Background
Non-exclusive breastfeeding (non-EBF) is a risk factor for a number of diseases, including infant mortality from diarrhoea, upper respiratory infections and other common infectious diseases [
1‐
3]. Nigeria – a developing country with 40 million children and the leading beneficiary of developmental assistance for health in Africa [
4] – has one of the highest rates of non-EBF among infants aged 0–5 months (84% on average between 1999 and 2013) [
5,
6]. This is despite the introduction and implementation of various national and subnational initiatives to reduce non-EBF practice [
5,
7]. The Global Burden of Diseases, Injuries, and Risk Factors study 2016 (GBD 2016) ranked non-EBF among the top ten risk factors for under-five deaths and disability-adjusted life years (DALYs) in Nigeria, accounting for 25,300 under-five deaths and 2.1 million DALYs lost in 2016, respectively [
8,
9]. This study also found that a large proportion of childhood wasting and malnutrition can be attributed to diarrhoea and other common infectious diseases, which also reflects the role of non-EBF in childhood morbidity.
In Nigeria, various facility- and population-based studies have identified key modifiable factors associated with non-EBF, including socio-economic factors (i.e., low maternal education and poor household wealth) and health service factors (i.e., fewer antenatal care visits, home delivery and delivery assisted by non-health professionals) [
5,
10‐
15]. Additional factors identified in Nigeria include a lack of family support, mother’s employment, and myths and belief systems held for breastfeeding [
16‐
18]. At the population level, the effect of a particular risk factor is dependent on the strength of association of the risk factor with a given outcome, and the prevalence of the risk factor in the population of interest [
19,
20]. Changes in the prevalence of non-EBF in Nigeria may be affected by a range of strategic initiatives, including policy development (such as socio-economic reforms to reduce poverty and improve female education); facility-based interventions (such as strengthening the baby friendly hospital initiative) and community-based initiatives (such as the baby friendly community initiative and involvement of family members in infant feeding counselling) [
5,
10].
The population attributable fraction (PAF) is the proportional reduction in the incidence of a disease or outcome over a specified time interval that would be achieved by eliminating the exposure(s) of interest from the population under an alternative or counterfactual, more favourable distribution of risk factors [
21]. Thus, the PAF measures the proportion of a given health outcome attributable to an exposure in the study population and the hypothetical reduction if the exposure prevalence could be reduced to zero. The PAF also assumes, that the exposure is related to the outcome; measurement of the risk prevalence and association is unbiased and the reduction of the exposure will have no effect on the distribution of other risk factors [
21,
22]. The PAF of risk factors for a range of diseases and/or outcomes (including breast cancer, diabetes and child mortality) has been measured globally [
23,
24] and for many countries, including Australia [
19], United States of America [
25,
26], and Nigeria [
27].
From a public health viewpoint, estimation of the PAF is of most relevance when the exposure is causally related to the outcome and the factor is amenable to strategic initiatives [
21]. Based on nationally representative population-based data, the strongest modifiable risk factors (in terms of the magnitude of effect size) associated with non-EBF in Nigeria include: (i) low or no maternal education, (ii) middle or poor household wealth, (iii) a lack of antenatal care visits, (iv) delivery at home, and (v) delivery assistance from a non-health professional [
5,
10]. In Nigeria, no published studies have assessed the attributable risk of important modifiable risk factors associated with non-EBF, nor has there been an investigation of scenarios based on feasible impact of strategic community-based initiatives in reducing exposure prevalence. Thus, this study aims to quantify and compare the burden of non-EBF attributable to key modifiable risk factors in Nigeria to inform strategic policy responses and initiatives.
Results
A total sample (
N = 34,653) of maternal responses of infants aged 0–5 months in relation to EBF were examined in Nigeria for the period (1999–2013). More than three quarter (84.4%) of infants 0–5 months of age were not exclusively breastfed. That is, infants who received other water-based liquids in addition to breast milk [Table
1]. The analysis showed that 22.8% (95% CI: 9.2–37.0%) of all estimated cases non-EBF in Nigeria could be attributed to maternal primary level of education and no maternal education [Table
2]. Approximately 25.0% (95% CI: 9.5–39.5%) of all estimated cases of non-EBF in Nigeria for the years 1999–2013 was attributable to lower household wealth. For antenatal care visits, 9.7% (95% CI: 1.7–18.1%) of all cases of non-EBF was attributable to fewer numbers (1–3) of antenatal care visits and no antenatal care visits of Nigerian mothers. Of the estimated cases of non-EBF, 18.8% (95% CI: 6.9–30.8%) was attributable to home birthing. Similarly, 16.6% (95% CI: 3.0–31.3%) of non-EBF in Nigeria was attributable to delivery assistance from non-health professionals (traditional birth attendants and untrained health personnel).
Table 2
PAF and PIF for selected modifiable exposures associated with non-EBF in Nigeria (1999–2013)
Low and no maternal education | 4222 | 22.8 (9.2–37.0) | 0.5 (−7.0; 7.7) |
Middle and poor household wealth | 4730 | 24.7 (9.5–39.5) | – |
Lower number (1–3) and no antenatal care visits | 3368 | 9.7 (1.7–18.1) | 5.4 (−3.0; 13.9) |
Home delivery | 3987 | 18.8 (6.9–30.8) | 2.8 (−3.6; 8.9) |
Delivery assistance from non-health professional (traditional birth attendants and untrained personnel) | 4012 | 16.6 (3.0–31.3) | 2.3 (−4.6; 9.2) |
Joint PAF and PIF combinations (in descending order) | Cases of non-EBF (a) | Joint PAF% (95%CI) | Joint PIF% (95%CI) |
Mat. Edu + H. wealth + ANC + Pl. delivery + Del. assistance | 20,319 | 64.5 (50.5–76.2) | – |
Mat. Edu + H. wealth + ANC + Pl. delivery | 16,307 | 57.4 (42.4–70.2) | – |
Mat. Edu + ANC + Pl. delivery + Del. assistance | 15,589 | 52.8 (37.4–66.4) | 10.5 (−5.4; 24.7) |
Mat. Edu + H. wealth + ANC | 12,320 | 47.8 (31.2–61.7) | – |
Mat. Edu + H. wealth | 8952 | 41.9 (25.3–56.9) | – |
ANC + Pl. delivery + Del. assistance | 11,367 | 38.9 (23.5–53.2) | 8.0 (−2.2;17.9) |
Maternal education | 4222 | 22.8 (9.1–37.4) | – |
In combination, the joint PAF showed that 64.5% (95% CI: 50.5–76.2%) of all cases of non-exclusive breastfeeding in Nigeria for the years 1999–2013 could be attributed to the modifiable risk factors of low or no maternal education, poor or medium household wealth, lower frequency of antenatal care visits, home delivery and delivery assistance from non-health professionals (Table
2). Key socioeconomic indicators (i.e., low or no maternal education and poor or middle household wealth) accounted for the largest PAF of non-EBF in Nigeria, followed by the health service factors (home delivery and fewer/no antenatal care visits).
Potential impact fractions assuming improvements in maternal education [
39], the number of antenatal visits, and a higher proportion of deliveries in a health facility with increased training of health professionals [
31,
32] suggested that 10.5% (95% CI: -5.4; 24.7) of cases of non-EBF could be avoided in Nigeria (Table
2). This included 0.5% (95% CI: -7.0; 7.7) associated with a 5% relative increase in the number of women finishing secondary school; 5.4% (95% CI: -3.0; 13.9) associated with a 17% reduction in the proportion of women who had less than four antenatal care; and 2.8% (95% CI: -3.6; 8.9) and 2.3% (95% CI: -4.6; 9.2) associated with a 15% decrease in the proportion of deliveries occurring outside a health facility, with untrained personnel and traditional birth attendants, respectively.
Discussion
This study estimated the number of cases of non-EBF in Nigerian mothers attributable to the key modifiable risk factors of maternal education, household wealth, antenatal visits, home delivery, and delivery assistance from a non-health professional. The largest population attributable fraction was associated with no maternal education, followed by poor household wealth, home delivery and delivery assistance from non-health professionals. Assuming similar impacts of published community-based interventions in improving health service contacts of Nigerian women [
31,
32] and continued improvements in maternal education [
39]; this study also suggests that approximately 11% of cases of non-EBF could potentially be avoided in Nigeria.
This study has a number of limitations. First, non-EBF practice was based on self-report and this could result in measurement bias as mothers may incorrectly recall how the baby was fed in the period referred to in the surveys. Nonetheless, attempts were made to reduce recall bias in the survey through the use of standardised questionnaires, and in the analyses by restricting analyses to the youngest living child aged < 24 months living with the mother, consistent with previous studies [
45,
46]. Second, misclassification bias in the exposure variables may also have occurred – for example, underestimation or overestimation of the number antenatal care visit – which may result in overestimation or underestimation of the association between antenatal care visits and non-EBF.
Third, the study was based on cross-sectional data, and the establishment of a clear causal relationship between exposure variables and non-EBF practice is difficult. However, the associations reported in this study were consistent with previously published studies from other developing countries which used nationally representative prevalence data (Demographic and Health Survey data) [
47,
48]. Additionally, it is possible that the data reflected in this analysis may not represent the most current situation in Nigeria given changes in the socio-demographic and economic status [
49].
Fourth, the burden of non-EBF attributable to key modifiable risk factors (examined in this study) was based on nationally representative data; however, this does not take into account other modifiable factors associated with non-EBF (for example, mothers employment status, a lack of family support and socio-cultural belief systems held for breastfeeding) as reported in published studies from regional areas of Nigeria [
16,
17,
50].
Fifth, unmeasured confounding factors (such as maternal health problems in the early postnatal period or twin birth) are also likely to affect the study findings. The wealth index used in the NDHS is an indirect measure of household wealth, since it is challenging to obtain reliable income and expenditure data in Nigeria.
Sixth, there are no adequate annual data on non-EBF in Nigeria, nor is there contemporaneous data on relevant breastfeeding policy implementation and evaluation in order to estimate the PIF with the prevalence of non-EBF. Information on annual non-EBF practice and long-term breastfeeding policy implementation and evaluation would have provided a specific representation on the avoidable of cases of non-EBF in the present Nigeria.
Seventh, the applicability of the potential impacts of community-based interventions in the Nigerian context may be challenging given the influence of geopolitical differences on infant and young child feeding practices in Nigeria [
10]. However, both population attributable fractions and potential impact fractions were presented in this study to provide an indication of the range of feasible impacts of key community- and facility-based interventions to reduce non-exclusive breastfeeding practice in Nigeria.
Despite these limitations, the study has a number of strengths. Firstly, the prevalence data used (NDHS, 1999–2013) are nationally representative and provide important information on infant feeding practices (including risk factors categories) that could be used to estimate relative risks, as previously reported [
19,
51,
52]. Secondly, selection bias is unlikely in the data as samples were drawn from the 1991 and 2006 national census frames, with response rates in the surveys ranging from 92 to 98%. Additionally, the risk factors assessed in this study are amenable to strategic policy responses and initiatives as suggested by Rockhill et al. [
21], and this study provides context-specific evidence on interventions that can potentially reduce non-EBF practice in Nigeria using summary population health metrics that have been used in a number of previously published studies [
19,
24,
51,
53,
54]. For example, evidence from some of these studies, including the Global Burden of Diseases, Injuries, and Risk Factors Study, have been used to describe and prioritise key health initiatives and policies in many countries, including the United Kingdom, Australia, Rwanda, South Africa, United States of America and Columbia [
55].
Evidence from the current study showed that more than half (65%) of all cases of non-EBF in Nigeria are attributable to a range of modifiable risk factors amenable to public health initiatives and socio-economic reforms – areas where Nigeria receives substantial support from the international community in terms of funding [
4]. For example, in 2015, the World Bank approved USD500, 000 million for Nigeria to provide assistance for the “saving one million lives” project – an initiative to improve key areas of maternal and child health, including nutrition [
56]. A similar large-scale public health intervention conducted in Bolivia, Ghana and Madagascar (funded by an international donor) resulted in a significant increase in not only exclusive breastfeeding, but also improvements in early initiation of breastfeeding [
31]. As the Nigerian government and donor agencies work toward the achievement of this “saving one million lives” project, the Global Nutrition Targets [
57] and the Sustainable Development Goals (SDGs) in Nigeria; considerable improvements in child nutrition and health can be made if specific and measureable objectives are set towards reducing these observed modifiable risk factors in the Nigerian context to reduce non-EBF.
Additional modifiable risk factors associated with non-exclusive breastfeeding in Nigeria not examined in this study (because of a lack of data) include: poor family support, pressure on the mother to resume work, and myths and belief systems held for breastfeeding; for example, the infant is perceived to still be hungry after breastfeeding [
16,
17] or the baby requires additional water to quench thirst [
58]. Although these risk factors were not assessed, large scale community- and facility-level initiatives that specifically target the modifiable risk factors examined in this study can also reduce the unassessed risk factors associated with non-EBF in Nigeria as reported in previous studies [
31,
32]. Another important factor that may be problematic to measure is political context, with studies suggesting that a high-level of political will and commitment plays a major role in reducing non-EBF in other context [
32]. Given the current health initiatives in Nigeria and increasing international funding for maternal and child health initiatives [
4], sustained political resolve at all levels is also needed to achieve substantial reduction in non-EBF in Nigeria.
Geopolitical region, culture and religion play major roles in policy formulations in Nigeria – a country that is largely divided into Muslim-north and Christian-south [
59,
60]. Accordingly, strategic interventions and policies responses that are designed to reduce non-exclusive breastfeeding practice in Nigeria should be context-specific, and must consider the impacts of geopolitical differences, culture and religion on optimal infant and young child feeding practices in Nigeria. A detailed description of locally-relevant interventions and policies to improve EBF in Nigeria has been reported elsewhere [
6]. These include stronger political will and funding, strengthening community and facility-based participation, and refining standalone/integrated infant and young child policy implementations.