Effectiveness of programmes and interventions to support optimal breastfeeding among children 0-23 months, South Asia: A scoping review.

Abstract Most children in South Asia are breastfed at some point in their lives; however, many are not breastfed optimally, including the early initiation of breastfeeding (EIBF) within 1 hr of birth, avoidance of prelacteal feeds (APF), exclusive breastfeeding (EBF) for 6 months, and continued breastfeeding (CBF) up to 2 years of age or beyond. This review identifies and collates evidence on the effectiveness of interventions to support optimal breastfeeding in five countries in South Asia: Afghanistan, Bangladesh, India, Nepal, and Pakistan. A scoping review was conducted of peer‐reviewed and grey literature. The 31 eligible studies included randomized trials and quasi‐experimental designs that were conducted between 1990 and 2015. Data were collated regarding intervention design, characteristics, and effectiveness to support EIBF, APF, EBF, and CBF. Most studies reported a positive impact on breastfeeding outcomes, including 21/25 studies that examined EIBF, 15/19 studies that examined EBF, and 10/10 studies that examined APF. The only study that examined CBF reported no effect. Education, counselling, and maternal, newborn, and child health initiatives were common intervention types with positive effects on breastfeeding outcomes. Interventions were delivered in health facility, community, and home/family environments. Programmes and interventions that reached women and their families with repeated exposure and beginning during pregnancy were more likely to improve EIBF and EBF outcomes. Interventions with no impact on breastfeeding were characterized by short duration, irregular frequency, inappropriate timing, poor coverage, and targeting.

Most children in South Asia (96%) are breastfed at some point in their lives; however, the majority of children are not breastfed optimally from birth to 2 years (Dibley et al., 2010;UNICEF, 2016). Trends in breastfeeding practices in the five largest South Asian countries found that over half of children do not benefit from early initiation of breastfeeding (EIBF), despite recent improvements in Bangladesh, India, and Nepal (UNICEF, 2016). Further, there have been recent declines in exclusive breastfeeding (EBF) in several South Asian countries, and continued breastfeeding (CBF) is significantly lower at 2 years than 1 year of age (UNICEF, 2016).
The WHO Global Strategy for Infant and Young Child Feeding recommends a set of interventions to improve breastfeeding (WHO, 2003). Existing evidence suggests that these interventions should be delivered concurrently and in multiple settings, including households, communities, workplaces, and health systems (Haroon, Das, Salam, Imdad, & Bhutta, 2013;Sinha et al., 2015). Despite this, there is substantial heterogeneity in breastfeeding practices within and across countries in South Asia (Benedict, Craig, Torlesse, & Stoltzfus, 2018;UNICEF, 2016), reflecting the varying impact of policy and programme action to protect, promote and support breastfeeding in the region.
Scoping reviews map a broad range of literature and present an overview of evidence on a topic and, unlike systematic reviews, do not evaluate the quality of studies (Arksey & O'Malley, 2005). In this review, we explore and summarize evidence from peer-reviewed and programme evaluation literature on interventions to improve breastfeeding practices in Afghanistan, Bangladesh, India, Nepal, andPakistan between 1990 and2015 to inform research, programming and policy efforts in the region. We highlight areas for future research and note the nuanced factors that influence intervention effectiveness and optimal breastfeeding practices in South Asia.

| Definitions
Standard WHO breastfeeding indicator definitions were used for EIBF, EBF, and CBF (WHO, 2017), with some exceptions in order to capture the breadth of evidence on what works to improve breastfeeding practices. WHO defines EIBF as the initiation of breastfeeding within 1 hr of birth; this review also included studies that reported EIBF within 30 min, 3 hr, 5 hr, and 24 hr of birth. EBF in infants aged 0-5 months was the primary focus of this review; however, studies that report EBF for narrower groups within this age range were included, as well. Avoidance of prelacteal feeding (APF) was also included and defined as feeding no food/drink to the infant other than breast milk during the first 3 days of life; studies that assessed intake of prelacteal feeds (PF) were also considered. If a study assessed EBF within 3-4 days of birth, the breastfeeding practice was categorized as APF. CBF was examined at 1 year (12-15 months) and 2 years (20-23 months).
Effectiveness to support breastfeeding was defined as a measured improvement (e.g., proportions or odds ratios) of the breastfeeding practice in the study population.

| Literature search
A PICOT search strategy, including elements of the Population, Intervention, Comparison, Outcomes, and Time, was used to identify research articles published between January 1990 and December 2016 from PubMed, Web of Science (CORE Collection), GenderWatch, POPLINE, and Sociological Abstract. Search terms were applied with various Boolean operators for country location, programmatic approach, breastfeeding practice, and date of publication ( Figure S1). In addition, grey literature published between was January 1990 and December 2016 on programme evaluations was sourced from experts and organizational websites.

| Study selection
Eligibility criteria included study location (Afghanistan, Bangladesh, India, Nepal, Pakistan), study design (individual and cluster randomized controlled trials and quasi-experimental designs for programme evaluations), year conducted , at least one measured breastfeeding practice of interest (EIBF, APF or PF, EBF, CBF), and full-text availability in English. Duplicate citations were removed, and titles and abstracts were screened to identify relevant studies.
Two researchers independently screened the full texts of all potentially relevant articles to assess eligibility. Any discrepancies between reviewers were resolved by discussion and consultation with a third researcher.

Key messages
• Multiple interventions, including education and counselling, community mobilization, mass media, and maternal and newborn health initiatives, delivered across multiple implementation environments (health facility, community, and home/family), are effective in improving breastfeeding practices in South Asia.
• Evidence from studies that did not report a positive impact on breastfeeding practices indicate that intervention timing, frequency, duration, and targeting influence effectiveness.
• Evidence gaps on the impact of interventions on breastfeeding outcomes were identified for specific countries (Afghanistan, Nepal, and Pakistan), breastfeeding outcomes (avoidance of prelacteal feeds and continued breastfeeding), and implementation environments (policy and workplace). These intervention types are not mutually exclusive, for example, several MNCH initiatives involved education and counselling.

| Characteristics of included studies
After excluding duplicates, 1,157 articles were screened and 31 studies were included ( Figure S2). All included studies were peer-reviewed publications or programme evaluations. The majority of included studies were conducted in India (n = 14) and Bangladesh (n = 11). Fewer studies took place in Pakistan (n = 4) and Nepal (n = 2), and no eligible studies were conducted in Afghanistan. In total, 22 individual or cluster randomized controlled trials and nine quasi-experimental programme evaluations were included (  (Carvalho et al., 2014;More et al., 2012;Sikander et al., 2015;Talukder et al., 2016;Vir et al., 2014).
Included studies were characterized by the implementation environment (health facility, community, and home/family) and type of intervention (education and counselling, community mobilization, mass media, and MNCH initiatives; Table S3). Several studies examined multiple implementation environments and intervention types. The results are presented by implementation environment and type of intervention.

| Early initiation of breastfeeding
Twenty-five studies reported EIBF practices, of which 21 studies showed a positive effect. The most common intervention type was interpersonal education and counselling (n = 18), followed by MNCH initiatives (n = 7), community mobilization (n = 4), and mass media (n = 1). These interventions were implemented in the home/family (n = 12), community (n = 11), and the health facility (n = 9) environments.
Education and counselling in the home/family environment was effective in improving EIBF outcomes in Bangladesh (Haider et al., 2000;Talukder et al., 2016) and India (Vir, 2013), and in combination with community mobilization in India (Vir et al., 2014) and Pakistan (Bhutta et al., 2008). However, a cognitive-behavioural counselling intervention offered to pregnant women during home visits in Pakistan was not effective in improving EIBF (Sikander et al., 2015); although women received a total of seven visits, only one session was delivered before birth and one session was delivered immediately after delivery (Sikander et al., 2015). A study in Nepal reported no effect on EIBF after a sanitation and maternal and child care education intervention was delivered to women and other household members at home for 3 hr fortnightly for 6 months (Jha et al., 2006). The same study found no significant findings for six other behavioural outcomes, except care during pregnancy and introduction of complementary foods, suggesting the message content may have been too broad (Jha et al., 2006).
Community-level education and counselling showed mixed results on EIBF proportions. In India, a women's group programme that provided community education and counselling fortnightly for 6 months with community mobilization activities had no effect on EIBF (More et al., 2012). Poor coverage, non-specific target groups and membership attrition were characteristic of the programme and may have contributed to this finding (More et al., 2012). On the other hand, positive effects on EIBF were reported by community-level education and counselling interventions in Bangladesh (Fottrell et al., 2013) and India (Agrawal et al., 2012;Balakrishnan et al., 2016), and when combined with community mobilization in India (Bhandari et al., 2003). All studies that examined interventions to deliver education and counselling at health facility level showed positive effect on EIBF in Bangladesh (Akter et al., 2012;Jahan et al., 2014;Thakur et al., 2012) and India (Prasad & Costello, 1995), as well. A programme in Bangladesh that combined education and counselling on EIBF at home/family level with mass media interventions also had a positive impact on EIBF (Menon et al., 2016).
MNCH initiatives, such as the Improving Maternal, Neonatal, and      individually or in small groups in at least 5 contacts. Staff were trained on the benefits and feasibility of early breastfeeding, and the dangers of prelacteal feeds, together with instructions on explaining this information to mothers. Midwives and nurses were trained to motivate, persuade, and help mothers to EIBF.
At 6 months following the intervention 36% of women reported receiving education vs. 64% that did not report receiving education.
EIBF proportion (within 12 hr): At 6 months following the intervention, higher among mothers who received education (78%) vs. women that received no education 17% (P < 0.001). APF proportion: At 6 months following the intervention, higher among mothers who received education (58%) vs. women that received no education 3% (P < 0.001  • Trained staff in improving health worker skills for case management of neonatal and child illness.
• Health system strengthening to implement IMCI including improved supervision of CHWs and performance-based incentives.
Three monthly women's group meetings were conducted to improve community awareness of available services.
CG: Routine care provided by AWW, ASHA, ANW, or primary health care physician.
EIBF proportion: The intervention resulted in a larger effect on EIBF in poorer families (difference in inequity gradients 3.0%, CI [1.5, 4.5], P < 0.001), in lower caste and minorities families, and in infants of mothers with fewer years of schooling.
Health facility: Education and counselling IG: Facility-based nutrition education on EBF was provided twice a month for 2 months after delivery. Education emphasized APF, EIBF, breast attachment, positioning, EBF for 6 months, increasing frequency and quality of maternal diet during lactation, food hygiene, personal hygiene, and necessary family assistance for breastfeeding.
One IMNCI intervention in India that was delivered across all three implementation environments reported a higher increase in EIBF in intervention clusters among poorer families, lower caste and minority families, and among infants of mothers with fewer years of schooling . Skin-to-skin contact (SSC) also showed positive impacts on EIBF through home/family and community-level interventions in India (Mahmood et al., 2011) and health facility-level interventions in Pakistan (Kumar et al., 2008;Mahmood et al., 2011). In India, both the Janani Suraksha Yojana (JSY) conditional cash transfer programme and Yashoda birth companion programme promoted maternal and newborn health through health systems changes (Carvalho et al., 2014;Varghese et al., 2014). Though JSY showed a positive effect on EIBF proportions (Carvalho et al., 2014), the Yashoda programme showed no overall effect on EIBF except EIBF within 5 hr among infants delivered via caesarean section (Varghese et al., 2014).

| Prelacteal feeding (APF)
All 10 studies in Bangladesh, India, and Pakistan that examined APF (or PF) reported a positive impact of the programme intervention. The most common intervention type was interpersonal education and counselling (n = 7), followed by MNCH initiatives (n = 3), community mobilization (n = 1), and mass media (n = 1). These interventions were delivered in the home/family (n = 5), community (n = 4), and health facility (n = 3) environments.
Education and counselling interventions delivered at home/family level improved APF in in Bangladesh (Akter et al., 2012;Haider et al., 2000;Menon et al., 2016;Talukder et al., 2016) and Pakistan (Sikander et al., 2015). Only one study in India reported positive effects from community-level education and counselling (Bhandari et al., 2003). Two studies, one in Bangladesh (Akter et al., 2012) and the other in India (Prasad & Costello, 1995), reported that education and counselling in the health facility setting had positive effects on APF. Only two studies reported positive effects of community mobilization (Bhandari et al., 2003, India) and mass media interventions (Menon et al., 2016, Bangladesh) on APF.
MNCH initiatives that promoted SSC and kangaroo mother care (KMC) in the home/family setting in India (Kumar et al., 2008), the community in Bangladesh and India (Ahmed et al., 2011;Kumar et al., 2008), and health facilities in India (Srivastava et al., 2014) all impacted positively on APF.

| Exclusive breastfeeding
In total, 19 studies reported effectiveness to support EBF, and all but four studies showed a positive effect. The most common intervention type was interpersonal education and counselling (n = 14), followed by MNCH initiatives (n = 6), community mobilization (n = 3), and mass media (n = 1). These interventions were delivered in the home/family (n = 9), community (n = 7), and health facility (n = 9) environments.
In India, community-level education counselling through women's groups combined with community mobilization that addressed several health and nutrition issues showed no effect on breastfeeding practices (neither EIBF nor EBF at 4 weeks; More et al., 2012;Vir et al., 2014). A programme that used community volunteers to conduct family-level counselling and community mobilization also reported no impact on EBF at 4-5 months (More et al., 2012, Vir et al., 2014. There was no effect on EBF of training community volunteers and traditional birth attendants to deliver education and counselling services to women at home in Bangladesh; however, the number of visits following delivery was irregular (Talukder et al., 2016). Other studies reported on the positive effects of education and counselling at home/family level on EBF in Bangladesh and Pakistan (Haider et al., 2000;Khan et al., 2016;Sikander et al., 2015) and when combined with mass media in Bangladesh (Menon et al., 2016) or community mobilization in Pakistan (Bhutta et al., 2008). Community-level education and counselling was also effective in Bangladesh (Fottrell et al., 2013) and when combined with community mobilization in India (Bhandari et al., 2003).

MNCH initiatives, including SSC and KMC interventions, also
showed mixed results. One KMC intervention in India encouraged mothers of very low birth weight infants to initiate KMC immediately after delivery; the intervention reported positive EIBF findings, yet there was no impact on EBF at 6 months (Gavhane et al., 2016). The JSY programme in India showed positive effects on child health outcomes including EIBF, but no effect on EBF (Carvalho et al., 2014).
The authors speculate that CHWs may have provided more counselling on topics such as immunizations rather than breastfeeding (Carvalho et al., 2014). On the other hand, an SCC intervention at facility level in India reported a positive and significant effect on EBF at 6 weeks among healthy infants (Srivastava et al., 2014) as did other studies providing IMNCI packages at home/family and health facility level in India (Arifeen et al., 2009;Mazumder et al., 2014) and community and health facility level in Bangladesh (Arifeen et al., 2009, Mazumder et al., 2014.

| Continued breastfeeding
One study in India reported CBF practices: an IMNCI intervention package delivered by CHWs to breastfeeding mothers at home, combined with CHW supervision, found no significant effect on CBF at 1 year of age (Mazumder et al., 2014). sessions. These findings support existing research that demonstrates the effective use of education and counselling services to promote optimal breastfeeding practices (Haroon et al., 2013;Sinha et al., 2015). Expanding programmatic reach of these interventions through community mobilization and mass media interventions also demonstrated effectiveness to promote optimal breastfeeding practices;

| DISCUSSION
however, these community-based strategies were not implemented as often as other intervention types. The findings highlight key components of successful interventions and suggest the need for more community-level interventions to maximize the benefits education and counselling interventions.
MNCH initiatives were effective at supporting EIBF and APF but had mixed results for EBF. Most of these initiatives, including promotion of KMC and SSC, offered breastfeeding support immediately following delivery and during the postnatal period; as such, these initiatives had the most opportunity to impact on early feeding behaviours. These findings support existing guidelines supporting Among studies that reported no impact or negative effects of interventions on breastfeeding practices, several common characteristics were identified. These included inadequate or irregular frequency and/or duration of contacts for counselling visits and poor timing of intervention delivery relative to the mother's gestational period and the child's age to ensure breastfeeding practices are implemented early (EIBF and APF) and maintained for the recommended periods (EBF and CBF). In addition, poor coverage and targeting of interventions to pregnant and breastfeeding women, particularly for community mobilization programmes, and intervention packages that lacked context-specific messages or covered too many health and nutrition behaviours contributed to suboptimal outcomes. The findings illustrate weaknesses in critical factors of intervention design and implementation and suggest that addressing these factors may increase the effectiveness of breastfeeding interventions in the region.
Several key knowledge gaps were identified in this review. The lack of evidence from Afghanistan is noteworthy as recent trends in Afghanistan show declining proportions in optimal breastfeeding practices (Benedict et al., 2018). There were also few studies conducted in Nepal and Pakistan, and limited evidence on APF and CBF outcomes across the South Asia region. In addition, there were no studies that examined interventions delivered in the policy or workplace environments. Additional research in these countries and on these specific breastfeeding practices and implementation environments is required to inform regional policy and programme recommendations. Furthermore, there are knowledge gaps on how interventions improved breastfeeding practices; more research and documentation on implementation pathways will contribute evidence to design more effective and scalable interventions and programmes.

| LIMITATIONS
In line with scoping review methodology, our review did not grade the quality of evidence but instead described knowledge on the effectiveness of interventions to strengthen breastfeeding practices in South Asia. In order to provide the highest level of evidence on intervention effectiveness, we limited our eligibility criteria to include only randomized controlled trials and quasi-experimental studies as these designs provide the most precise estimates of the likely effects of an intervention with limited risk of bias.

| CONCLUSION
Mothers require active support to establish and sustain appropriate breastfeeding practices from birth until a child's second birthday. This Notable evidence gaps were identified for specific countries (Afghanistan, Nepal, and Pakistan), breastfeeding practices (APF and CBF), and implementation environments (policy and workplace). Addressing these gaps will generate additional evidence on effective interventions and pathways to protect, promote, and support optimal breastfeeding practices in South Asia.