Promoting child health—Structural and policy level interventions.

Of the twenty-five reviews that cover interventions at the structural and policy levels, seven looked at interventions promoting child health. These covered two main types of interventions: contracting out health promotion services to non‐governmental service providers, and financial and non-financial assistance and incentives. All of these interventions had a targeted approach, focusing on the impacts on the most disadvantaged in society.

Contracting out health promotion services (1 review). One high-quality review covering two primary studies explored the effects of contracting out health services to non‐governmental service providers to improve service utilization and health outcomes in rural, under‐resourced communities [49]. One study covered by this review includes contracting out child health promotion services in Cambodia, which addressed two factors of disadvantages: residence (rural) and income (under‐resourced communities). The review found low-quality evidence that the intervention had no effect on diarrhea incidence in children from these vulnerable groups. Therefore, it can be considered as having a neutral effect on health disadvantages.

Financial and non-financial assistance and incentives (6 reviews). Six reviews reported on the effects of financial and non-financial assistance and incentives to poor or vulnerable families and communities on child immunization coverage and uptake [62, 64, 72, 78, 83, 85]. One review by Owusu-Addo & Cross [83] also reported on the effect of conditional cash transfer on morbidity risks in children from selected infectious diseases. Two of these reviews were assessed as high methodological quality [62, 64], three as low [78, 83, 85] and one as moderate quality [72].

Among the factors of vulnerabilities addressed by these reviews, five covered interventions targeting poverty or low socio-economic status, with a particular focus on rural communities. Higher health needs were also found as factors of disadvantages in studies across all reviews.

Despite covering five of the same programs and four of the same primary studies, the findings of these reviews are mixed or even contradictory. The same trial from India—covered by three reviews [62, 72, 78]—shows moderate-quality evidence that food incentives can improve child immunization coverage. This would suggest that food incentives may be effective in addressing the health disadvantage these communities face in terms of immunization coverage. Referring to this trial, Jarrett et al. [72] pointed out that addressing the basic needs of the community as this intervention did, could be particularly relevant to building trust and expanding services in underserved and disadvantaged communities. Another study of non-monetary incentives at immunization visit covered by Johri et al. [62] also found this type of incentive effective in addressing the health disadvantage of the rural communities covered by improving immunization coverage.

Five reviews covered monetary incentives such as cash transfers or vouchers, with or without additional services [62, 64, 78, 83, 85]. Owusu-Addo & Cross [83] concluded that conditional cash transfers were effective both in improving child immunization coverage and in reducing morbidity risks from diarrhea and acute respiratory infections (ARI) in the targeted communities. Johri et al. [62] and Munk et al. [85] also concluded that selected types of incentives were effective in improving immunization coverage, although they highlighted a high risk of bias in the included studies and that there was heterogeneity in the results. In contrast, the reviews by Bright et al. [78] and Oyo-Ita et al. [64] found either mixed or no effect of monetary incentives on child immunization. The cause for these contradictions, including in the appraisal of the same evidence, is unclear. However, it should be noted that the reviews by Bright et al., Johri et al. and Oyo-Ita et al. [62, 64, 78] included a wider variety of programs, including non-conditional cash transfer and vouchers, which may partially explain the discrepancies and heterogeneity. The review by Oyo-Ita et al. [64] also noted that in three of the programs, children above two years old had no improvement in immunization coverage, thus suggesting a potential neutral or detrimental effect on inequalities between age groups. Looking specifically at the two high-quality reviews [62, 64], both highlighted the heterogeneity and uneven quality of the included studies and their findings. This may partially explain the contradictory conclusions these two reviews drew on the effect of incentive-based programs. Therefore, we cannot conclude at this point on the effectiveness of such interventions in addressing health disadvantages.

Promoting child health—Community, household and individual level interventions.

Of the forty-four reviews that cover interventions at the community, households and individual levels, twelve looked at interventions promoting child health. These covered two main types of interventions: nutrition supplements and health promotion education given to caregivers. Of the nine reviews covering nutrition supplements, six included interventions adopting a universal approach to health equity and four included interventions adopting a mixture of universal and targeted approaches. The reviews covering health promotion education to caregivers also adopted a mixture of universal and targeted approaches.

Nutrition supplements (9 reviews). Nine reviews addressed the effect of nutrition supplements on child morbidity and mortality from selected infectious diseases [48, 60, 61, 63, 66, 67, 70, 103, 107]. Seven of these reviews were assessed as high methodological quality [48, 60, 61, 63, 66, 67, 70] and two as critically low quality [103, 107].

Three of the reviews covered eighteen individual studies on zinc supplementation given in different formats and dosage to children, with or without other nutrition interventions [48, 63, 107]. Two reviews assessed the effect of this intervention on respiratory infection morbidity (including pneumonia) [63, 107], one on diarrhea incidence [107] and one on otitis morbidity [48] in the general population. Lassi et al. [63] and Tam et al. [107] reached opposite conclusions on the effect of zinc supplementation on respiratory infections in children. While Lassi et al. [63] found low-quality evidence that zinc supplementation reduced child morbidity from pneumonia, Tam et al. [107] found no significant effect of the intervention on lower respiratory tract infections (LRTI—including pneumonia, evidence quality unknown). Many factors may explain this contradiction, including the definition of the outcome observed, the quality of the reviews (high vs. critically low, respectively) or the year of the search and number of studies: Tam et al. is more recent and includes 10 new trials compared to Lassi et al. However, the lack of details on the studies included in Tam et al.’s meta-analysis makes further comparison difficult. Since Lassi et al. also highlighted the low-quality of the evidence available, we conclude that further research is needed to assess the effect of zinc supplementation on respiratory infections in children living in different contexts. Tam et al. [107] also assessed the effect of this intervention on diarrhea incidence and found a positive effect of zinc supplement on this outcome, although the meta-analysis showed high heterogeneity. Finally, Gulani and Sachdev [48] found no effect of the supplements on otitis morbidity in the general population. If we are to focus exclusively on the findings of high-quality reviews, there is low-quality evidence that zinc supplementation reduces pneumonia morbidity but has no effect on otitis [48, 63]. Assessing their impacts on health inequalities, these two reviews reported results for some vulnerable or disadvantaged populations. Lassi et al. [63] found low-quality evidence that zinc supplementation reduced child morbidity from pneumonia in specific groups such as populations living in slums, low-economic status neighborhoods or populations affected by HIV. Gulani and Sachdev [48] found studies which suggested a positive effect of the supplement for infants or malnourished children. However, Gulani and Sachdev [48] concluded that this evidence is currently insufficient to draw definitive conclusions on the benefits of zinc supplementation for specific vulnerable groups.

Three reviews [60, 61, 107] cover eleven individual randomized controlled trials (RCTs) of mixed quality giving vitamin A to neonates, infants or children alongside other vitamin supplements or selected vaccines. None of the reviews, including the two-high-quality reviews, found any effect from the interventions on either morbidity nor mortality from diarrheal infections, meningitis or respiratory infections (ARI or LRTI). The reviews did not report disaggregated effects of the interventions for different population groups, although the trials captured by Haider et al. [60]’s review all took place in populations with high prevalence of either vitamin A deficiencies or HIV. Whether this review’s finding could inform targeted nutrition interventions in these groups would require further research.

As for other micronutrient supplementation interventions in the general child population, Tam et al. [107] found no effect of multiple micronutrient (MMN) supplementation capsules or tablets, or of iron or iron/Folic acid supplementation on diarrhea morbidity. Finally, Eaton et al. [66] concluded that there was insufficient evidence to assess the effect of animal‐source foods on infectious disease morbidity outcomes in infants and children across different settings. Neither reviews reported disaggregated effects of the intervention for different population groups.

Four reviews included nutrition supplementation interventions that adopted a mixture of universal and targeted approaches. Three reviews, covering two cross-sectional studies and twelve RCTs, examined the effects of point-of-use multiple micronutrient powder (POU-MNP) sachets or fortified flour provided to caregivers to be mixed with infants and children’s food before consumption [67, 103, 107]. The findings of these reviews were mixed. While Carroll et al. [103] found that POU-MNP reduced diarrhea incidence amongst refugee children, Tam et al. [107] found evidence that the intervention actually increased diarrhea prevalence in healthy children. Meanwhile, Suchdev et al. [67] found no effect of POU-MNP in infants living in a variety of settings. The difference in settings and populations may explain these mixed findings. The reviews by Carroll et al. [103] and Suchdev et al. [67] included studies in vulnerable populations—such as children living in refugee camps, slums or malaria-endemic areas, while Tam et al. [107] focused on the general population. Looking at the only high-quality review, Suchdev et al. [67] concluded that the evidence was currently insufficient to conclude on the effect of POU-MNP on diarrhea morbidity or on other infectious diseases outcomes the review explored, such as malaria or upper respiratory tract infections. The fact that all three reviews acknowledge some contradictions or heterogeneity in their results further supports Suchdev et al.’s conclusions on insufficient evidence.

The last review looking at nutritional interventions is Imdad et al. [70]’s meta-analysis on neonatal oral probiotics/synbiotics supplementation added to breast milk and/or formula at various dosages and frequency. Although the review did not limit its search to specific groups of newborns, all twenty-one RCTs reporting on neonatal sepsis took place amongst low-birth-weight or preterm babies, primarily in hospital settings in middle-income countries. The authors found high-quality evidence that probiotics supplementation reduced the risk of sepsis amongst these vulnerable groups of neonates. Whether these findings could be extended to the general population or community settings requires further research. The review did not disaggregate the results further according to other factors of vulnerabilities or inequalities.

Health promotion education to caregivers (3 reviews). Three reviews covering 13 individual studies addressed the effect of health promotion education to caregivers of children under five years old by health workers at community or individual levels [62, 78, 92]. Two reviews [62, 78], assessed as low and high methodological quality respectively, show mixed results on immunization uptake and coverage. Both reviews cover the same types of delivery mechanism (such community health workers, community groups, information session). Johri et al. [62], where both the review and the studies it covers were assessed as high quality, found these interventions had a positive effect and there was low heterogeneity among the studies covered. Conversely, Bright et al. [78], where both the review and its studies were assessed as lower quality, showed mixed results. It should be noted that several of these studies combined health promotion education with other interventions, which may explain the discrepancies. Finally, two studies covered by Flórez et al. [92] showed a positive effect of the intervention on morbidity from diarrhea and one study showed a positive effect on malaria morbidity. As all three reviews brought together both targeted and universal interventions, it is difficult to draw definitive conclusions on the effect of these interventions on health inequalities.

Preventing infectious diseases in children—Structural and policy levels.

Of the twenty-five reviews that covered interventions at the structural and policy levels, seventeen looked at infectious disease prevention interventions. These covered three main types of interventions: improving the water supply infrastructure, introducing new vaccines into national immunization programs and contracting out preventive health services.

The reviews covering water supply infrastructure and new vaccine introduction adopted a universal approach to health equity while the review on contracting out preventive health services adopted a targeted approach, focusing on a disadvantaged group in the countries of intervention.

Water supply infrastructure (2 reviews). Two reviews [55, 90], assessed as high and critically low methodological quality respectively, assessed the effect of improving water supply infrastructure on childhood diarrhea morbidity. Covering a total of six different studies, both reviews concluded that there was insufficient evidence to decide on the potential effectiveness of such interventions on childhood diarrhea risk. Neither of them reported results disaggregated by sub-groups or the effect of the interventions on inequalities.

Introduction of new vaccines (14 reviews). Fourteen reviews covered the effect of new vaccines, that have been introduced into national immunization programs [58, 68, 81, 86–88, 91, 93, 98, 101, 102, 104, 108, 110]. Among those, two covered Haemophilus influenzae type b (Hib) vaccines [86, 91], six covered pneumococcal conjugate vaccines (PCV) [58, 81, 87, 91, 104, 110] and eight rotavirus vaccines [68, 88, 91, 93, 98, 101, 102, 108]. With the exception of the review by de Oliveira et al. [58] and Soares‐Weiser et al. [68], all these reviews were assessed as low [81] or critically low methodological quality [86–88, 91, 93, 98, 101, 102, 104, 110].

The Hib vaccine was found effective in reducing either Hib meningitis mortality [91] or morbidity [86] in infants (the group targeted by vaccination) in fourteen studies from Bangladesh, Pakistan and Brazil. DeAntonio et al. [91] also found that another study from Uruguay suggested a potential herd effect of the introduction of the Hib vaccine among children under three years not targeted by the vaccine. However, the evidence was currently insufficient to confirm such an effect. The quality of the primary studies covered by the reviews was unclear and the reviews themselves were assessed as critically low quality, thus calling for caution in the interpretation of these results. The potential health inequalities effect of the Hib vaccine introduction is unknown, as neither review covering this vaccine reported any health inequalities measure.

The six reviews covering various types of PCV vaccines from thirty-one primary studies also found that further to their licensing and introduction into national programs, these vaccines were effective in reducing mortality [58] or morbidity [81, 87, 91, 104, 110] from pneumococcal infections in children. Bonner et al. [87], de Oliveira et al. [58] and Ngocho et al. [81] assessed the studies as fair to high quality. However, it should be noted that half of these reviews focus exclusively on Latin America [58, 87, 91] while the other half focused on the African continent only [81, 104, 110]. Hence the generalizability of these results across all LMICs and continents cannot be confirmed. Looking at the health equity implications of PCV vaccines, five of the reviews included some measure of health inequalities. Three reviews [58, 81, 104] found a stronger decline of pneumococcal diseases in younger children, the group carrying a higher burden of the disease, suggesting that the vaccine may help reduce inequalities between age groups in the burden of these diseases. The review by Bonner et al. [87] also focuses on children above the vaccination age and concludes that the implementation of catch-up doses of PCV vaccines for children above the targeted age at the time of vaccine introduction was effective in reducing the burden of pneumococcal disease. However, the review by Vardanjani et al. [110] highlights that PCV vaccines may have a detrimental impact on health inequalities in another domain. Reviewing case-control studies from South Africa bringing together HIV infected and uninfected children, Vardanjani et al. not only found that vaccine effectiveness was lower amongst HIV positive children but that this effect may also reverse the positive effect on the gap between age groups. Whether these conclusions can be extended to settings and HIV-affected populations outside of South Africa would require further research. Looking at the only high-quality review covering PCV vaccine, findings by de Oliveira et al. [58] confirm a positive effect of the PCV vaccine on pneumococcal meningitis mortality and on the gap between younger and older children.

Finally, all eight reviews covering different types of rotavirus vaccines found the vaccine effective in reducing child mortality [88, 91, 93, 98] or morbidity [68, 88, 93, 98, 101, 102, 108] due to diarrhea or gastroenteritis, with or without confirmation of rotavirus infection. These eight reviews covered 104 individual primary studies. DeAntonio et al. [91] also found evidence suggesting a potential herd effect of the vaccine among children under five years above the vaccination age but concluded that there was currently insufficient evidence to confirm this finding. However, all but one of these reviews were assessed as critically low quality and the quality of the primary studies they covered was also mixed. Four out of six of these reviews were also exclusively focused in Latin America [88, 91, 98, 101], hence the number of overlaps in the vaccination programs and studies covered. This region is therefore over-represented in the results. Yet, Soares‐Weiser et al. [68], the only high-quality review for this intervention, found moderate- to high-quality evidence that all three types of rotavirus vaccines approved by WHO were effective in reducing the burden of rotavirus or all-cause diarrheal diseases across fifteen low- and middle-income countries. However, the meta-analyses showed statistical moderate to high heterogeneity and variations in the impact of the vaccine according to the type of diarrheal disease considered. Looking at the health equity implications of these interventions, four reviews noted differences in the effectiveness of the vaccine across age groups and/or contexts and one review reported disaggregated results for selected vulnerable groups. Three reviews noted a higher decline of gastroenteritis morbidity/mortality among infants, the age group carrying a higher disease burden, suggesting a potential reduction of disease burden inequalities between age groups [68, 88, 102]. Three reviews also highlighted both uneven efficacy and effectiveness of the vaccination between countries and regions [88, 93, 102], with Chavers et al.’s [88] results suggesting that the vaccine was most effective in countries with comparatively lower child mortality. These findings may point towards the potential increase in global inequalities. Yet, the mixed quality of evidence available and uneven geographical coverage of the reviews suggest that there is currently insufficient evidence to assess the effect of rotavirus vaccines on global health inequalities. As for the effect of rotavirus vaccination on disadvantaged groups, the review Soares‐Weiser et al. [68] reveals opposite results depending on the factor of disadvantage considered. While rotavirus vaccination significantly reduced the burden of diarrheal diseases amongst malnourished children, it had no significant effect amongst HIV-exposed or -infected children. However, the number of studies supporting these findings was very small, with one RCT reporting on malnourished children and three on HIV-exposed or -infected children.

Contracting out preventive services (1 review). The review by Odendaal et al. [49] on contracting out health services to non‐governmental service providers also covered the contracting out of preventive services. The authors found moderate-quality evidence that such an intervention had no effect on immunization uptake in children in neither country of intervention—Cambodia and Guatemala. This suggests that this intervention had a neutral effect on health disadvantages when it comes to immunization uptake.

Preventing infectious diseases in children—Community, households and individual levels.

Of the forty-four reviews that cover interventions at the community, households and individual levels, thirty-one included infectious disease prevention interventions. The reviews looked at interventions aiming to increase the use or coverage of preventive tools (immunization, bed nets, infant HIV prophylaxis) as well as water, sanitation and hygiene (WASH) interventions, and hospital-based interventions to prevent infections in vulnerable neonates.

The reviews covering WASH interventions adopted a universal approach to health equity while the reviews on hospital-based interventions towards vulnerable neonates adopted a targeted approach. Finally, the reviews covering interventions aiming to increase the use or coverage of preventive tools synthesized together interventions with targeted and universal approaches to health equity.

Increasing child immunization (17 reviews). Seventeen reviews covered interventions at community or individual levels aiming to improve child immunization uptake and/or coverage either by increasing the demand for or the supply of immunization services [50, 59, 64, 65, 72–74, 82, 85, 89, 92, 95–97, 99, 106, 109]. These reviews synthesized 100 individual primary studies, almost all of which were from Asia and Sub-Saharan Africa, with only three studies from Latin America. Three of these reviews were assessed as high quality [59, 64, 65] and three as moderate [72–74]. The other eleven were assessed either as low [50, 82, 85] or critically low methodological quality [89, 92, 95–97, 99, 106, 109].

All fourteen reviews covering intervention aiming to increase the demand for child immunization -such as immunization education, communication, social mobilization and information campaigns, and reminder tools with or without outreach services—found some positive effect on the coverage or uptake of one or more vaccines [50, 64, 65, 72–74, 82, 85, 89, 92, 95–97, 106]. However, eight of them also reported at least one study or outcome where the intervention was ineffective [65, 74, 85, 89, 95–97, 106] and seven meta-analyses reported moderate to high heterogeneity in their results [64, 65, 73, 74, 95, 96, 106]. Neither the type of vaccine outcome measured nor the sub-type of intervention seem to explain alone these variations. The quality of the primary studies was also very heterogeneous. A sub-analysis of high- and moderate-quality reviews leads to the same findings, namely, that while this type of intervention is found effective in improving childhood vaccine uptakes and coverage in most studies, all five reviews raised the issues of the quality of the evidence base and the heterogeneity of pooled results [64, 65, 72–74]. In terms of health equity, three of these reviews discussed the effect or suitability of these interventions among different groups or settings [50, 72, 97] and six reviews focused specifically on a disadvantaged group [64, 65, 72, 82, 89, 97]. The factors of disadvantage targeted by these interventions included higher health needs (especially low vaccination rates) [64, 65, 72, 82], poverty [64, 65, 72, 89], place of residence and type of dwelling (e.g., rural communities, urban slums, informal settlements or small dwellings) [50, 64, 65, 89, 97], ethnicity [97], and low literacy or education [50, 64]. Overall, the findings suggest that such interventions may be effective both at population level and in addressing the targeted populations’ health disadvantage. However, the results’ heterogeneity warrants further research to confirm such claim. The type of interventions covered also seems to affect their impact on health equity. For example, three reviews highlighted targeted communication or education towards pre-identified disadvantaged groups as an effective approach to raising immunization uptake in those groups [50, 72, 97]. Conversely, the evidence on the potential effect of mass communication promoting immunization on health inequalities is less clear. While Yuan et al. [50] concluded that the evidence was unclear and limited (only one study), Jarrett et al. [72] highlighted the fact that such ‘top-down’ approaches to communication may not be effective across all groups, depending on the level of vaccine hesitancy–thus potentially raising inequalities issues.

The eight reviews reporting on interventions targeting the supply of child immunization services, found mixed results, with variations in effectiveness according to the specifics of the intervention, the type of vaccine considered or even amongst a similar type of intervention [59, 64, 72, 82, 85, 97, 99, 109]. The quality of the primary studies included was also uneven [59, 64, 72, 82, 85, 97, 99]. They covered interventions such as the integration of immunization with other health services, improving community or outreach services, the introduction of new technology or strategies, the provision of tools and material to support immunization health professionals or health professional training, either on their own or as part of multi-component interventions. The findings of the three high- and moderate-quality reviews were consistent with those highlighted above. In terms of health equity, two of the eight reviews reported on the comparative effect of service integration interventions in rural versus urban areas [97, 99]. Smith et al. [99] reported on one poor-quality study showing that while the integration of HIV and immunization services had a positive effect on overall population immunization uptake in urban sites, its effect was detrimental in rural sites due to stigma and concerns over discrimination. Nelson et al. [97] also highlighted the potential risk of increased rural/urban disparities with integrated services, reporting on one low-risk-of-bias study showing that such integration could not address distance issues, thus making it more suitable and effective to urban contexts rather than rural ones. Both studies took place in Sub-Saharan African countries. While the evidence from these reviews is mixed, they do point towards a potential risk for this type of intervention to increase inequalities in service uptake between rural and urban areas. As for the impact of supply-side interventions on addressing the health disadvantage, the five reviews reporting on interventions adopting a targeted approach to health equity suggested that the effect of these interventions on addressing various health disadvantages would require further high-quality studies to be confirmed [64, 82, 97, 99, 109]. Four reviews focused on areas or communities with higher health needs (e.g., low vaccination rate or high infectious disease prevalence). The other factors of disadvantages addressed by these interventions were the place of residence (i.e., urban versus rural communities) [64, 97, 99, 109], poverty [64] or families identified as vulnerable in their individual context [99, 109].

Increasing bed net use (3 reviews). Three reviews covered interventions aiming to increase bed net use for children under five [54, 76, 96]. The review by Naugle & Hornik, assessed as critically low methodologically, reported on a single study from Cameroun covering the effect of a mass media campaign to promote consistent bed net use [96]. It concluded that the evaluation of this campaign provided strong evidence that this intervention was effective in increasing bed net use for children under five. This review only reported the overall population impact of the intervention and did not report any disaggregated effect of the invention on different groups.

Two reviews, assessed as high and moderate quality methodologically, synthesized six studies on community outreach education activities on the use of insecticide treated nets (ITNs) and community health services to mothers and households with children under five [54]. All but one study took place in Sub-Saharan African countries, in malaria-endemic areas. The reviews found low-quality evidence that such interventions may be effective in increasing under-5 use of ITNs in these communities. This suggests that this intervention may reduce the health disadvantage faced by communities in malaria-endemic areas regarding the burden of that disease.

Preventing HIV transmission (3 reviews). Three reviews assess the effects of preventive HIV services for a targeted group of infants—those exposed to HIV, with the HIV status of the mother as the main factor of disadvantage [75, 84, 99]. One review focuses on HIV prophylactic treatment given to breastfed, HIV-exposed infants [75] while the other two focus on how HIV prophylactic treatment is delivered to HIV-exposed infants. All but one of the twelve included studies occurred in Sub-Saharan African countries.

The moderate-quality review by White et al. [75] found that compared to standard care, extending HIV infant prophylaxis throughout the breastfeeding period reduces the rate of HIV infections amongst HIV-exposed infants, even when the mother is not receiving antiretrovirals. While these findings are based on a small number of studies in urban sites, the authors assessed the evidence as moderate to high quality, thus recommending that HIV infant prophylaxis must continue until weaning.

The two reviews by Puchalski Ritchie et al. [84] and Smith et al. [99], assessed as low and critically low methodologically respectively, show mixed results on the uptake and coverage of preventive HIV services for infants exposed to HIV [84, 99]. Both reviews suggested that home visits by community health workers may be effective in addressing the health disadvantage faced by infants born to HIV positive mothers, leading to an increase in uptake of HIV prophylaxis. However, the quality of the two primary studies reporting on this type of intervention was mixed.

As for the integration of prevention of mother-to-child transmission (PMCT) with mother and child healthcare services, the reviews drew conflicting conclusions. Puchalski Ritchie et al. [84] found this intervention had either no or even detrimental effect on HIV prophylaxis uptake at birth from two trials assessed as having a high risk of bias. Smith et al. [99] found fair to good-quality evidence that integrating PMCT into post-partum care or immunization services increased the uptake of HIV prophylaxis. It is unclear whether the setting or the study design may explain these differences. This leads us to conclude that the effectiveness of service integration in addressing some of the health risks of HIV exposed children is currently unclear.

Puchalski Ritchie et al. [84] also covered three other interventions targeting HIV positive mothers and their infants, including peer-to-peer education, task-shifting and service quality improvement, none of which showed any effect on HIV prophylaxis uptake at birth. This suggests that these interventions seem to have no effect on addressing the health disadvantage of HIV exposed infants. The authors also noted the high risk of biases of the three trials testing these interventions.

WASH interventions (7 reviews). Seven reviews covered WASH interventions at the community or household level from 90 individual studies, reporting on their impacts on the overall population [55, 69, 77, 79, 80, 90, 94]. Two reviews were assessed as high quality [55, 69], one as moderate review [77] and four reviews as low [79, 80] or critically low methodologically [90, 94].

Three reviews covered 39 individual studies on water quality improvement at point-of-use [55, 77, 90]. Despite the difference in the quality of each review (rating from critically low to high), all three reviews found very similar results for individual type of interventions (i.e., chlorination, flocculation, solar disinfection). They concluded that there was low- to moderate-quality evidence that these interventions were effective in reducing diarrhea morbidity in children under five years old at population level. However, the reviews also highlighted the high heterogeneity of the results. None of the reviews reported disaggregated results by population groups for these types of interventions, thus providing no information on the potential impact of the intervention on health inequalities. However, Soboksa et al. [77]’s analysis by world regions shows significant differences in the effect of solar disinfection between Latin American and Asian countries, thus raising questions as to the potential implications of such interventions for global health inequalities.

Five reviews covered sanitation and hygiene interventions, such as hygiene promotion and education, community mobilization and campaigns, with or without the development of a WASH infrastructure [69, 79, 80, 90, 94]. These reviews showed mixed results of these interventions on morbidity from selected infectious diseases in children under five. Of the four reviews reporting on the effect of sanitation and/or hygiene interventions on diarrhea morbidity, only one review found them effective and reported high-quality evidence supporting their conclusions [79]. The other three showed mixed results depending on the context and the details of the intervention, also highlighting the low quality and lack of conclusive evidence [69, 90, 94]. The findings of the only high-quality reviews covering these types of interventions were consistent with those highlighted here [69].

The two reviews reporting on respiratory infections and pneumonia also had mixed findings. One review by Morita et al. [94] found no conclusive evidence on the effect of hygiene interventions on this outcome while the second by McGuinness et al. [80] showed mixed results. Both reviews highlighted the low quality of evidence currently available and were assessed as low and critically low quality themselves. However, McGuinness et al. [80] highlighted the result of one high-quality cluster RCT from Pakistan that had found hygiene education and the provision of hygiene product effective in reducing pneumonia illness rate. Finally, three reviews explored the effect of these interventions on other infectious diseases [69, 79, 94]. With the exception of trachoma cases–for which Freeman et al. [79] found moderate to high quality evidence that sanitation and hygiene interventions were effective in reducing trachoma morbidity, the reviews found either no effect [69, 79] or insufficient evidence [94] that sanitation and hygiene interventions had any impact on parasitic diseases; or on dysentery prevalence [69].

While some of this heterogeneity in results may be explained by variations in intervention settings, targeted population and the quality of evidence available, these variations also offer some findings relevant to health inequalities research. McGuinness et al. [80] highlighted that the effectiveness of hygiene interventions varied according to intervention settings and characteristics (urban child care-based vs. rural home-based interventions) as well as compliance but was unable to attribute these variations to a specific factor with available evidence. Majorin et al. [69] reported on one evaluation where the education and hygiene promotion intervention showed a reduction of diarrhea prevalence in rural sites but not in urban slums. Morita et al. [94] suggested that the effect of these interventions vary according to the child’s age. As only Majorin et al. [69]’s review was assessed as high quality, a sub-analysis according to the quality of the reviews cannot provide further insights into the implications of these interventions for health inequalities. Hence, while some of these reviews suggest that hygiene and sanitation interventions may be relevant when studying health inequalities in the burden of child infectious diseases, the evidence is currently insufficient to conclude on their effect on health inequalities or the relevant factor of inequalities they affect.

Preventing infections in vulnerable neonates (2 reviews). Two reviews, for which were assessed as high quality, synthesized the effect of targeted interventions in hospital settings to reduce the risk of infections in low-birth weight or premature neonates [56, 57]. The higher health needs of low-birth weight or premature babies was the main factor of disadvantage covered in these reviews. As all the interventions took place in hospital settings, including neonatal specialized units, the generalizability of the results beyond this specific setting is not possible.

Conde-Agudelo & Diaz-Rossello [57] included seven RCTs from India, Colombia, Ecuador and Madagascar relevant to this umbrella review. The review assessed the effect of kangaroo mother care on the risk of illnesses (as defined in individual studies) or selected infectious diseases at six months follow-up as compared to routine care. Based on these trials, the authors found moderate-quality evidence that this intervention had a positive effect the risks of severe illness, nosocomial infection/sepsis and lower respiratory tract diseases but no effect on mild/moderate infection and illness or diarrhea. Thus, this intervention may be effective in addressing the higher health risks of selected infectious diseases faced by this vulnerable group of infants.

The review by Cleminson & McGuire [56] assessed the effect of topical application of moisturizing emollients (ointments, creams, or oils) to increase the protective barrier function of the skin in premature infants and so, prevent infections in this vulnerable group. Of the synthesis of six trials relevant to this review, the authors found no effect of such an intervention on the incidence of invasive infections, regardless of the type of emollient used. Therefore, this intervention does not seem effective in addressing the higher health needs of this vulnerable population.

Preventing infectious diseases in children—Multilevel interventions.

We found four reviews [62, 69, 85, 97] covering preventive, multilevel interventions. All of them reported the overall population impacts of these interventions–with no equity aspects.

Combined demand-and supply-side interventions for child immunization (3 reviews). Three reviews [62, 85, 97] covering ten individual studies assessed the effect of combined demand-and supply-side interventions across levels to improve child immunization coverage and/or uptake. The quality of the reviews themselves were uneven as Johri et al. [62] was assessed as high methodological quality while Munk et al. [85] and Nelson et al. [97] were appraised as low and critically low respectively. Despite these differences, all three reviews found these interventions effective in improving child immunization but highlighted the heterogeneity of the result and issues with the quality of some of the studies.

While these interventions themselves did not all follow a targeted population approach by design, the reviews were focusing on urban populations [97], rural areas [85] or communities with lagging health or social indicators [62]. Hence the findings may only be relevant to these groups specifically. Then, these findings may inform targeted approaches addressing the health disadvantage of urban populations or communities with lagging health or social indicators regarding immunization. However, further research would be needed to confirm it. None of the reviews disaggregated their results by subgroup among the study population.

Combined hygiene education and water supply/sanitation infrastructure provision (1 review). The high-quality review by Majorin et al. [69] synthesized the results of two studies in Bangladesh where rural populations received an intervention combining water and sanitation infrastructure development with hygiene education and water quality improvement. The review found very low-quality evidence that such a combined intervention reduced dysentery and diarrhea prevalence.