Our findings suggest that SMC is effective in preventing clinical malaria among children 3–59 months during the high transmission season in northern Mozambique. Results of the nRCT show an estimated protective effect of SMC of 86%. This reflects similar results shown in a recent study of similar design conducted in Uganda [18], which found a protective effect of 92%, and with the pooled result of a case-control study conducted in five countries showing a protective effect of 88% [19]. The observed results also align with those of other clinical trials showing the high coverage achieved by SMC, and reporting good adherence at day 2 and 3, as per results of the EoR coverage survey conducted in our study [16]. Large scale clinical trials administering SP + AQ as SMC [20–22] have shown significant protective effectiveness against clinical malaria during the transmission season. While some work has been done [23], the impact of seasonality on SMC effectiveness needs to be fully understood in east and southern Africa, and further work is required to document this, especially as new locations in ESA, including our study sites, do not necessarily completely fit the WHO seasonality criteria for SMC. Some recent work done using dynamical modelling found that the effect size of SMC is highest when baseline incidence is lowest, suggesting the need to account for seasonality in programming [24].
During the implementation of SMC in four cycles the most frequent adverse event reported was vomiting and fever. These adverse effects were also found as the most frequently reported in other studies [25–27]. All the adverse events were resolved without any other intervention or administration of other medications.
The absence of Pfdhfr 164L marker is reassuring as this is often associated with pyrimethamine failure [28]. Similarly, no Pfrcrt mutations were detected, suggesting absence of amodiaquine resistance mediated by Pfcrt. Pfmdr1 mutants are found in approximately half of the samples processed for 184F. Change over one round of any marker is not significant and based on the data presented here, there is no statistical evidence that the observed difference is due to SMC introduction. The combination of SNPs of relevant Pfdhps-dhfr mutants is notable. However, long term implications of this observation will be much better understood once the molecular and chemoprevention efficacy component results of phase 2 study become available.
The study also revealed a generally high acceptance of SMC among communities, with caregivers reporting a significant reduction in malaria among their children and an improvement in their quality of life. Community members learned about SMC campaign from different sources of information such as community health workers, their leaders, their neighbours, and mass media. This multi-source approach in spreading the information about SMC and its objectives in this context worked effectively. The involvement of community members has been identified as a key advantage in delivering SMC and malaria messages at community level by other studies [29–31]. Benefits of SMC have been already reported by caregivers of children less than five years of age, holding the view that the combination drug was very useful in preventing malaria [32, 33]. Despite the general acceptance and positive perception of SMC, caregivers, community distributors and stakeholders reported several challenges, such as mistrust, lack of partner approval, fear for side effects and local beliefs. These barriers have been also reported in other contexts when delivering a community intervention [33]. However, it has been suggested that continued health education can increase the acceptability of SMC [32], as well as to drive the delivery method and applying the earned trust [30] and the delay in payment incentives to staff [33]. On the other hand, caregivers and key informants provided clear views about enablers for the acceptability of the campaign, that as already noted here above were the involvement of community members and community leaders, both in spreading the information and distributing the medication, the door-to- door delivery, the free access to the medication and the community network, as observed also in other contexts [30, 34]. Participants’ suggestion to expand the intervention to older age groups illustrates the high acceptability of the intervention and reflects the perception on the burden of malaria as a health concern that affects not only young children but also older age groups. Some studies indicate that expanding SMC to older children can contribute towards reducing the incidence of malaria [20, 25], and the most recent WHO guidelines for malaria recognise that the target age for SMC should be selected based on risk of severe malaria [1]. Overall, these results provide valuable insights into the implementation of SMC at community level, emphasizing the importance of involving the community, utilizing the natural network to increase the transparency about the objectives and the ownership of all the beneficiaries.
Strengths
The study design includes a variety of methods, combining a non-randomised controlled trial, cross-sectional surveys, and qualitative interviews. This broad scope allows for a comprehensive, initial understanding of the intervention’s effects. Secondly, the inclusion of a comparison (Lalaua) and intervention districts (Malema and Mecubúri) allows for comparison, increasing the reliability of the results. Thirdly, the follow-up visits after each cycle and detailed recording of confirmed malaria cases provide robustness to the study’s ability to accurately assess the intervention. Finally, the inclusion of a diverse range of stakeholders at any level up, from the community to the central level, provides a more exhaustive perspective of the feasibility and acceptability of SMC.
Limitations
Despite the encouraging results provided by this study, however there are several limitations. First, the absence of randomization in the trial design may have introduced bias as the arms were not evenly balanced with respect to potential confounding factors. However, after analysis of baseline characteristics of children, caregivers, and households in the two follow-up groups, we found differences in only two variables: children’s use of bednets the night before the base line survey and incidence of fever in the previous month before the baseline survey among children. This baseline imbalance was addressed by use of random effects at the community level to account for community-level differences in malaria transmission. Also, as there was some loss to follow-up (~ 10%), which could have impacted on the overall power of the final analysis; however, the study was powered conservatively to detect a significant difference in hazard of malaria cases of only 40%. Secondly, bias may have been introduced in the qualitative data through the selection of respondents who had received SMC. Thirdly, caregivers’ report on fever or adherence on day 2 and 3 may have introduced recall bias in the way caregivers reported the following month to community distributors, especially if caregiver’s recall is influenced by their perceptions of SMC. Fourthly, this study was conducted in a specific area of Nampula province, affecting the generalisability of the findings to other regions of Mozambique or other countries. This will be addressed by further studies, such as the rapid assessments [35]. Finally, Sanger sequencing was used to analyse molecular markers of antimalarial resistance, which has lower sensitivity to detect mixed infections (resistant and mutant alleles) when compared to next generation sequencing.