This study represents the first multi-country analysis of the combined effectiveness of ITNs and IRS on parasitaemia and under-five mortality under routine conditions. The effect of these interventions was found to vary across malaria transmission levels, such that ITNs are associated with a significant reduction in malaria morbidity in high and medium transmission settings, while IRS appears to be most effective in medium and low transmission areas. The use of both interventions together shows more protection than each intervention on its own; particularly in medium transmission settings, the present study’s results demonstrate a synergistic effect of ITNs and IRS. No statistically significant effects were found for the interventions on child mortality; however, this is likely largely due to the small sample sizes and the very small observed number of deaths in each intervention category.
It is particularly interesting that the risk of parasitaemia was significantly reduced when both interventions were used jointly in areas of medium and high transmission, 53% (95% CI 37% to 67%) and 31% (95% CI 11% to 47%) correspondingly. Further, an additional 34% (95% CI 7% to 53%) effect against parasitaemia was accrued by having both ITNs and IRS compared to the sum of the individual protectiveness provided by ITNs and IRS in medium transmission areas. When the analysis was conducted stratified by location, risk reduction for parasitaemia was also significant for children receiving both interventions in rural and urban areas alike, 57% (95% CI 48% to 65%) and 39% (95% CI 10% to 61%), respectively. The additional protectiveness of having both ITNs and IRS, as compared to each intervention on its own, was not statistically significant in this analysis.
The results for combined intervention approach were less conclusive in low transmission areas. It is likely that the analysis was underpowered to detect significant effects if they existed. For example, under low transmission settings, only 22 children tested positive for malaria parasites for the combined intervention variable. While this finding aligns with this setting's transmission classification (i e, relatively low malaria transmission risk), having so few observations for the health outcome of interest makes detecting a significant effect, if it exists, very difficult. The 95% confidence intervals are quite wide and overlap substantially across intervention categories, further demonstrating the uncertainty in the findings of the present study in low transmission settings. Further, it is possible that other types of interventions are used and that the ways in which ITNs and IRS are deployed in low transmission settings vary (e g, active case detection, focal application of IRS) from the ways that ITNs and IRS are used for areas with greater malaria transmission . Subsequently, the analytical focus of the present study may not have been optimally aligned with the interventions most widely used in areas with low malaria transmission.
For the mortality analyses, it is likely that the small sample sizes of children who had both interventions limit the power of the analysis to detect a significant effect, if one existed. Under high transmission settings, for example, the 95% confidence intervals for relative risk reduction associated with having both interventions (ITNs and IRS) ranged from −67% to 79%, reflecting the very small number of child deaths that occurred during joint intervention exposure (n = 4). Similarly, very few child deaths were recorded under IRS only exposure across all malaria risk categories, ranging from three to thirteen fatalities in high and medium transmission areas, respectively. It is thus important for future studies to include additional survey data or consider alternate data sources for analysing the combined effect of ITNs and IRS, as well as the singular effect of IRS, on child mortality.
In many ways, it is not surprising to find relatively small sample sizes for the joint use of ITNs and IRS, regardless of transmission risk or urbanicity. Few malaria control programmes and development agencies in sub-Saharan Africa have actually scaled-up the coverage of both ITNs and IRS. This tendency to invest in one or the other of two interventions, rather than both, likely stems from financial and logistical constraints, as well as the lack of scientific evidence supporting a combined approach . On the other hand, much more data have been collected on the effectiveness of exposure to ITNs and IRS individually. There is substantial evidence both from pooled randomized-control trials (RCTs)  and from observational studies that use and ownership of ITNs results in reductions in child mortality and parasitaemia [32, 41, 42]. Less data are available on the effectiveness of IRS ; however, a meta-regression analysis across a range of study types recently showed a 62% (95% CI 54% to 69%) reduction in malaria prevalence . The present study’s findings show that comparable protectiveness against malaria morbidity can be achieved under routine conditions with IRS in areas with low malaria transmission, 66% (95% CI 17% to 86%), as well as in rural areas, 47% (95% CI 31% to 55%). With little existing evidence systematically demonstrating the association between IRS application and mortality reduction , further research is needed, especially as more country programmes expand their IRS operations.
The idea that the joint exposure to ITNs and IRS could provide significantly more protection than either intervention alone is not a new one , but little was known about the effects of these interventions when used together under routine settings. Until now, no multi-country analysis had previously shown significant results in favour of the combined use of ITNs and IRS against malaria morbidity. The operational and biological mechanisms by which the combined use of ITNs and IRS may provide greater protection than each intervention alone is supported by field studies and modelling exercises. ITNs provide physical protection against mosquitoes and malaria transmission, which is largely conferred to the individual using the net. These nets’ insecticidal properties further deter transmission to the individual, but may also help prevent continued transmission to other household members and nearby community members. IRS is less directly protective of any given individual but garners greater protection to larger groups of people, repelling malaria-transmitting vectors from entering households in the first place or killing the mosquitoes if they rest post-feeding on recently sprayed walls. Subsequently, the combined deployment of ITNs and IRS targets mosquitoes at multiple, complementary transmission points and is likely to most effectively minimize the number of opportunities for malaria vectors to reach any given individual than a singular intervention . Also, when used in combination, the insecticidal protection provided by ITNs and IRS may last longer than when only one insecticide-based intervention is used . Based on results from two areas in Mozambique, which would be classified as medium malaria transmission risk , Kleinschmidt and colleagues documented that having both ITNs and IRS provides a multiplicative protective effect against malaria infection beyond the added benefits accrued by having each intervention alone . The present study’s findings support these results and strengthen the evidence base for viewing the protection offered by having both ITNs and IRS as synergistic, especially in settings under which medium malaria transmission are experienced.
In sub-Saharan Africa today, countries face challenging decisions about the financing and prioritization of interventions to prevent malaria, upholding their fragile successes in malaria control, and the possibility of eliminating malaria from their borders. It is not enough to know whether malaria interventions are reaching the populations who need them. Understanding if, and to what extent, these interventions are related to health outcomes under every-day, routine conditions is essential. The present study’s findings offer a critical step toward this understanding, but they must be balanced with timely information on countries’ malaria programme needs and changing epidemiological profiles. The need to further assess the protectiveness of ITNs and IRS in low malaria transmission settings only intensifies, as is understanding how different intervention combinations work over time as countries shift their programmatic strategies along the control to elimination spectrum [39, 45]. At a time when donor financial assistance is potentially flat-lining  and increasing challenges face the malaria community (e g, the documentation of insecticide resistance in some areas in the world), trying to expand programmes, let alone potentially “doubling-up” the receipt of intervention, involves substantial political will, evidence-based cost-effectiveness analyses, and strategic use of resources.
The present study’s findings need to be interpreted in light of the limitations of the analysis. First, the small sample sizes especially in mortality, limit the power to detect statistical significance. Second, attempts were made to control for as many confounders as were analytically appropriate and plausible; nonetheless, residual confounding may still be present since this is a non-randomized study. Third, the effect of intervention integrity was not investigated as it was associated with health outcomes. For ITNs, data were not available for whether the nets had been washed, which potentially compromises the net’s insecticide potency, or had sizeable holes . For IRS, survey data did not indicate the type of insecticide used for spraying, which is a potentially important factor because the duration of IRS efficacy varies by insecticide applied and countries do not generally use only one type for IRS . Finally, the present study focused on individual exposure to ITNs and IRS and did not further consider community-level effects accrued by either intervention. However, studies show that communal effects for ITNs and IRS may be conferred within smaller ranges than what surveys can capture (e g, neighbouring households within 100 to 300 m, as opposed to the expanse of a whole village, which is represented by the survey PSU) [47–50].
Using publicly available survey data, the present study has sought to quantify the association between having a combination of insecticide-based interventions and child health outcomes. Overall the findings suggest that the combined use of IRS and ITNs provides greater protection against malaria than the use of IRS or ITNs alone. In addition, in medium malaria transmission areas, these results suggest that there may be a synergistic effect of using ITNs and IRS together. While continued work is necessary in order to fully understand how ITNs and IRS are related to child mortality, these findings provide a scientific basis for viewing the combination of ITNs and IRS, under medium malaria transmission settings, as more protective against malaria morbidity than having singular intervention and begin filling the knowledge gap considering the differential effectiveness of ITNs and IRS in sub-Saharan Africa.