 42, 45, 49, 58, 60, 65,66,67, 71, 72, 75]. Although there were no observed differences in these studies compared to those that demonstrated benefits of using multiple methods, several factors could potentially explain these findings from the review. First of all, the effects of certain malaria prevention methods such as ITNs and IRS which dominated in the various studies are likely influenced by the behaviour and resistance status of the primary malaria vectors. Indeed, mosquitoes could develop resistance to certain insecticides leading to reduced efficacy of deployed methods [82, 83]. In addition, while innovative mixes of interventions could achieve large reductions in disease burden [45], there are concerns about the short residual life of the insecticides used in IRS and the related need for additional rounds of spraying. These concerns could pose cost-effectiveness issues, excessive demand on the spray programme, and households’ non-compliance with re-spraying of their houses [75]. Acceptability and efficacy challenges have also been reported regarding IRS use in some communities [84] hence could explain some of the non-significant results.

It is also worth noting that there is need for more evidence on the effectiveness of some of the methods in malaria prevention such as spatial repellents, larvivorous fish, and larval habitat modification as recommended by the WHO [81]. Future studies on individual and combinations of various prevention methods (exploring possible synergistic effects) are needed to add to the much needed literature on malaria control beyond common interventions such as ITNs and IRS. Such research should be context specific, considering how various attributes such as mosquito density and behaviour, existing evidence on proposed designs, and coverage of other malaria prevention approaches impact the mosquito or disease related outcomes. For instance, some studies highlight the importance of net coverage in determining the effect of the IRS plus ITNs combination [27], and that high LLIN coverage is sufficient to protect people against malaria in areas of low or moderate transmission. Findings indicate that where ITN coverage is low, additional control with IRS could be needed [49, 66] as IRS is considered a secondary measure, and only crucial when ITNs have not been effective [23]. Indeed, in settings where ITN coverage is optimal, the addition of IRS may add minimal benefit in reducing malaria morbidity and mortality [81]. There are also concerns regarding the fidelity of some interventions in the integrated approach such as environmental management being cumbersome [85] which could also have influenced the findings. More evidence is therefore needed to further explain why combinations of certain malaria prevention methods may not be as effective in certain studies and contexts.

A limitation of this review is that outcomes in the included studies were measured differently which may have affected the results. This review also considered only articles published in English which could have led to publication bias. In addition, having included studies from only LMICs could have omitted interesting findings from other countries. Furthermore, there were few RCTs and many studies were generally of fair quality which can impact the level of evidence obtained. Nevertheless, this review is the first to synthesize evidence on integrated malaria prevention which should contribute to malaria control efforts in endemic countries. The 20-year period used in the review provided a sufficient duration to include not only recent studies but also evidence on combinations of various malaria prevention methods used many years ago. However, it is recognized that there could have been changes in malaria control recommendations and guidelines during this period which could have influenced the quality and extent of interventions used at various times and resultant findings.

Conclusions

Use of multiple malaria prevention methods was effective in reducing malaria incidence, prevalence, human biting and entomological inoculation rates, as well as increasing mosquito deterrence and mortality in comparison with single methods. However, a few studies showed mixed results or no benefits on using multiple approaches to prevent malaria. More evidence is needed on the effectiveness of some malaria prevention methods for malaria control used individually or in combination. Results from this systematic review could inform future research, as well as practice, policy and programming on integrated malaria prevention in endemic countries to add to existing national and global control efforts.