Should we prioritise children 6–23 months of age for vitamin A supplementation? Case study of West and Central Africa

Background Vitamin A (VA) supplementation has been associated with reductions of all-cause child mortality. Child mortality amenable to VA, particularly related to infectious diseases, may be age dependent; hence, the beneficial effect of VA supplementation may differ between younger and older children. We aimed to estimate the all-cause child mortality disaggregated by younger and older than 2 years of age and estimate the contribution of VA supplementation in preventing child death in West and Central Africa. Methods Using the most recent (post-2010) cross-sectional Demographic and Health Surveys and Multiple Indicator Cluster Surveys, we analysed child-level data (n=187 651) from 20 West and Central African countries. Age-specific (all-cause) mortality rates were estimated using survival analyses. Age-specific VA supplementation coverage was linked with the age-specific all-cause child mortality to estimate the contribution of the supplementation in averting child death. Results The cost per averted child death was also estimated using an average cost of US$1.2/child and VA supplementation coverage which ranged from 14% in Cote d’Ivoire to 81% in the Gambia. About 75% of the under-5 mortality occurred in the first 2 years of life. The share of excess (all-cause) mortality averted by VA supplementation was significantly higher in the first 2 years of life. A mean reduction of 7.1 deaths/1000 live births was estimated for children 6–23 months, compared to a reduction of 2.5 deaths/1000 live births for older children (24–59 months). The mean cost/averted child death for the 20 countries was 2.8 times lower for the 6–23 than the 24–59 months age group. Conclusion Prioritising VA supplementation for children in the first 2 years of life could be more cost-effective than when implemented among 6–59 months of age.


INTRODUCTION
Vitamin A (VA) supplementation is and remains one of the key nutrition interventions implemented at scale. 1 Every year, billions of VA capsules are distributed through the health system to reduce VA deficiency, reducing child morbidity and averting child death. 2 However, with the increasing number of countries implementing interventions such as oil fortification and biofortification that can prevent VA deficiencies, continued VA supplementation is now primarily justified by its beneficial effect on reducing child morbidity and mortality 3 4 Indeed, the anti-infectious vitamin was named after its immune-regulating function and beneficial effect in reducing diarrhoea and measles mortality. 5he most recent meta-analysis by Imdad et al suggests that VA supplementation reduces the overall risk of death by 12% for children aged 6-59 months. 6The certainty of the evidence was rated as high for mortality reductions related to diarrhoea, but to a lesser extent for measles and meningitis.A disaggregated analysis also suggested that the effect of VA supplementation may vary by child age and geography (eg, Asia vs Africa). 6Furthermore,

WHAT IS ALREADY KNOWN ON THIS TOPIC
⇒ Vitamin A supplementation reduces the overall risk of death in children 6-59 months of age.⇒ A disaggregated analysis suggests that the effect of vitamin A supplementation may vary by child age and geography.
WHAT THIS STUDY ADDS ⇒ The majority of child deaths in West and Central Africa occur in children aged 6-23 months.⇒ Vitamin A supplementation is more effective and less costly in averting child deaths in children aged 6-23 months compared with older children.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
⇒ The findings suggest that additional benefits are likely with a focus of vitamin A supplementation in children 6-23 months than the current targets (6-59 months).⇒ Implementation research on when, where and how to make this age prioritisation is needed.

BMJ Nutrition, Prevention & Health
a recent study by Baye et al 4 indicated that the estimated averted child deaths related to VA supplementation are highly variable among countries in the sub-Saharan African region.This can be partly explained by varying levels of all-cause mortality and VA supplementationamenable child morbidity and mortality.In addition, the age distribution of child morbidity and mortality can significantly influence the effectiveness of the intervention, as VA supplementation is likely to avert more child deaths in age groups where infection-related morbidity and mortality occur the most. 7conomic downturns and the low and declining coverage of VA supplementation are raising difficult questions about how best to use the limited funds available to implement the supplementation programmes.Both from an economic and programme effectiveness standpoint, focusing the VA supplementation programme on a narrower age range of 6-23 months may be promising if the bulk of child death and health interventions are concentrated in this age group. 8However, the pros and cons of such a prioritisation should be evaluated against its impact on child mortality outcomes.
In West and Central Africa (WCA), over 1.8 million children under 5 years of age died in 2019. 9A non-negligible share of these deaths could have been prevented with optimal and effective coverage of VA supplementation, but coverage remained not only suboptimal but also showed a declining trend in several countries. 10Revitalising VA supplementation programmes is timely, but the potential impact of such prioritisation needs to be investigated.The current study aimed to disaggregate child mortality estimates for the West and Central African region as well as estimate the potential impact of focusing VA supplementation on averting death in various age groups in WCA in order to improve VA supplementation programming.

METHODOLOGY Data source
The study used the most recent Demographic and Health Survey (DHS) and Multiple Indicator Cluster Surveys (MICS) conducted in WCA.Countries that did not have a nationally representative household survey after 2010 as well as those with incomplete data on date of birth, age at death and VA supplementation coverage were excluded.A total of 20 countries (16 from the DHS and 4 from MICS) out of the 23 countries in the WCA region met the above criteria.This yielded a total sample size of 187 651 children with complete data on VA supplementation (online supplemental figure S1 and online supplemental table S1).Data on child mortality and population distribution were extracted from the United Nations Inter-Agency Group for Mortality Estimation (UN IGME) and the UN Population Division, respectively.

Patient and public involvement
The study did not have patient and public involvement.

Mortality estimates by age
First, the total number of child death was estimated from the UN IGME.Age-specific mortality rates were then estimated.Age-specific mortality rates are traditionally estimated by DHS and MICS using synthetic cohort life tables, in which mortality probabilities are first computed for smaller age categories based on a real cohort experience.However, this approach is confronted with incomplete cohorts, which affects its accuracy.Instead, we estimated the age-specific mortality using survival analyses as described by Victora et al. 11 The survival analyses compute the death over person-years (m-rates) first, which are then converted to probabilities (q-rates) by calculating the average time spent by those who died in the interval, as described in more details in Victora et al. 11 Age-specific child mortality was estimated for the following age groups: 0-5, 6-11, 12-23 and 23-59 months.
VA supplementation coverage and reductions in all-cause child mortality, by age Age-specific VA supplementation coverage was calculated using the most recent surveys.The coverages were then linked to the calculated age-specific all-cause child mortality at country level to estimate the supplementation's effectiveness in lowering all-cause child mortality.The study used estimates of all-cause mortality reductions of 41% (6-11 months) and 32% (12-59 months) related to the implementation of VA supplementation (vs no implementation), based on estimates from Imdad et al's study. 6or each country, the number and proportion of child deaths averted were calculated as follows: Death averted=coverage×effectiveness×affected proportion …. (1).
Where, Coverage: VA supplementation coverage from the most recent DHS/MICS survey (scenario 1) or a hypothetical near-universal coverage of ≥90% (scenario 2).
Effectiveness: all-cause child mortality reductions associated with VA supplementation, based on Imdad et al's study. 6ffected proportion: proportion of the population in the age range considered.

Cost of VA supplementation per child averted per year
Several studies that estimated the cost of VA supplementation programmes were examined.Cost estimates (per child/year) were available for Senegal (US$1.25),Cameroon (US$1.14) and India (US$1.20),leading to a mean cost of US$1.2/child/year, which corresponded to the UNICEF estimates for the regions of South Asia and Sub Saharan Africa. 2 12-14 With a coverage of 90%, the cost per averted child death was estimated for each country: Population: total population of the targeted age groups.
Cost: estimated cost for the delivery of VA, including supply and modalities to administer the vitamin to children.

RESULTS
Our analyses of the latest DHS/MICS data included 187 651 children with data on coverage of VA supplementation from 20 West and Central African countries.VA supplementation coverage ranged from 14% in Cote d'Ivoire to 81% in the Gambia (figure 1).8 of the 20 countries had coverage less than 50%, and only 1 country (Gambia) had coverage above 80%.With these coverage levels, the 20 countries saved an estimated 234 505 children's lives per year (table 1).A scale-up of VA supplementation to ≥90% would have saved an additional 107 981 child lives (342 486 lives).Figure 2 presents the distribution of under-5 child mortality by smaller age groups.The pooled analysis of data from the 20 countries suggests that about 75% of the under-5 mortality occurs in the first 2 years of life (0-23 months of age).Child mortality among infants 0-23 months of age ranged from 64% in Niger to 90% in Senegal.Consequently, the share of excess all-cause mortality averted by VA supplementation was significantly higher between 6 and 23 months compared with later, 24-59 months (figure 3).The country with the highest proportion of VA supplementation-averted child deaths between 6 and 23 months was Nigeria, where above 90% of child mortality avertable by VA supplementation occurred.The lowest VA supplementation-averted child death was reported for Gabon, where a little over 25% of child deaths that could be averted by VA supplementation occurred from 6 to 23 months.
Under 90% coverage, the mortality reduction rate (per 1000 live births) and the cost per child death avoided were calculated (table 2).In line with the absolute number of child deaths averted in the 20 countries, a mean reduction of 7.1 deaths per 1000 live births was estimated for children 6-23 months of age, compared with a reduction of 2.5 deaths per 1000 live births for children 24-59 months of age.The average cost per averted child death for the 20 countries was about three times (×2.8)lower for the 6-23 age group than the 24-59 age group.

DISCUSSION
Analysing nationally representative data from 20 West and Central African countries, this study showed that the coverage of VA supplementation was highly variable between countries but remained low relative to the WHO minimum target of 80%.Under-5 child mortality was concentrated in the first 2 years of life, resulting in the greatest VA supplementation-related mortality reductions for younger (6-23 months) children versus older (24-59 months).The cost of life savings per child was also much lower for younger children (6-23 months).
The reduction in under-5 child mortality witnessed over the past decades, the increasing availability of alternative VA interventions (eg, fortification), the perceived risk of excessive VA intake and the increasing competition over limited resources for other programmes have led to a call for a fresh look at VA supplementation programmes. 15he study by Baye et al 4 recently showed that the excess child mortality that can be averted by VA supplementation has declined over the years, but also illustrated that it remains significant for several countries in sub-Saharan Africa.According to the study, some countries could benefit from a scaleback or a more targeted approach.Indeed, another finding suggested that perhaps focusing resources on younger children could help increase effective coverage through more focus and better integration of VA supplementation with other preventive nutrition and health interventions. 7ur study has shown that about two-thirds of child deaths in the West and Central African countries included in the analyses occur for children 6-23 months, a finding in line with a recent global analysis. 8The recently updated meta-analyses also showed that all-cause mortality reductions related to VA supplementation are slightly higher for younger (6-11 months) than older (12-59 months) children (32%). 6The cost per child death avoided was also in favour of prioritising the programme for younger children (6-23 months) over older children (three times lower).Further reductions in delivery costs can be expected with shifts to routine programme delivery through the health system, though it has been difficult to estimate the savings. 13In addition, campaign fatigue is already leading countries and donors to transition to routine delivery.Although theoretically more sustainable, such shifts have already unmasked the weaknesses and inequities of the health system, where the more remote and vulnerable often have limited access. 15 16In addition, in places where VA supplementation was bundled with polio vaccination, coverage also showed a drop due to the phasing out of the polio vaccination campaigns as polio  This study has several limitations that need to be considered when interpreting our findings.First, this is a cross-sectional study; hence, causality cannot be established.Our estimates of excess mortality avoidable through VA supplementation, on the other hand, were based on evidence from meta-analyses of randomised controlled trials.Second, the study used VA supplementation coverage estimates from population-based surveys such as DHS/MICS, which only capture exposure over the previous 6 months.Given that the relative risk of child mortality related to VA supplementation is established for two doses, the finding assumed a similar coverage for the second semester.Third, the cost of delivery of VA supplementation was primarily estimated for programmes delivered through campaigns; hence, estimates for VA supplementation delivered through the routine health system are likely to be overestimated.

BMJ Nutrition, Prevention & Health
Notwithstanding the above limitations, this study showed that a higher proportion of child mortality in the region that could be prevented by VA is concentrated in the first 2 years of life, between 6 and 23 months.Prioritising VA supplementation for children in the first 2 years of life could be more cost-effective than when implemented among 6-59 months of age.While considering the lowering of the age range of the intervention, it is also important to strengthen links with the health system, and address existing inequalities in access to healthcare.With changing diets, disease epidemiology, overlapping Nutrition, Prevention & HealthAverted cases: number of child deaths averted considering a 90% coverage.

Figure 1
Figure 1 Vitamin A supplementation coverage among children 6-59 years of age in selected West and Central African countries.

Figure 3
Figure 3 Per cent share of averted deaths among children 6-59 months by age group.

Table 1
Comparison of actual mortality rates versus estimated mortality rates in absence of preventive vitamin A supplementation by age groups irradiation goals were getting within reach.Altogether, if shifts to routine delivery are the way forward, careful planning and alignment with other nutrition and health interventions will be critical.However, our results suggest the possibility to increase VA supplementation coverage and cost-effectiveness in reducing child mortality if the intervention is focused to children 6-23 months of age.

Table 1 Continued
Figure 2 Per cent distribution of child mortality (0-59 months) by age group.

Table 2
Estimated cost per averted death (in USD) for 90% coverage and child averted by vitamin A supplementation per 1000 children(6-59 months)