Background
Breastfeeding is widely recognized as one of the most costeffective investments to improve child survival (UNICEF, 2013). According to The Lancet Series on Maternal and Child Under Nutrition, the adoption of exclusive breastfeeding through six months of age and continued breastfeeding through age two in 36 high burden countries could avert 11.6% of all deaths in children under the age of one, and nearly 10% of all deaths in children under the age of two (Bhutta et al., 2008a). The benefits of breastfeeding stretch beyond a child’s survival. Breastfed children do better on cognitive and motor development tests, and generally achieve better academic outcomes than non-breastfed children (Horta et al., 2013).
Additionally, breastfeeding imparts critical benefits to the woman. When practiced exclusively, breastfeeding is associated with lactational amenorrhea, the natural postnatal infertility that occurs when a woman is amenorrheic (not menstruating). The lactational amenorrhea method (LAM) is considered a modern method of contraception that prevents a second pregnancy within six months of giving birth, and helps to conserve maternal iron stores (Dewey et al., 2001). Increased birth spacing improves both the health of the woman and her infant’s chances of survival, while suboptimal pregnancy intervals are associated with an increased risk of adverse perinatal and infant/child outcomes including preterm birth, low birth weight, small for gestational age (SGA), stunting, and underweight infants/ children (Bhutta et al, 2013; Conde-Agudelo, 2006; Conde- Agudelo et al., 2012; Rutstein, 2008).
The World Health Organization (WHO) recommends three primary breastfeeding practices: initiation of breastfeeding within one hour after birth (also referred to as immediate breastfeeding), exclusive breastfeeding (EBF) through six months of age, and continued breastfeeding until 24 months of age (Dyson et al., 2005; WHO, 2008). Evidence associated with the recommendation for immediate breastfeeding is limited and likely operates through the effect of exclusive breastfeeding (Bhutta et al., 2013). The Lancet, therefore, includes only EBF and continued breastfeeding in the modeling of optimal breastfeeding practices cited above.
Despite the promise of optimal breastfeeding practices, rates for the three WHO recommended breastfeeding practices remain low, and negligible progress has been made to increase these rates over the past two decades (UNICEF, 2013). According to an analysis of data from 78 low and middle income countries, rates of immediate breastfeeding range from a mean of 36% in Eastern Europe to a mean of 58% in Latin America, and the rate of exclusive breastfeeding in children one to five months of age is just 30% (Black et al, 2013). According to the UNICEF global database of national surveys from 2007-2011, the current rate of breastfeeding at one year is 76%, while the rate of breastfeeding at age two years is 58%.12
This chapter reviews the effectiveness of social and behavior change communication (SBCC) approaches on improving breastfeeding practices.
Search Results
Sixty-two peer-reviewed studies met the literature review’s inclusion criteria.13 A complete list of the studies reviewed, including a summary of the SBCC approach(es) utilized, the study design, practices measured, and levels of significance is presented in Table 3.2.4.
Forty-eight reported a statistically significant change in at least one of the priority breastfeeding practices. Each of those articles is described in greater detail below. Studies reporting a statistically significant change in more than one of the prioritized practices and/or utilizing more than one SBCC approach are cited and discussed multiple times. Where this occurs, we have provided a detailed description of the study in the first reference, and only discuss results in subsequent references.
The findings are organized by practices targeted, SBCC approach(es) utilized, and study design. This organization enables the reader to derive an independent judgment regarding findings from each study. In addition, implementation processes followed are also discussed.
Of the 62 studies reviewed for this chapter, six were literature reviews that included meta-analysis, 27 were randomized controlled trials (RCT), 14 were longitudinal studies, 13 were repeated cross-sectional studies, and two were cross-sectional studies. See Table 3.2.1.
The literature reviews were published between 2010 and 2012. Twenty-one of the 56 studies included in the previously published reviews met the inclusion criteria for the present review, and these studies are included in the summaries below.
Table 3.2.4: Studies reporting on breastfeeding practices
| Title | Approaches Used | Scale/Scope | Evaluation Design | Outcomes Measured | Results | P-values/CIs |
|---|---|---|---|---|---|---|
| Abolyan (2006); Russia | Counseling in facilities by providers, Support group in facilities by providers and/or lay volunteers, Small media | 8 maternity hospitals | Longitudinal – 741 healthy postpartum women (383 intervention, 358 control) |
% EBF from birth until discharge from hospital Median time until initiation of breastfeeding |
88.9% vs. 32.6% 2 hrs. vs. 12 hrs. |
NS NS(28) |
| Agrasada et al. (2005); Philippines | Counseling in homes by lay volunteers | 1 hospital | RCT – 179 postpartum women (60 IG1, 60 IG2, 59 control) |
% EBF during previous 7 days at 6 months Likelihood of EBF |
counseling group vs. “other groups”(29): 44% vs. 0% counseling group vs. “other groups”(29): OR=6.3 (based on generalized estimating equation models) |
-- p<0.001 (95% CI: 3.53- 11.3) |
| Aidam et al. (2005); Ghana | Counseling in facilities by providers, Counseling in homes by providers | 2 hospitals | RCT – 123 pregnant women attending prenatal clinics: 40 IG1, 39 IG2, 44 control)(30) |
% EBF during previous month at 6 months % EBF during previous 24 hours at 6 months % EBF “since birth” at 6 months |
IG1 vs. IG2 vs. control: 90.0% vs. 74.4% vs. 47.7% IG1 vs. IG2 vs. control: 90.0% vs. 79.5% vs. 54.5% IG1 vs. IG2 vs. control: Pooled intervention vs. control 39.5% vs. 19.6% |
p=0.008 p=0.001 p=0.02 |
| Aksu et al. (2011); Turkey | Counseling in homes by peers, Counseling in facilities by providers | 1 hospital | RCT – 60 women who gave birth in study hospital (30 intervention, 30 control |
% EBF at 6 months Duration of EBF (months) Mean duration of breastfeeding (months) % EBF at 2 weeks % EBF at 6 weeks |
43% vs. 23% 4.7 vs. 3.9 15.1 vs. 12.1 67% vs. 40% 60% vs. 33% |
p=0.04 p=0.014 p=0.001 p=0.0038 p=0.038 |
| Akter et al. (2012); Bangladesh | Counseling in facilities by providers, Group education in facilities by providers | 1 hospital | Longitudinal – 115 women at six months gestation (57 intervention, 58 control) |
% breastfeeding initiation within one hour of birth % EBF at 1 month |
75.4% vs. 34.5% 64.9% vs. 37.9% |
p=0.001 p=0.003 |
| Alam et al. (2002); Bangladesh | Counseling in facilities by providers, Support group in facilities by providers and/or lay volunteers, Small media | 2 hospitals (1 intervention, 1 control) | Longitudinal – 209 women who gave birth in study hospital (105 intervention, 104 control) |
% EBF at 1 month (30 days) % EBF at 2 months (60 days) % EBF at 3 months (90 days) % EBF at 4 months (120 days) % EBF at 5 months (150 days) % EBF > 5 months Median duration of EBF |
82.3% vs. 50.0% 63.2% vs. 40.0% 34.7% vs. 21.5% 20.8% vs.11.4% 16.2% vs. 6.5% 8.1% vs. 6.5% 69.74 vs. 48.4 days |
-- -- -- -- -- p<0.001 p<0.001 |
| Albernaz et al. (2003); Brazil | Small media, Counseling in facilities by providers, Counseling in homes using hotline | 1 city | RCT – 157 newborns (82 intervention, 75 control) |
% breastfeeding at 4 months (other food or milk in addition to breast milk) Likelihood of having stopped breastfeeding at 4 months % EBF at 4 months |
32% vs. 28% Prevalence ratio:: 1.85 40% vs. 31% |
NS p=0.04 (95% CI: 1.01- 3.41) NS |
| Albernaz et al. (2008); Brazil | Small media, Counseling in facilities by providers, Counseling in homes using hotline | 1 city | Longitudinal – 421 from 1982 cohort, 561 from 1993 cohort, 310 from 1997-1998 cohort, 317 from 2004 cohort |
% breastfeeding at 6 months % breastfeeding at 24
months % breastfeeding at 24 months % EBF at 1 month % EBF at 3 months % EBF at 6 months |
Post vs. immediately prior to intervention(31): 64% vs. 46% Post vs. immediately prior to intervention(31): 33% vs. 22% Post vs. immediately prior to intervention(31): 11% vs. 14% Post vs. immediately prior to intervention(31): 40% vs. 26% Post vs. immediately prior to intervention(31): 19% vs. 16% Post vs. immediately prior to intervention(31): no difference |
-- -- -- -- -- -- |
| Alderman (2007); Uganda | Group education in communities by lay volunteers, Mobilization of issue groups, Support group in communities by peers | 34 districts | Longitudinal – 2,250 households with a child under 6 (750 IG1, 750 IG2, 750 control) |
% of infants weaned before 6 months (~BF at 6 months) |
“significant decline” |
-- |
| Azad et al. (2010); Bangladesh | Group education in communities by lay volunteers | 18 clusters from three rural districts (9 intervention, 9 control) | RCT – 36,113 births (17,514 intervention, 18,599 control) |
% EBF for six weeks |
68·0% vs. 61·5% |
RR: 1·10 (0·98–1·24) Adjusted RR: 1·10 (0·98– 1·23) |
| Balaluka et al. (2012); Democratic Republic of Congo | Mobilization of issue groups, Counseling in communities by peers, Counseling in homes by peers | 2 districts (1 intervention, 1 control) | Longitudinal – 385 newborns (208 intervention, 178 control) |
% EBF at 4 months % EBF at 5 months % EBF at 6 months Median duration of EBF |
91.8% vs. 50.7% 81.3% vs. 9.6% 57.7% vs. 2.7% 6 vs. 4 months |
p<0.001 p<0.001 p<0.001 p<0.001 |
| Bang et al. (2005); India | Mobilization of issue groups, Counseling in communities by peers, Counseling in homes by peers | 86 villages (39 intervention villages, population: 38,998; 47 control villages, population: 42,149) | RCT – 2,361 neonates; baseline (1995-1996): 763; 1st follow-up (1996– 1997): 685; 2nd follow-up (1997–1998): 913 |
% breastfeeding initiation within six hours of birth |
2nd follow-up vs. 1st follow-up vs. baseline: 89.7% vs. 81.9% vs. 47.5% |
Significance of trend p <0.001 |
| Baqui et al. (2008a); Bangladesh | Counseling in homes by lay volunteers, Events targeting policy makers and regulators, Mobilization of campaign, event, or special 'days' | 24 unions (population: approximately 20,000 each) from three rural upazilas of Sylhet district (8 unions per group) | RCT – 14,769 live births from home care intervention, 16,325 live births from community care intervention, 15,350 from control |
% breastfeeding initiation within one hour of birth - Home care vs. control % breastfeeding initiation within one hour of birth - Community care vs. control |
81% vs. 57% 71% vs. 57% |
p<0.0001 p=0.0158 |
| Baqui et al. (2008b); India | Counseling in homes by lay volunteers, Counseling in homes by providers | 2 districts in Uttar Pradesh state (1 intervention, 1 control) | Repeated cross-sectional – baseline: 14,952 women who had had a live birth or stillbirth within the reference period (8,756 intervention, 6,196 control); endline:13,826 women who had had a live birth or stillbirth within the reference period (7,812 intervention, 6,014 control) |
%breastfeeding initiation within one hour of birth %breastfeeding initiation within one hour of birth |
At baseline: 3.1% vs. 2.4% At endline: 37.7% vs. 5.8% |
Difference in differences: p<0.00132 Difference in differences: p<0.00132 |
| Bashour et al. (2008); Syria | Counseling in homes by providers | 1 hospital in Damascus | RCT – 876 women who had recently given birth (285 in IG1, 294 in IG2, 297 in control) |
% EBF at 4 months (using no other fluids except breast milk for the baby) |
IG1 vs. IG2 vs. CG33: 28.5% vs. 30.1% vs. 20.2% |
p=0.023 |
| Berti et al. (2010); Ethiopia, Ghana, Malawi and Tanzania | Meetings targeting policy makers and regulators, Events targeting policy makers and regulators, Support group in communities, Mobilization of issue groups, Mass media | 4 countries, 4 million beneficiaries | Repeated cross-sectional – secondary data from 2-3 time points from program evaluations of moderate or high quality and Demographic Health Surveys (DHS); program survey sample sizes per country were between 900 and 4801 randomly selected households per survey |
% EBF for 6 months - Ethiopia % EBF for 6 months - Ghana % EBF for 6 months - Malawi % EBF for 6 months - Tanzania |
endline vs. midline vs. baseline: 49% vs. 38% vs. 25% endline vs. midline vs. baseline: 49% vs. 27% vs. 17% endline vs. midline vs. baseline: 70% vs. 47% vs. 15% endline vs. midline vs. baseline: 21% vs. 15% vs. --- |
-- -- -- -- |
| Bhandari et al. (2001); India | Counseling in homes by providers | 1 community in Delhi | RCT – 418 infants four months of age (96 in food supplement group; 95 in nutritional counseling group; and 96 in visitation group; 100 in control group); data collected when infants were 6, 9 and 12 months of age |
% breastfeeding… at 6 months % breastfeeding… at 9 months % breastfeeding…at 12 months |
Supplement + counseling vs. counseling alone vs. visitation vs. control: 93.8% vs. 97.9% vs. 99.0% vs. 93.0%, Supplement + counseling vs. counseling alone vs. visitation vs. control: 92.0% vs. 95.5% vs. 97.9% vs. 91.4% Supplement + counseling vs. counseling alone vs. visitation vs. control: 83.9% vs. 96.9% vs. 96.7% vs. 89.2% |
(Comparisons with final group (visitation)): NS, NS, NS (Comparisons with final group (visitation)): NS, NS, NS (Comparisons with final group (visitation)): p<0.05, NS, NS |
| Bhandari et al. (2003); India | Group education in communities by providers, Counseling in homes by peers | 8 communities (4 intervention, 4 control) | RCT – 1,115 infants born within 9 months of start of intervention; 3 month assessment: 895 infants (483 intervention, 412 control); 6 month assessment: 880 infants (468 intervention, 412 control) |
% EBF at 3 months % EBF at 4 months % EBF at 5 months % EBF at 6 months Median duration of EBF EBF among low-birth weight babies |
79% vs. 48% 69% vs. 12% 49% vs. 6% 42% vs. 4% 122 days vs. 41 days higher rates at every time point |
p-<0.0001 p-<0.0001 p-<0.0001 p-<0.0001 p-<0.0001 p<0.001 |
| Bhandari et al. (2005); India | Counseling in facilities by providers, Counseling in homes by peers, Mobilization of issue groups | 8 communities (4 intervention, 4 control) | RCT – 1025 newborns |
Mean duration of EBF (days) % EBF at 3 months % EBF at 4 months |
0 vs. 1-2 vs. 3+ channels: 114.1 vs. 127.3 vs. 128.7 days 0 vs. 1-2 vs. 3+ channels: 70.8% vs. 81.3% vs. 93.6% 0 vs. 1-2 vs. 3+ channels: data not reported |
p=0.090, adj. p= 0.064 p=0.0002, adj. p=0.002 p=0.001 |
| Bhutta et al. (2008b); Pakistan | Group education in communities by providers, Counseling in homes by lay volunteers, Mobilization of issue groups, Small media, Mid-sized media | 8 village clusters from 2 sub-districts in Sindh province (4 intervention with a total population of 138,600, 4 control) | Repeated cross-sectional – data were collected at baseline (Jun-Aug 2003) and three other time points; Endline data (Sept 2005) reported are from 395 and 375 women from the intervention and control village clusters |
% EBF for first 4 months % breastfeeding initiation within one hour of birth |
64.6% vs. 18.1% 66.1% vs. 21.1% |
Not reported Not reported |
| Bortolini and Vitolo (2012); Brazil | Counseling in homes by providers | 1 hospital | RCT – 397 newborns (163intervention, 234 control) |
% EBF at… < 1 month % EBF at…= or > 4 months % EBF at…= or > 6 months % breastfeeding at… - at 6 months % breastfeeding at…- at 12 months |
33.3% vs. 48% 45.1% vs. 28.6% 19.1% vs. 8.2% 66.3% vs. 55.6% 52.8% vs. 41.9% |
p=0.004 p=0.001 p=0.001 p=0.04 p=0.04 |
| Braun et al. (2003); Brazil | Counseling in facilities by providers, Support group in facilities by providers and/or lay volunteers, Small media | 1 hospital | Longitudinal – 437 newborns (intervention: 250 born in 1999, 2 years after BFHI implementation; control: 187 born in 1994) |
Duration of EBF |
2 vs. 1 month |
p=0.01 |
| Coutinho et al. (2005); Brazil | Counseling in homes by lay volunteers | 2 hospitals | RCT – 350 women who gave birth in target hospitals (175 intervention, 175 control) |
% EBF at 6 months % breastfeeding at 6 months |
45% vs. 13% 78% vs. 62% |
p<0.0001 p<0.0001 |
| Crookston et al. (2007); Cambodia | Community/social mobilization, Group education in communities by lay volunteers, Counseling in communities by lay volunteers, Small media, Support group in facilities by providers and/or lay volunteers | 4 districts (3 intervention, 1 control) | Repeated cross-sectional – baseline (summer of 2004): 440 women with a child less than 6 months of age (206 intervention, 234 control); 1 year followup: 467 women (222 intervention, 245 control) |
% breastfeeding initiation within one hour of birth % breastfeeding initiation within one hour of birth EBF in the previous 24 hours among infants < 6 months of age EBF in the previous 24 hours among infants < 6 months of age |
At baseline: RR=1.1 At follow-up: RR=1.62 At baseline: RR=1.54 At follow-up: RR=1.81 |
(95% CI: 0.74–1.68) (95% CI: 1.30–2.01) (95% CI: 1.21–1.96) (95% CI: 1.49–2.21) |
| de Oliveira et al. (2006); Brazil | Counseling in facilities by providers, Counseling in homes by providers | 1 hospital | RCT – 211 mother-infant pairs (74 intervention, 137 control) |
% EBF at 7 days postpartum % EBF at 30 days postpartum |
79.7% vs. 82.5% 60.8% vs. 53.3% |
NS NS |
| Duyan Camurdan et al. (2007); Turkey | Counseling in facilities by providers, Support group in facilities by providers and/or lay volunteers, Small media | 1 hospital | Cross-sectional – 297 babies born in the four months after BFHI (November 2002-February 2003) (intervention) and 258 babies born before BFHI (November 2001-February 2002) (control) |
% EBF at 15 days old % EBF at 2 months Mean duration of breastfeeding (months) % EBF-at 4 months % EBF at 6 months |
after-BFHI group vs. before-BFHI: 97.4% vs. 93.1% after-BFHI group vs. before-BFHI: 76.1% vs. 67.2% after-BFHI group vs. before-BFHI: 21.17 vs. 17.83 months OR=1.5 after-BFHI group vs. before-BFHI: 53.7% vs. 45.6% after-BFHI group vs. before-BFHI: 9.3% vs. 9.8% |
NS NS p=0.0036 (95% CI: 1.16–2.03) NS NS |
| Gogia and Sachdev (2010); India, Pakistan, Bangladesh | Counseling in homes by lay volunteers | varied | Literature review – four RCTs |
Breastfeeding initiation within one hour of birth |
Pooled RR : 3.35 |
p=0.012 (95% CI: 1.31- 8.59) |
| Guldan et al. (2000); China | Counseling in homes by peers | 4 townships (2 intervention, 2 control) | RCT – newborns were enrolled throughout the year of project implementation, at completion of the year infants had been monitored for 4-12 months; 495 infants born during study period analyzed (250 intervention, 245 control); infants 4-6 months old: 57 intervention, 69 control; infants 7-9 months old: 85 intervention, 76 control; infants 10-12 months old: 108 intervention, 100 control |
% breastfeeding initiation within one hour of birth % breastfeeding initiation within 24 hour or more of birth % gave colostrum to infant % breastfeeding at 4-12 months |
23% vs. 18% 23% vs. 35% 91% vs. 80% 83% vs. 75% |
NS p=0.031 p<0.004 p=0.034 |
| Gupta et al. (2004); Uganda | Mass media | 12 districts | Cross-sectional – 212 women aged 15-49 years whose youngest child was under six months of age |
EBF for 6 months EBF for 6 months EBF for 6 months |
Exposure to 1 BCC type: OR=1.00 Exposure to >1 BCC type: OR=2.79 No BCC message exposure: OR=2.32 |
NS p<0.05 p<0.05 |
| Guyon et al. (2009); Madagascar | Mobilization of issue groups, Mobilization of campaign, event, or special 'days', Mass media | 2 highland provinces; targeted population varied throughout project from 1.4 million in 6 districts to 6 million in 23 districts | Repeated cross-sectional – baseline (2000): 1,200 children under two years old; endline (2005): 1,760 children under two years old |
% breastfeeding of children 20-23 months of age % EBF of infants < 6 months old % breastfeeding initiation within one hour of birth % breastfeeding of children 12-15 months of age |
endline vs. baseline: 73% vs. 43% endline vs. baseline: 70% vs. 42% endline vs. baseline: 68% vs. 33% endline vs. baseline: 94% vs. 89% |
p<0.001 p<0.001 p<0.001 p<0.05 |
| Haider et al. (2000); Bangladesh | Counseling in homes by peers | 40 zones in Dhaka (20 intervention, 20 control) | RCT –573 women in last trimester of pregnancy who completed all five months of study (288 intervention, 285 control) |
% EBF at 5 months % breastfeeding initiation within one hour of birth |
70% vs. 6% 64% vs. 15% |
p<0.0001 p<0.001 |
| Hall (2011); India, Syria, Pakistan, and Bangladesh | Counseling in communities, Counseling in homes by peers, Group education in communities | varied | Literature review – four RCTs |
EBF in low and middle-income countries (effect of community-based interventions) |
pooled OR= 5.9 |
(95% CI: 1.81- 18.6) |
| Haque et al. (2002); Bangladesh | Counseling in facilities by providers | 3 counseling centers | Longitudinal – 114 mother-child pairs completed one-year follow-up (59 intervention, 55 control) |
% EBF for 5 months % gave colostrum to infant |
54.2% vs. 36.4% 100% vs. 98.9% |
NS -- |
| Harkins et al. (2008); Honduras | Mobilization of issue groups, Counseling in homes | 1 community | Repeated crosssectional(34) –Honduras: 300 households with children under two years old interviewed in August 2004 and December 2005 (post-intervention) |
EBF for 6 months |
endline vs. baseline: 19% vs. 11% |
p<0.05 |
| Imdad et al. (2011); various | Counseling, Group education/counseling | varied | Literature review – 53 RCTs and quasi-RCTs |
Breastfeeding at 6 months - in developed and developing countries EBF at 4-6 weeks - in developed and developing countries EBF at 4-6 weeks- in developing countries EBF at 6 months- in developed and developing countries EBF at 6 months- in developing countries |
RR = 1.12 RR=2.37 RR = 1.89 RR=1.43 RR = 6.32 |
(95% CI: 1.01 -1.24) (95% CI: 1.33- 4.24) p<0.001 (95% CI: 1.50- 2.37) p<0.001 (95% CI: 1.28- 1.60) p<0.001 (95% CI: 3.35- 11.93) p<0.001 |
| Jolly et al. (2012); various | Counseling in homes by peers, Counseling in communities by peers, Support group in communities by peers | varied | Literature review – 17 RCTs |
Risk of not breastfeeding at last study follow-up in low income countries (Peer support vs. none) Risk of not EBF at last study follow-up in low income countries (Peer support vs. none) |
RR = 0.70 RR = 0.63 |
(95% CI: 0.60 - 0.82) (95% CI: 0.52 - 0.78) |
| Khresheh et al. (2011); Jordan | Counseling in facilities by providers, Counseling in homes by providers (via phone) | 2 hospitals | RCT – 90 postpartum women (45 intervention, 45 control) |
% EBF at 6 months |
39% vs. 27% |
NS |
| Kramer et al. (2001); Republic of Belarus | Counseling in facilities by providers, Support group in facilities by providers and/or lay volunteers, Small media | 31 maternity hospitals and polyclinics (16 intervention, 15 control) | RCT – 16,442 mother-infant pairs (8,547 intervention, 7,895 control) |
% breastfeeding at 12 months % EBF at 3 months % EBF at 6 months |
19.7% vs. 11.4%; AOR=0.47 43.3% vs. 6.4%; AOR=0.47 7.9% vs. 0.6% |
(95% CI: 0.32- 0.69) p<0.001 (95% CI: 0.32- 0.69) p= 0.01 |
| Kumar et al. (2008); India | Group education in communities by lay volunteers, Counseling in homes by lay volunteers, Small media, Traditional media | 39 villages (population: 104,123) (13 villages in each groups)(35) | RCT – 2,601 live births (1,522 intervention, 1,079 control) |
% pre-lacteal feed % breastfeeding initiation within one hour of birth |
38·4% vs. 79·9% 70·6% vs.15·5% |
Rate ratio: 0·49 (0·42–0·57) p<0·0001 Rate ratio: 4·57 (3·38–6·15) p<0·0001 |
| Kupratakul et al. (2010); Thailand | Support group in facilities by providers and/or lay volunteers, Counseling in homes by providers, Counseling in homes by providers (via phone) | 5 hospitals | RCT – 80 pregnant women of more than 32 weeks’ gestation (40 intervention; 40 control) |
% EBF at …- 1 month % EBF at …- 2 months % EBF at …- 4 months % EBF at …- 5 months % EBF at …- 6 months % predominant breastfeeding at 6 months % partial breastfeeding at 6 months % EBF at … - 14 days |
77.5% vs. 52.6% 62.5% vs. 36.8% 35.0% vs. 7.9% 25.0% vs. 2.6% 20.0% vs. 0% 40.0% vs. 5.3% 15.0% vs. 15.8% 82.5% vs. 52.6% |
p = 0.021 p = 0.023 p = 0.008 p = 0.012 p = 0.005 p = 0.0002 p = 0.923 p = 0.005 |
| Lassi et al. (2010); various | Events targeting policy makers and regulators, Mobilization of issue groups, Counseling in homes by peers | varied | Literature review – 18 RCTs and quasi-RCTs |
BF initiation within one hour of birth |
RR = 1.94 |
p<0.001 (95% CI: 1.56 - 2.42) |
| Leite et al. (2005); Brazil | Counseling in homes by peers | 8 health centers | RCT – 1,003 women and their newborns (503 intervention, 500 control) |
% EBF at 4 months |
24.7% vs. 19.4% |
p = 0.044 |
| Li et al. (2007); China | Mobilization of issue groups, Mid-sized media, Small media | 10 villages (2 from each of five sampled townships in Luxi city (population of approximately 321,000) | Repeated cross-sectional – 352 infants born during reference period |
Average time until initiation of breastfeeding (hours) - Male child Average time until initiation of breastfeeding (hours) - Female child Average time until initiation of breastfeeding (hours) - Total % breastfeeding initiation within 30 minutes - Male child % breastfeeding initiation within 30 minutes - Female child % breastfeeding initiation within 30 minutes - Total Mean duration of breastfeeding (months) - Male child Mean duration of breastfeeding (months) - Female child Mean duration of breastfeeding (months) - Total |
endline vs. baseline: 17.7 vs. 27.8 endline vs. baseline: 20.3 vs. 29 endline vs. baseline: 19.0 vs. 28.5 endline vs. baseline: 47.4% vs. 37.6% endline vs. baseline: 42.4% vs. 36.8% endline vs. baseline: 44.9% vs. 37.2% endline vs. baseline: 12.0 vs. 12.6 endline vs. baseline: 12.9 vs. 12.9 endline vs. baseline: 12.5 vs. 12.8 |
-- -- -- -- -- -- -- -- -- |
| Mackintosh et al. (2002); Vietnam | Counseling in homes by peers, Group education in communities by lay volunteers | 5 communes (4 intervention, 1 control) | Repeated cross-sectional – 46 households in intervention (family had one child who previously participated in the project and one younger child who had not participated), 25 households in control (had no previous exposure to program activities and included an older and younger sibling in the age range of the intervention households) |
% currently breastfeeding(36) “younger” child Mean duration of breastfeeding (months) |
41.3% vs. 20% 16.8 vs. 16.6 |
NS NS |
| Manandhar et al. (2004); Nepal | Support group in communities by peers, Mobilization of issue groups, Counseling in homes by peers | 42 rural village development committees from one district (21 intervention, 21 control) | RCT – 6,714 (3,190 intervention, 3,524 control) |
% breastfeeding initiation within one hour of birth % discarded colostrum |
62% vs. 54%, AOR37 = 1.40 29% vs. 42%, AOR=0.55 |
(95% CI: 0.52–3.79) (95% CI: 0.27–1.10) |
| Matovu et al. (2008); Uganda | Counseling in facilities by providers, Group education in facilities by providers, Small media, Mid-sized media | 3 health facilities (2 hospitals, 1 center) | Retrospective crosssectional study – 139 HIV infected women with children 6-12 months old who adhered to EBF for six months, 139 HIV infected women who did not adhere to EBF for six months |
EBF for 6 months among HIV+ women (individual vs. group counseling) EBF for 6 months among HIV+ women (at least 4 ANC sessions attended vs. fewer than 4) % EBF for 6 months among HIV+ women (attended at least 6 postnatal counseling sessions vs. fewer than 6) EBF for 6 months among HIV+ women (discussed with health worker decision regarding EBF vs. did not) |
OR=3.43 OR=5.95 OR=3.34 OR= 4.97 |
p<0.05 p<0.05 p<0.05 p<0.05 |
| Morrow and Guerrero (2001); Mexico | Counseling in homes by peers, Small media | 1 community (population of approximately 30,000) | RCT –130 women 3 months postpartum (44 IG1, 52 IG2, 34 control) |
% EBF at 3 months |
IG1 vs. IG2 vs. control group: 67% vs. 52% vs. 12% |
p < 0.001, log rank test |
| Omer et al. (2008); Pakistan | Traditional media, Counseling in homes by peers | 10 communities from 3 districts (5 intervention, 5 control), 8 of which were rural | RCT – 1,070 women who had been pregnant and delivered during the three years prior to the intervention (529 intervention, 541 control) |
% gave colostrum to infant |
79.2% vs. 65.3%; OR=1.60 |
(95% CI: 1.17–2.18) |
| Pachón et al. (2002); Vietnam | Group education in communities by lay volunteers | 12 communes from one rural Vietnam (6 intervention, 6 control) | Longitudinal – 238 children 5-25 months old (119 intervention, 119 control) |
% breastfeeding previous day at month 0 % breastfeeding previous day at month 12 |
72.3% vs. 65.5% 13.4% vs. 13.9% |
NS NS |
| Piwoz et al. (2005); Zimbabwe | Counseling in facilities by providers, Counseling in communities by providers, Group education in facilities, Group education in communities, Mid-sized media | 14 maternity clinics from greater Harare | RCT – 14,110 women from the Zimbabwe Vitamin A for Mothers and Babies Project clinical (ZVITAMBO) clinical trial; 11,362, 1311, and 1437 were enrolled during the pre, partial, and full intervention period; data were collected at multiple time points |
% EBF for at least 3 months % EBF for at least 3 months % EBF for at least 3 months |
Partial participation: AOR=1.75 Full participation: AOR=8.43 No participation vs. group only vs. individual counseling only vs. group + individual 5.5% vs. 13.4% vs. 15.9% vs. 23.7% |
p<0.001 (95% CI: 1.31- 2.35) p<0.0001 (95% CI: 6.13–11.59) p<0.0001 |
| Piwoz et al. (2007); Zimbabwe | Group education in facilities by providers, Small media, Mid-sized media, Counseling in facilities by providers | 14 maternity clinics from greater Harare | RCT – 437 HIV+ postintervention women from the ZVITAMBO clinical trial (20 received 3 visits, 108 received 2 visits, 234 received 1 visit, 75 received no visits); used data collected at 6 weeks postpartum, 3 months postpartum, and every 3 months thereafter for up to 24 months |
% EBF at 6 months % EBF at 6 weeks % EBF at 3 months |
Dose response: 0 vs. 1 vs. 2 vs. 3 visits: 4.0% vs. 6.0% vs. 4.6% vs. 5.0% Dose response: 0 vs. 1 vs. 2 vs. 3 visits: 8% vs. 15% vs. 14.8% vs. 30% Dose response: 0 vs. 1 vs. 2 vs. 3 visits: 6.7% vs. 12.8% vs. 17.6% vs. 25% |
p=0.48 p=0.02 p=0.005 |
| Quinn et al. (2005); Bolivia, Madagascar and Ghana | Individual counseling, Support group, Mass media, Small media, Mobilization | Bolivia - 3 ecoregions in rural and urban areas (1 million people) Ghana – 30 districts (3.5 million people) Madagascar - 10 focus districts of 2 highland provinces (expanded and contracted due to political crises) | Repeated cross-sectional – large repeated surveys in each country |
% EBF of infants < 6 months old - Bolivia % EBF of infants < 6 months old - Madagascar % EBF of infants < 6 months old - Ghana % timely initiation of breastfeeding - Bolivia % timely initiation of breastfeeding - Madagascar % timely initiation of breastfeeding - Ghana |
65% vs. 54% 68% vs. 46% 79% vs. 68% 74% vs. 56% 78% vs. 34% 50% vs. 32% |
p<0.001 p<0.001 p<0.001 p<0.001 p<0.001 p<0.05 |
| Renfrew et al. (2012); various | Individual counseling, Group education/counseling | varied | Literature review – 52 RCTs or quasi-RCTs; 37 in high-income countries, 12 in upper-middle-income countries, 2 in low-middleincome countries, and 2 in low-income countries(38) |
Risk of having stopped EBF at six months (all countries) Risk of having stopped EBF at 4-6 weeks (all countries) Duration of breastfeeding (all countries) Risk of stopping breastfeeding before six months (all countries) |
risk ratios following all forms of extra support together: RR = 0.86 risk ratios following all forms of extra support together: RR = 0.74 risk ratios following all forms of extra support together: positive effect risk ratios following all forms of extra support together: RR = 0.91 |
(95% CI: 0.82 - 0.91) (95% CI: 0.61 - 0.89) -- (95% CI: 0.88 - 0.96) |
| Salud et al. (2009); Philippines | Counseling in homes by peers | 15 zones from one city (population 25,499) | Longitudinal – 312 women with infants less than 2 months old who were not exclusively breastfeeding at baseline |
% EBF of infants < 3 months old |
endline vs. baseline: 53.5% vs. 1% Note: baseline = infant < 2 months, endline = 3 weeks later |
p<0.001 |
| Santiago et al. (2003); Brazil | Counseling in facilities by providers, Counseling in facilities by lay volunteers, Group education in facilities by providers | 1 clinic | RCT – 101 babies (35 in IG1, 33 in IG2, and 33 in control) (39) |
% EBF of infants 4 months old |
IG1 vs. IG2 vs. control: 82.9% vs. 66.7% vs. 30.3%: OR=1.00 vs. OR=0.39 vs. OR=0.11 |
p<0.001: (95% CI: 0.12- 1.36) vs. (95% CI: 0.03- 0.39) |
| Saowakontha et al. (2000); Thailand | Mobilization of campaign, event, or special 'days', Mid-sized media, Small media | 90 villages (target population was approximately 33,000 women of reproductive age) | Repeated cross-sectional – 700 |
% gave colostrum to infant |
endline vs. baseline: 97.2% vs. 65.3% |
p=0.000 |
| Sule et al. (2009); Nigeria | Small media, Group education in communities by providers, Counseling in homes by providers | 2 rural communities (1 intervention with a population of 5,202, 1 control with a population of 8,121) | Longitudinal – 300 mothers of children aged 0-18 months (150 intervention, 150 control) |
% EBF for 6 months among infants were ≥ 6 months % EBF for 6 months among infants were ≥ 6 months |
baseline:59.0% vs. 47.9% endline (6 months later): 66.7% vs. 52% |
NS p=0.01 |
| Sun et al. (2011); China | Small media, Mass media, Mid-sized media, Counseling in facilities by providers | 2 counties in Shan’xi province (approximately 6,000 children 6-24 months of age) | Repeated cross-sectional – baseline (June 2008): 226 caregivers and their children 6-24 month-old; endline (January 2010): 221 caregivers and their children 6-24 month-old |
% breastfeeding at 24 months % early initiation of breastfeeding % breastfeeding at 12 months |
endline vs. baseline: 37.0% vs. 42.3% endline vs. baseline: 16.8% vs. 8.6% endline vs. baseline: 66.7% vs. 76% |
NS p=0.01 NS |
| Syed et al. (2006); Bangladesh | Counseling in facilities by providers, Social and community mobilization/ engagement/ empowerment, Counseling in homes by providers | 11 upazilas (subdistricts) (10 intervention, 1 control) | Repeated cross-sectional – baseline: 3,325 women with children aged less than one year (2,989 intervention, 336 control); endline: 3,110 women with children aged less than one year (2,782 intervention, 323 control) |
% breastfeeding initiation within one hour of birth % breastfeeding initiation within one hour of birth % gave colostrum to infant % gave colostrum to infant % gave colostrum to infant % breastfeeding initiation within one hour of birth |
endline: 76.2% vs. 46.7% Difference in differences: 23.6 baseline: 86.3% vs. 91.7% endline: 96.5% vs. 93.5% Difference in differences: 8.4 baseline: 38.6% vs. 32.7% |
-- p=0.01 -- -- -- -- |
| Thompson and Harutyunyan (2009); Armenia | Counseling in homes by peers, Support group in communities by peers, Small media, Mid-sized media | 16 villages across Martuni region of Gegharkunik marz | Repeated cross-sectional – 600 with at least one child aged 24 months or younger (300 preintervention; 300 postintervention) |
% breastfeeding initiation within one hour of birth % breastfeeding initiation within one hour of birth % EBF in the past 24 hours among infants < 6 months old % EBF in the past 24 hours among infants < 6 months old |
endline vs. baseline: endline vs. baseline: 66.9% vs. 70.4% endline vs. baseline: exposed vs. unexposed: 72.5% vs. 56.3% endline vs. baseline: endline vs. baseline: 48.1% vs. 16.7% endline vs. baseline: exposed vs. unexposed: 58.1% 37.5% |
p<0.05 p<0.05 p<0.05 p<0.05 |
| Tripathy et al. (2010); India | Support group in communities by peers | 36 clusters (18 intervention, 18 control) with a population of approximately 228,186 | RCT – live births identified over study period; analysis of breastfeeding indicators was limited to 16,926 infants alive at 1 month (8,807 intervention; 8,119 control) |
% breastfeeding initiation within four hours of birth % EBF for 6 weeks |
61% vs. 61%; OR= 1.01(40) 80% vs. 69%; OR=1·82 |
(95% CI: 0·38– 3·11) (95% CI: 1.14–2.92) |
| Tylleskar et al. (2011); Burkina Faso, Uganda, South Africa | Counseling in homes by peers | 24 communities in Burkina Faso, 24 in Uganda, and 34 in South Africa (villages randomly assigned in a 1:1 ratio to control or intervention clusters) | RCT – 2,579 motherinfant pairs Burkina Faso: 392 intervention, 402 control Uganda:396 intervention, 369 control South Africa: 535 intervention, 485 control |
% EBF during previous 7 days at 6 months (24 weeks) - Burkina Faso % EBF during previous 7 days at 6 months (24 weeks) - Uganda % EBF during previous 7 days at 6 months (24 weeks) - South Africa % EBF during previous 7 days at 3 months (12 weeks) - Burkina Faso % EBF during previous 7 days at 3 months (12 weeks) - Uganda % EBF during previous 7 days at 3 months (12 weeks) - South Africa |
71% vs. 9%; OR=3.27 51% vs. 11%; OR=2.30 2% vs. <1%; OR=1.98 77% vs. 23% 77% vs. 34% 8% vs. 4% |
(95% CI: 2.13–5.03) (95% CI: 2.00–2.65) (95% CI: 1.30- 3.02) -- -- -- |
Overview of the Evidence, by Practice
This chapter is organized around the three WHO recommended breastfeeding practices: early initiation of breastfeeding, EBF through six months of age, and continued breastfeeding until 24 months of age. The feeding of colostrum is included in the early initiation section. Specific indicators used to measure these practices varied considerably among studies.
Of the studies reviewed, 22 reported findings related to early initiation of breastfeeding (including six which reported on feeding colostrum), 41 reported on findings related to EBF, and 16 reported on findings related to continued breastfeeding. See Table 3.2.1.
Early initiation of breastfeeding
Of the 22 studies that measured breastfeeding initiation, 17 reported statistically significant results. Fifteen of those looked at early initiation of breastfeeding (defined as initiating breastfeeding within one hour of birth) and three reported statistically significant results related to the feeding of colostrum to infants.
Literature reviews with meta-analysis
Gogia and Sachdev (2010) conducted a literature review and meta-analysis to “determine whether home visits for neonatal care by community health workers (CHW) can reduce infant and neonatal deaths and stillbirths.” The review included four studies that assessed breastfeeding counseling provided during home visits.14 All of these studies have been included and are described in the present review. The authors found that women in the intervention group were more than three times more likely to initiate breastfeeding compared with women in control groups (pooled RR 3.35).
Lassi et al. (2010) conducted a review of 18 cluster-RCTs/ quasi-RCTs of community-based interventions that involved training outreach workers in maternal care during pregnancy, delivery and in the postpartum period as well as newborn care. Six of the 18 studies assessed the impact of these interventions on initiation of breastfeeding within one hour after birth. The meta-analysis from these studies revealed that following community-based interventions, the practice nearly doubled (Relative Risk (RR) = 1.94). All but one of the six studies included in the meta-analysis matched the inclusion criteria for the present review. These studies are included in the summaries that follow.15
Table 3.2.1: Number of studies reviewed (with number of studies reporting statistically significant results in parentheses),
by recommended practice and study design
| Practice | Literature reviews with meta-analysis | RCT | Longitudinal studies |
Repeated crosssectional studies |
Crosssectional studies |
Total |
| Early initiation of breastfeeding | 2 (2) | 8 (6) | 2 (1) | 10 (8) | 0 | 22 (17) |
| Exclusive breastfeeding | 4 (4) | 20 (14) | 12 (8) | 7 (5) | 2 (2) | 41 (32) |
| Continued breastfeeding | 2 (1) | 7 (6) | 3 (1) | 4 (1) | 0 | 16 (9) |
* Note that columns do not sum to the “total” since some articles reported on multiple practices.
Randomized controlled trials
Six studies reporting statistically significant changes in early initiation of breastfeeding were RCTs. Two of these were included in the meta-analyses described above.16 Bang et al. (2005) looked at early initiation of breastfeeding as an outcome associated with delivery of a home-based neonatal care package in Gadchiroli District of India. The intervention included the training of female village health workers (VHWs) and traditional birth attendants (TBAs) to provide health education to mothers and grandmothers about care of pregnant women and of neonates through advice, demonstration, and assistance. Messaging addressed traditional beliefs and practices as well as barriers to care identified during the first two years of the program, and was delivered during group meetings conducted once every four months, during home visits in the eighth and ninth months of pregnancy, and on the first day after delivery. By the end of the intervention period, the percentage of women initiating breastfeeding within six hours of birth had increased from 47.5% to 89.7%. The authors did not report on the standard indicator of initiation of breastfeeding within one hour of birth.
Baqui et al. (2008a) conducted a large-scale RCT in Sylhet District, Bangladesh. Twenty-four clusters, each with a population of approximately 20,000 people, were randomly assigned to one of three study groups – a community-care (CC) group, a home-care (HC) group, or a control group. In both intervention groups (CC and HC), “male and female community mobilisers were recruited to hold group meetings for the dissemination of birth and newborn-care preparedness messages…” Each female community mobilizer held community meetings once every four months, while each male community mobilizer did so every 10 months. In the CC arm, female volunteers identified pregnant women and encouraged them to attend community meetings organized by the community mobilizers, receive routine antenatal care, and seek care for signs of serious illness in mothers or newborns. In the HC arm, CHWs “promoted birth and newborn care preparedness through two scheduled antenatal and three early postnatal home visits.” Furthermore, female community mobilizers conducted group meetings with women in the HC group once every eight months. After nearly three years, breastfeeding initiation within one hour of birth was dramatically higher in both intervention arms compared with the control group. While a total of 71% of women in the CC group and 81% in the HC group reported initiating breastfeeding within one hour of birth, only 57% in the control group reported the same. Differences between both intervention groups and the control group were statistically significant.
Guldan et al. (2000) looked at Infant and Young Child Feeding (IYCF) practices in rural Sichuan, China. The study included two intervention and two control groups. The intervention included counseling by CHWs during monthly growth monitoring and promotion sessions, and visits to the homes of pregnant women and women with children under the age of one. At the conclusion of the one year intervention, the percentage of women reporting feeding newborns colostrum was higher in the intervention group than in the control group (91% vs. 80%). However, the authors explained that complete randomization in group selection was not possible as the “[intervention] and control group townships could not be contiguous and needed to be roughly equal geographically and socioeconomically” and there were important differences between the women in the intervention and control groups in terms of mean years of education and main income source, making it difficult to draw conclusions about the effect of the intervention or the generalizability of the findings.
Haider et al (2000) looked at early initiation and EBF in Dhaka, Bangladesh. The intervention included home-based peer counseling over a period of five months. Trained counselors conducted a total of 15 home visits in the five month study period: two in the last trimester of pregnancy, four in the first month postpartum, and every fortnight afterward until the completion of five months. At the conclusion of the study period, the percentage of women initiating breastfeeding within the first hour after birth was 64% in the intervention group, compared with 15% in the control group.
Kumar et al. (2008) conducted a cluster-RCT looking at early initiation of breastfeeding in the context of a neonatal mortality intervention in Uttar Pradesh, India. The intervention included training CHWs to hold community meetings with folk songs and conduct home visits targeting community leaders, priests, teachers, birth attendants, unqualified medical care providers, healthcare workers, fathers-in-law, husbands, mothers-in-law, pregnant women or mothers, neighbors, and relatives. After 16 months of implementation, the rate of breastfeeding within one hour of birth was 70.6% among women in intervention areas compared with 15.5% among women in non-intervention areas.
Finally, Omer et al. (2008) conducted an RCT focused on reducing heavy workload (such as lifting) during pregnancy, attending prenatal check-ups, and feeding colostrum to newborns in Sindh Province of Pakistan. The intervention trained lady health workers (LHW) to present and discuss an embroidered cloth panel portraying key messages during routine home visits with pregnant women. Following the interventions, the authors found that women advised by a LHW were more likely to feed their newborn colostrum than those who were not advised by a LHW to do so (79.2% vs. 65.3%).
Longitudinal studies
Akter et al. (2012) looked at breastfeeding initiation in a longitudinal study conducted in one hospital in Bangladesh. The intervention involved four group counseling sessions during the last trimester of pregnancy. Following the intervention, 75.4% of women in the intervention group reported initiating breastfeeding within one hour of birth, compared with 34.4% of women in the control group.
Repeated cross-sectional studies
Eight of the 15 studies reporting statistically significant results were repeated cross-sectional studies. Baqui et al. (2008b) looked at the impact of an intervention that included counseling during home visits on preventive care practices for mothers, newborn care practices, and the use of health care services in India. In intervention areas, CHWs recruited additional community volunteers (“change agents”) to expand the reach of the program. Following intervention, the number of women initiating breastfeeding within one hour of birth increased by 31.2%.
Crookston et al. (2007) looked at breastfeeding behaviors in rural Cambodia. The intervention “assess[ed] the impact of Buddhist nuns and wat (pagoda) grannies on optimal breastfeeding behaviours.” Young nuns were trained to promote breastfeeding in their villages using leaflets, visual aids and home visits. In addition, the nuns mobilized community groups, provided one-on-one counseling, and conducted educational sessions with groups of 6-10 women. After one year, women in the intervention group were 62% more likely to initiate breastfeeding within one hour after birth than those in the control group.
Quinn et al. (2005) conducted a repeated cross-sectional study looking at timely initiation and EBF in three countries: Madagascar, Bolivia, and Ghana. The interventions varied slightly by country, but all were implemented at scale as part of the USAID-funded Linkages project. In all three countries, “women were reached through small- and large-group activities, one-on-one counseling in homes and at local health posts, breastfeeding promotion songs performed by women’s groups and musical troupes, and community mobilization events such as local theater, health fairs, and festivals celebrating breastfeeding and child health days.” Mass media was used to complement this community-based work. After four years of implementation, rates of early initiation of breastfeeding increased from 34% to 78% in Madagascar, from 56% to 74% in Bolivia, and from 32% to 50% in Ghana. All changes were statistically significant.
Guyon et al. (2009) reported on the same intervention in Madagascar after five years of implementation, finding an increase from 33% to 68% in rates of early initiation of breastfeeding. Neither Quinn et al. (2005) nor Guyon et al. (2009) included a control group.
Saowakontha et al. (2000) looked at the feeding of colostrum to newborns in three districts in northeastern Thailand. The intervention included training health officials, health care workers, and villagers, and disseminating messaging through broadcasting systems, events and exhibitions, and bulletin boards. By the end of the study period, the percentage of women reporting feeding colostrum to their newborns increased from 63.5% to 97.2%. There was no control group in this study.
Sun et al. (2011) looked at infant and young child feeding practices as part of an intervention to promote consumption of Ying Yang Bao (YYB), a multiple micronutrient powder, in Shan’xi Province, China. The intervention included formative research and message testing as well as behavior change communication (BCC) targeting health workers and family members. BCC materials included handbooks about IYCF for parents, booklets about YYB for health workers, and television spots for the public. After 20 months of project implementation, the percentage of women reporting early initiation of breastfeeding had increased from 8.6% to 16.8%.17 There was no control group in this study.
Syed et al (2006) looked at the provision of newborn care in 10 upazilas (sub-districts) in Bangladesh. The intervention involved “increasing the coverage of health workers and community-based caregivers trained and competent in providing essential newborn care and promoting positive maternal and newborn-care practices.” Activities targeted pregnant mothers as well as family decision-makers, such as husbands, mothers-in-law, and village leaders. After less than two years of the intervention, the percentage of women initiating breastfeeding within one hour of birth had increased from 38.6% to 76.2%. Though this difference was statistically significant, the lack of a control group makes it difficult to conclusively attribute the change to the intervention.
Finally, Thompson and Harutyunyan (2009) looked at EBF through six months of age in Martuni region in northeastern Armenia. The intervention was delivered through a community- based Integrated Management of Childhood Illness (IMCI) platform, and included home visits for pregnant and lactating women, group education classes for new parents, and a local mass media campaign. By the end of the study period, the number of women initiating breastfeeding within one hour of birth was 16.2% greater among women exposed to the IMCI campaign compared with those who were not exposed (72.5%. and 56.3%, respectively).
Exclusive breastfeeding
Of the 45 studies reporting on EBF practices, 33 reported statistically significant results. Five of these studies reported on duration of EBF, 14 reported on EBF for six months, five reported on EBF among those under six months of age, 19 reported on EBF at other ages, and two reported on EBF at unspecified ages.
Literature reviews with meta-analysis
Hall (2011) conducted a meta-analysis of four RCTs to “assess the effectiveness of community-based interventions to improve the rates of exclusive breastfeeding at four to six months in infants in low- and low-to-middle income countries.” Community-based interventions were defined as interventions accessible locally to the woman (whether in her own home or a local building such as a clinic or school); delivered by a health professional or trained lay person; and provided either individually or in a group. Interventions could be provided antenatally, postnatally or both, and participants were limited to women who were pregnant or currently breastfeeding an infant less than six months of age. The meta-analysis found a significant effect of community-based interventions on rates of EBF (pooled OR=5.9). Three of the four studies reviewed by Hall met the inclusion criteria of the present review and these studies are summarized below.18
Imdad et al. (2011) conducted a meta-analysis of a total of 53 RCTs and quasi-RCTs looking at the impact of education and support strategies on breastfeeding outcomes. Fourteen of those studies were conducted in developing countries. The interventions were defined as “breastfeeding education and/or additional support given to mothers through counselors (be they doctors, nurses, midwives, lactation consultants or peer counselors) in individual or group sessions.” The authors found a six-fold increase in rates of EBF at six months of age and an 89% increase in rates of EBF at 4-6 weeks of age following interventions in developing countries. In sub-group analyses, the authors found that changes in prenatal counseling, postnatal counseling and a combination of both were statistically significant, with “the highest impact being that of prenatal counselling.” Furthermore, “group counselling had a greater impact (67% increase in EBF rate at 4-6 weeks), compared with individual counselling (38% increase). The results were statistically significant at all levels of care (community, facility and both combined).” Five of the fourteen studies included in the Imdad et al. (2011) review met the inclusion criteria of the present review and these studies are included in the summaries below.19
Jolly et al. (2012) conducted a systematic literature review and meta-regression analysis of RCTs conducted in high-, middle- and low-income countries “to examine the effect of setting, intensity, and timing of peer support on breastfeeding practices.” The review included 17 RCTs; however, only six were conducted in low income countries. In a sub-analysis of low income countries, those who received peer support were significantly less likely to have stopped EBF at the last study follow up (RR=0.63). Four of the six low-income country studies included in this meta-analysis met the inclusion criteria for the present review and are included in the summaries below.20
Finally, Renfrew et al. (2012) conducted a Cochrane Review looking at the effect of extra support for women on EBF at six months of age. Extra support was defined as reassurance, praise, information, and the opportunity to discuss questions with professionals, trained lay people or both during home visits or facility-based counseling sessions. This meta-analysis used data from 52 RCTs or quasi-RCTs, 16 of which were conducted in middle- to low-income countries. Extra support led to a lower risk of having stopped EBF at 4-6 weeks (average RR=0.74) as well as a lower risk of having stopped EBF at six months (average RR=0.86). In a sub-group analysis, the authors found a greater treatment effect on EBF cessation at six months “in settings where there were high background rates of breastfeeding initiation (average RR=0.83) compared with areas where there was intermediate (average RR=0.89) or low background initiation rates (average RR = 1.00), […] Results were even more pronounced for cessation of exclusive breastfeeding at up to four to six weeks with interventions seeming to be most effective for women living in areas with high background initiation rates (average RR=0.61) compared with areas with intermediate (average RR=0.81) or low rates (average RR=0.97)” (Renfrew et al., 2012). Fourteen of the 16 middle- to low-income studies included in Renfrew et al. (2012) met the inclusion criteria for the present review and these studies are included in the summaries below.21
Randomized controlled trials
Fifteen of the 33 studies reporting statistically significant results were RCTs. Thirteen of these were included in the meta-analyses described previously.22 Agrasada et al. (2005) looked at the efficacy of postnatal peer counseling conducted in a hospital in Manila, Philippines. The study included two intervention groups – one in which peer counselors were trained to provide breastfeeding counseling and the other in which peer counselors were trained to provide general childcare counseling. Women in the control group did not receive any counseling. Both intervention groups included eight counseling sessions conducted at home at the following child ages: 3-5 days, 7-10 days, 21 days, and 1.5 months. Women then received counseling monthly until their children reached the age of 5.5 months. Data on breastfeeding practices were collected during well-child hospital visits at child age two and four weeks and monthly until the child was six months old. The authors found that women in the breastfeeding counseling intervention group were more likely to report EBF in the previous seven days at six months of age (44%) than those in the childcare counseling group (7%) or the control group (0%). Statistical modeling revealed that women in the breastfeeding counseling group were 6.3 times more likely to EBF in the previous seven days at six months than those in the other groups.
Aidam et al. (2005) conducted an RCT looking at EBF at six months of age in Thema Township, Ghana. The intervention involved provider counseling and home visits, and included two intervention groups: the first intervention group (IG1) included two prenatal counseling sessions, one counseling session 48 hours postpartum, and six postpartum home visits. The second group (IG2) excluded the prenatal counseling, but included the rest. Following the intervention, the rate of EBF at 6 months of age was 90% in IG1, 79.5% in IG2, and 54.5% in the control group.23 The differences were statistically significant for both intervention groups.
Aksu et al. (2011) looked at breastfeeding practices among women who gave birth in a hospital following a Baby-friendly Hospital Initiative (BFHI) intervention in Aydın, Turkey. In addition to education in the first few hours after delivery, which was provided to all women in the hospital, women in the intervention group received breastfeeding education at home three days after delivery and were then monitored for six months. At each time point, women in the intervention group were more likely to exclusively breastfeed than women in the control group – 67% vs. 40% at two weeks, 60% vs. 33% at six weeks, and 43% vs. 23% at six months. At six months postpartum, women in the intervention group exclusively breastfed an average of 0.8 months longer more than women in the control group (4.7 vs. 3.9 months).
Bashour et al. (2008) explored the impact of home visits on EBF in Damascus, Syria. Women who had recently given birth at the Maternity Teaching Hospital in Damascus, Syria were randomly assigned to one of two intervention groups – women in IG1 received four home visits (on days 1, 3, 7, and 30 following delivery), while women in IG2 received one home visit on the third day after delivery. Registered midwives were trained to provide information, education, and support to women during the home visits. Following the intervention, EBF at four months was significantly higher among women in both intervention groups – 28.5% among women in IG1 and 30.1% among those in IG2 compared with those who received no home visits (20.2%). Following a sub-analysis of findings, the authors found that “the effect was major among the groups of women who had normal vaginal delivery but not the women who had a c-section or were primigravidae,” and suggested that other factors such as hospital policy regarding the practice of rooming-in (the practice of mothers and babies staying together after birth) may play a role and that such subgroups may “require a special package of intervention.”
Bhandari et al. (2003) looked at median duration of EBF, and EBF at three, four, five, and six months of age in Haryana, India. The intervention included counseling by traditional birth attendants at birth, monthly home visits by community-based health workers during the child’s first year of life, and counseling during weighing sessions every three months. By the end of the study period, median duration of EBF was significantly higher in the intervention group compared with the control group (122 days vs. 41 days), the proportion of women reporting EBF at four months was 69% in the intervention group compared with 12% in the control group, and EBF at six months was 42% in the intervention group compared with 4% in the control group. In a follow-up article, Bhandari et al. (2005) reported that women exposed to three or more communication channels were more likely to be practicing EBF at three months than those exposed to just one or two communication channels or none at all (≥3 channels: 93.6%, 1-2 channels: 81.3%, no channels: 70.8%).
Bortolini and Vitolo (2012) looked at EBF practices among infants born in one hospital in Brazil. The intervention group received home visits during the children’s first year of life on a monthly basis up to 6 months, and at 8, 10 and 12 months. Results showed that children in the intervention group were more likely to be EBF for four or more months than those in the control group (45.1% vs. 28.6%) and at six months (19.1% vs. 8.2%).
Coutinho et al. (2005) looked at EBF at six months of age in Pernambuco, Brazil. Following the training for health care providers on the BFHI,24 the proportion of infants exclusively breastfed during the hospital stay increased significantly, but EBF was not sustained beyond the hospital stay. In an effort to increase the length of EBF, researchers randomly assigned women who gave birth in two BFHI hospitals to the intervention group, which received ten postnatal home visits by trained community health agents, or to the control group, which did not receive home visits but still received the breastfeeding support provided during the hospital stay. At six months post-partum, rates of EBF among women in the intervention group was 45% compared with 13% among women in the control group.
Haider et al (2000) looked at breastfeeding practices in Dhaka, Bangladesh. Following a five month peer counseling intervention, the proportion of children exclusively breastfed at five months of age was 70% in the intervention group compared with 6% in the control group.
Kramer et al. (2001) looked at breastfeeding practices in 31 hospitals in Belarus. The objective of the Promotion of Breastfeeding Intervention Trial (PROBIT) was to “assess the effects of breastfeeding promotion on breastfeeding duration and exclusivity and gastrointestinal and respiratory infection and atopic eczema among infants.” The intervention was “modeled on the Baby-Friendly Hospital Initiative of the World Health Organization and United Nations Children’s Fund, which emphasizes health care worker assistance with initiating and maintaining breastfeeding and lactation and postnatal breastfeeding support, or a control intervention of continuing usual infant feeding practices and policies.” At the end of the study period, rates of EBF at three months were 43.3% in the intervention group and 6.4% in the control group. Rates of EBF at six months were 7.9% in the intervention group and 0.6% in the control group.
Kupratakul et al. (2010) looked at EBF in the context of an intervention aimed at improving antenatal and postnatal support in Bangkok, Thailand. The intervention included Knowledge Sharing Practices with Empowerment Strategies (KSPES) during antenatal education (storytelling, demonstration, and practice) as well as postnatal support. At the end of the six month intervention period, 20% of the women in the intervention group reported EBF through six months, while none in the control group reported EBF through six months.
Leite et al. (2005) assessed “the effectiveness of homebased peer counselling on breastfeeding rates for unfavourably low birthweight babies” in Brazil. Women and their newborn babies identified in the maternity services that met the inclusion criteria were randomly assigned to the intervention or control group. The intervention group received six home visits from trained lay counselors at 5, 15, 30, 60, 90 and 120 days following birth. In addition to counseling, home visits included “interviews with the mother; observation of the home environment; observation of all aspects involved with the breastfeeding, including technical ones, as well as the mother-child relationship; and identification of the difficulties faced by the mother during breastfeeding.” Following the intervention, rates of EBF at four months were significantly higher among the intervention group than the control group (24.7% vs. 19.4%).
Morrow and Guerrero (2001) conducted an RCT to determine the effectiveness of three vs. six home visits by peer counselors on EBF rates among women in a peri-urban area of Mexico. The intervention was informed by an ethnographic study identifying maternal beliefs, practices, and needs. Lay counselors were recruited from each intervention community and trained by La Leche League. Women were enrolled during pregnancy and randomly assigned to one of two intervention groups –women in IG1 received six home visits and women in IG2 received three home visits. In IG1, “mothers were visited by [peer counselors] twice during pregnancy (mid and late pregnancy), immediately postpartum, and at weeks 2, 4, and 6 postpartum.” In IG2, women were visited once late in pregnancy, once immediately postpartum, and once two weeks after delivery. The authors found that women in IG1 who received six home visits had higher rates of EBF than women in IG2 who received three visits (67% vs. 52%) or no visits (12%).
Santiago et al. (2003) explored the role of pediatricians in improving EBF rates at four months in Brazil. Healthy, full term infants born at the pediatrics outpatient clinic in the Triângulo Mineiro teaching hospital in Minas Gerais were randomly divided into three groups. Women in IG1 received advice from a multidisciplinary breastfeeding support team (including a pediatrician trained in breastfeeding, a social worker, a psychologist, a dentist and a nurse) during which time breastfeeding difficulties were discussed and potential solutions were offered as a group. Women in IG2 received advice from the same trained pediatrician, but in individual consultations, while those in the control group received advice from a pediatrician with no breastfeeding training. Following the intervention, 82.9% of women in IG1, 66.7% of women in IG2, and 30.3% of women in the control group reported EBF at four months. The difference between IG1 and IG2 was not statistically significant; however, the difference between IG1 and the control was statistically significant.
Finally, Tripathy et al. (2010) conducted an RCT looking at EBF in Orissa and Jharkhand, India. In intervention areas, female facilitators conducted monthly meetings with new mothers for a period of 20 months. The meetings revolved around participatory learning and action, and led to the development and implementation of strategies to address maternal and newborn health problems. Following adjusted analysis of surveillance data from years one and three, the authors reported that the proportion of infants exclusively breastfed at six weeks was higher in intervention areas than in control areas (OR=1.82).25
Longitudinal studies
Eight of the 33 studies reporting statistically significant results were longitudinal studies. Akter et al. (2012) studied an intervention which involved four group counseling sessions during the last trimester of pregnancy. Following the intervention, 64.9% of women in the intervention group reported EBF at one month after birth, compared with 37.9% of women in the control group.
Balaluka et al. (2012) conducted a longitudinal study looking at median duration of EBF, and EBF at four, five, and six months in two health districts in the Democratic Republic of Congo. The intervention involved a combination of home visits, community based counseling, and community mobilization, including monthly community “weighing sessions.” After approximately three years of implementation, the authors compared EBF practices in the intervention health district with another health district that resembled the intervention district. They found that median duration of EBF was two months longer in the intervention district than in the control district (six vs. four months). Likewise, the prevalence of EBF was higher in the intervention district compared with the control district at all of the following points in time: at four months (91.8% vs. 50.7%), at five months (81.3% vs. 9.6%), and at six months (57.7% vs. 2.7%). All differences were statistically significant; however, the lack of baseline values makes it difficult to attribute these differences to the intervention.
Braun et al. (2003) looked at the effect of the BFHI on EBF by following two cohorts of babies born at a hospital in Porto Alegre, Brazil – one cohort of babies born before the BFHI initiative had been introduced and the other born after the BFHI initiative had been introduced. Mothers were interviewed at the end of the first, second, fourth, and sixth months after birth. The authors found that the median duration of EBF increased from one month to two months after BFHI had been introduced.
Alam et al. (2002) conducted a longitudinal study comparing EBF rates in two hospitals in Dhaka, Bangladesh. The intervention hospital was certified “baby-friendly,” while the control hospital was not. The authors assessed EBF practices at 30, 60, 90, 120, and 150 days postpartum. At each time point, EBF rates were higher among women who delivered in the baby-friendly hospital compared with those who delivered in the control hospital, but the statistical significance of these differences was not reported. The likelihood of EBF for more than five months was statistically greater in the intervention hospital compared with the control hospital (8.1% vs. 6.5%) as was the median duration of EBF (69.74 days vs. 48.4 days).
Piwoz et al. (2005) conducted a longitudinal study within the Zimbabwe Vitamin A for Mother and Babies Trail (ZVITAMBO), in which 14,110 mother-baby pairs were enrolled within 96 hours of delivery. “Formative research was undertaken to guide the design of the program that included group education, individual counselling, videos, and brochures. Exclusive breast-feeding was recommended for mothers of unknown or negative HIV status, and for HIV+ mothers who chose to breast-feed.” Comparisons were made between women who were enrolled in the trial before, during, and after the education and counseling program was implemented. The authors found that women who enrolled in ZVITAMBO while the counseling and education program was being fully implemented were 8.4 times more likely to EBF at three months compared with those who enrolled in ZVITAMBO before the program began and, therefore, only partially participated in the program.
In a follow-up article, Piwoz et al. (2007) assessed the association between exposure to the counseling and education program described directly above and postnatal HIV transmission among a subset of 437 HIV+ women, 365 of whom did not know their HIV status. They found that EBF was higher in the intervention group at six weeks and three months, but no such effect was found at six months.
Salud et al. (2009) evaluated a peer counseling intervention aimed at increasing EBF rates in the Philippines. The intervention targeted 312 “mothers with infants less than 2 months of age who were not exclusively breastfeeding or had difficulty breastfeeding.” Each woman then received three peer counseling visits: one at baseline, another after one week, and the third after two weeks. Following intervention, EBF had increased from 1% to 53.5%.
Finally, Sule et al. (2009) evaluated the impact of nutritional education on IYCF knowledge, attitude and practices (KAP) of women in southwest Nigeria. The intervention included group-based nutrition education and demonstrations of complementary food once every two weeks for a period of six months. Participating women were also visited in their homes monthly for six months to monitor their EBF practices. After six months, women with children six months or older in the intervention community were significantly more likely to report EBF for six months than women in the control community (66.7% and 52.0%, respectively).
Repeated cross-sectional studies
Crookston et al. (2007) looked at breastfeeding behaviors in a repeated cross-sectional study conducted in rural Cambodia. At baseline, women in program communities were 54% more likely than women in control communities to have breastfed exclusively in the previous 24 hours. At followup, they were 81% more likely to do so. After one year, the RR associated with EBF through six months of age had increased from 1.54 to 1.81.
Harkins et al. (2008) evaluated the effectiveness of IMCI interventions on EBF for six months in San Luis, Honduras. The authors described the community IMCI approach as using the “social-actor methodology” to create and promote “linkages between partners that create synergies in local and regional social mobilizations.” Teachers and Red Cross volunteers trained in community IMCI conducted home visits to promote prenatal care, exclusive breastfeeding, vaccinations, and the healthy growth and development of infants and children.” After one year, the percentage of women reporting EBF for six months increased from 11% to 19%.
Thompson and Harutyunyan (2009) looked at EBF through six months of age in Martuni region in northeastern Armenia. The intervention was delivered through a community-based IMCI platform, and included home visits for pregnant and lactating women, group education classes for new parents, and a local mass media campaign. By the end of the study period, rates of EBF through six months increased from 16.7% to 48.1%. The finding was statistically significant; however, the lack of data from a control group limits claims of attribution.
Quinn et al. (2005) looked at EBF in three countries: Madagascar, Bolivia, and Ghana. Following a four year multi-channel behavior change program, rates of exclusive breastfeeding among children under six months of age increased from 46% to 68% in Madagascar, from 54% to 65% in Bolivia, and from 68% to 79% in Ghana.
In a follow-up analysis to the Quinn et al. (2005) article, Guyon et al. (2009) reported an increase in EBF from 42% to 70% after five years of implementation in Madagascar. Neither Quinn et al. (2005) or Guyon et al. (2009) included control communities.
Cross-sectional studies
Gupta et al. (2004) “evaluated the extent to which exposure to BCC messages in the media determined improvements in exclusive breastfeeding knowledge and practices in areas targeted by the Delivery of Improved Services for Health (DISH) Project of Uganda.” The DISH Project developed and disseminated materials that targeted both men and women of reproductive age and promoted EBF for the first six months and appropriate complementary feeding practices. The intervention used radio, television, video, posters, and print materials (newspapers, magazines, or leaflets). Multiple logistic regression analysis showed that women exposed to multiple BCC messages were more likely to exclusively breastfeed their infant for six months than women exposed to just one message. However, women with no exposure were more likely to EBF than those with exposure to any one message. The authors cautioned that “even the positive effects of self-reported BCC exposure on the outcomes of interest do not necessarily imply a direct causation, since precise information on the timing of changes in knowledge and practices of individuals with respect to exposure to the mass media was lacking.” They also discussed confounding factors which may be related both to exposure and to EBF practices, including age of respondent, marital status, parity and other socio-demographic factors, and suggested that the effects of media on breastfeeding practices “were less conclusive possibly because of the short interval between the launch of the BCC campaign and survey implementation.”
Matovu et al. (2008) conducted a retrospective study comparing adherence to EBF recommendations among HIV+ Ugandan women attending individual client-provider counseling at a health facility and receiving education with those who only received group counseling or education. At enrollment all women opted to EBF. At the conclusion of the study period, women who had received individual counseling were more likely to EBF at six months than those who had received only group counseling or education (OR=3.43). Furthermore, those who had attended at least four antenatal care (ANC) visits were even more likely to EBF at 6 months (OR=5.95, Adjusted Odds Ratio (AOR)=3.86), while those who attended at least six postnatal counseling sessions were more likely to EBF at six months than those who received less (OR=3.34,AOR=12.52). Finally, those who consulted a health worker regarding breastfeeding problems were also more likely to EBF at six months (OR=4.97, AOR=13.11). However, the retrospective nature and the lack of randomization make it impossible to attribute these differences to the intervention.
Continued breastfeeding
Of the 16 studies reporting on continued breastfeeding practices, nine reported statistically significant results. Continued breastfeeding was measured at six months (two studies), at 12 months (four studies), and at 24 months (one study). Statistically significant results with regard to the duration of any breastfeeding were reported in two studies.
Literature reviews with meta-analysis
Imdad et al. (2011) conducted a meta-analysis of 53 RCTs and quasi-RCTs looking at the impact of education and support strategies on breastfeeding outcomes. The authors reviewed twenty studies which reported on breastfeeding at six months, two of which were from developing countries. They found a 12% increase in any breastfeeding at six months (RR =1.12).
Randomized controlled trials
Six of the nine studies reporting statistically significant results were RCTs. Aksu et al. (2011) looked at the impact of both facility and home-based counseling on continued breastfeeding rates in Turkey. Following the intervention, women in the intervention group breastfed for an average of three months longer than those in the control group (15.1 vs. 12.1 months).
Bhandari et al. (2001) conducted an RCT in which women were randomized to one of four groups: the first received a micronutrient-fortified food supplement along with monthly nutrition counseling, the second received monthly nutrition counseling alone, the third received a twiceweekly home visit, and the fourth received no intervention. Following intervention, breastfeeding at 12 months was actually lower among those who received the supplement and counseling compared with those who received no intervention (83.9% vs. 96.7%). The authors reported no other statistically significant findings.
Bortolini and Vitolo (2012) conducted a study of infants born in Rio Grande do Sul, Brazil. Newborns and their mothers were randomly assigned to the intervention or control group. Women in the intervention group received home visits during the children’s first year of life once a month through the age of six months, and then every other month until the child reached 12 months. Following the intervention, the proportion of children breastfed at both six and 12 months was higher in the intervention group than the control group (66.3% vs. 55.6% at 6 months and 52.8% vs. 41.9% at 12 months).
Guldan et al. (2000) looked at the impact of home-based peer counseling on continued breastfeeding practices. Following a year-long intervention, the percentage of women reporting breastfeeding infants between the ages of 4-12 months was higher in the intervention group than in the control group (83% and 75%, respectively). The findings were statistically significant; however, as noted before, the lack of baseline and/or randomization of the intervention and control groups makes attribution of change to the intervention challenging.
Kramer et al. (2001) looked at the impact of facility and lay counseling and small media on continued breastfeeding practices in Belarus. Following the intervention, breastfeeding rates at 12 months were 19.7% in the intervention group compared with 11.4% in the control group (adjusted OR=0.47).
Kupratakul et al. (2010) found that the proportion of women practicing “predominant” breastfeeding at 6 months, defined as “infants … fed with the mothers’ breast milk and water, sweetened water and juices without formula” was significantly higher among women in the intervention group compared with those in control group (40.0% vs. 5.3%).
Longitudinal studies
Duyan Camurdan et al. (2007) conducted a longitudinal study looking at the impact of the BFHI initiative on breastfeeding practices in Gazi University Hospital in Turkey. All women delivering in the hospital were invited to participate in periodic well-child visits. During each visit, providers reinforced the benefits of breastfeeding and provided counseling. The average duration of breastfeeding was 21.17 months among women giving birth after the introduction of BFHI, compared with 17.83 months among those giving birth prior to the introduction of BFHI. Following regression analysis that controlled for variables that might affect breastfeeding practices (mother’s occupation, educational status, age, way of birth, parity, duration of pregnancy, birth weight) the breastfeeding rate was 1.5 times higher among women giving birth after the introduction of BFHI, compared with those giving birth prior to the introduction of BFHI.
Table 3.2.2: Number of studies reviewed (with number of studies reporting statistically significant results in parentheses), by SBCC approach and study design
| SBCC approaches and activities |
Reviews with meta-analysis |
RCTs | Longitudinal studies |
Repeated cross-sectional studies |
Cross-sectional studies |
TOTAL |
| Interpersonal Communication | 6 (6) | 27 (21) | 14 (9) | 11 (9) | 1 (1) | 59 (46) |
| One-on-One Counseling | 6 (5) | 25 (20) | 12 (9) | 10 (9) | 1 (1) | 48 (44) |
| Counseling in facilities | 0 | 8 (5) | 9 (6) | 2 (2) | 1 (1) | 20 (14) |
| Counseling in communities | 2 (2) | 0 | 2 (2) | 2 (2) | 0 | 6 (6) |
| Counseling in homes (home visits) | 4 (4) | 23 (18) | 4 (3) | 5 (3) | 0 | 32 (28) |
| Counseling in other settings | 2 (2) | 0 | 0 | 2 (2) | 0 | 4 (4) |
| Group Education | 3 (3) | 6 (5) | 6 (3) | 4 (4) | 1 (1) | 17 (16) |
| Education in facilities | 0 | 1 (1) | 3 (3) | 0 | 1 (1) | 5 (4) |
| Education in communities | 1 (1) | 5 (4) | 4 (2) | 3 (3) | 0 | 12 (10) |
| Education in other settings | 2 (0) | 0 | 0 | 1 (0) | 0 | 3 (0) |
| Support Group | 1 (1) | 4 (3) | 5 (3) | 5 (3) | 0 | 14 (10) |
| Support group in facilities | 0 | 2 (2) | 4 (3) | 1 (1) | 0 | 7 (6) |
| Support group in communities | 1 (1) | 2 (1) | 1 (0) | 2 (0) | 0 | 5 (2) |
| Support group in other settings | 0 | 0 | 0 | 2 (0) | 0 | 2 (0) |
| Media | 0 | 5 (5) | 8 (6) | 9 (6) | 2 (2) | 24 (19) |
| Mass media | 0 | 0 | 0 | 4 (3) | 1 (1) | 5 (4) |
| Mid-sized media (community radio / video, local billboards | 0 | 0 | 2 (2) | 5 (3) | 1 (1) | 8 (6) |
| Small media (posters, flyers, calendars, reminder stickers | 0 | 4 (4) | 7 (5) | 7 (5) | 1 (1) | 19 (15) |
| Traditional media (songs, drama) | 0 | 3 (3) | 0 | 0 | 0 | 3 (3) |
| Social media (Twitter, Facebook, etc.) | 0 | 0 | 0 | 0 | 0 | 0 |
| Community/Social Mobilization | 1 (1) | 4 (3) | 2 (1) | 9 (6) | 0 | 16 (11) |
| Campaign, event, special "days" | 0 | 0 | 0 | 2 (2) | 0 | 2 (2) |
| Issue groups | 1 (1) | 3 (2) | 2 (1) | 5 (2) | 0 | 11 (7) |
| Other | 0 | 1 (1) | 0 | 3 (3) | 0 | 4 (4) |
Repeated cross-sectional studies
Guyon et al. (2009) conducted a repeated cross-sectional study looking at breastfeeding practices in Madagascar. Following an intervention that included community mobilization, campaign events, and mass media, rates of breastfeeding among children 12-15 months of age increased from 89% to 94%, and among children 20-23 months of age the rate increased from 43% to 73%.
Summary of the Evidence
SBCC approaches utilized
For the purposes of analysis, we organized SBCC activities into three primary categories or approaches: interpersonal communication, media, and community/social mobilization. These approaches and the specific activities associated with each are described in further detail in the introductory chapter.
The studies reviewed included a broad range of approaches to promote optimal breastfeeding practices. Most included more than one specific SBCC activity and many utilized more than one approach.
All six literature reviews focused on interpersonal communication (IPC) approaches. One review included findings related to community/social mobilization.
Among the 42 other studies reporting statistically significant results, 40 included interpersonal communication (IPC) approaches, 19 included some form of media, and 10 included community/social mobilization. One-on-one counseling in the home was the most commonly used IPC approach, small media was the most commonly used media approach and the gathering of issues groups was the most commonly used community/social mobilization approach (see Table 3.2.2).
Nine of the reviewed studies explored the impact of the BFHI on breastfeeding practices. In some cases an SBCC intervention was added to the standard BFHI package, while in other cases, facilities implementing BFHI were compared with those who were not doing so.
There is a wide breadth of evidence regarding the effectiveness of various SBCC approaches in changing breastfeeding practices. Table 3.2.3 provides a summary of these findings.
Interpersonal communication
Six reviews with meta-analysis establish the effectiveness of IPC approaches with respect to uptake of optimal breastfeeding practices. All but one looked solely at IPC interventions. Gogia and Sachdev (2010) reported a positive effect of IPC with respect to early initiation of breastfeeding (RR=3.35), while Imdad et al. (2011) reported a six-fold increase in EBF rates at six months following breastfeeding promotion in developing countries (RR=1.37 overall). Renfrew et al. (2012) found that extra support led to a lower risk of having stopped EBF at six months (RR=0.86), Hall (2011) and Jolly et al. (2012) reported similar effect related to EBF at other times, and Imdad et al. (2011) reported a positive effect of IPC on continued breastfeeding at six months (RR=1.12). Lassi et al. (2010) reported an RR of 1.94 with respect to early initiation of breastfeeding, but the intervention package included both IPC and community/social mobilization, making attribution to IPC difficult. None of the reviews reported findings associated with breastfeeding at two years of age, the universally recommended practice.
In addition to the reviews, 41 primary research studies reported statistically significant outcomes related to the effect of IPC interventions on breastfeeding initiation, 14 reported significant outcomes with respect to breastfeeding initiation, 27 reported on EBF, and nine reported significant outcomes with respect to continued breastfeeding rates.
Eighteen studies measured the effectiveness of IPC independent of other SBCC approaches, including six which were not included in the previously described reviews: three RCTs, two longitudinal studies, and one repeated cross-sectional study. Among the RCTs, Bortolini and Vitollo (2012) reported the most compelling findings, a difference between intervention and control communities of 10.7 percentage points with respect to breastfeeding at six months and 10.9 percentage points with respect to breastfeeding at 12 months.
In the only study looking at continued breastfeeding at two years of age, Guyon et al. (2009) reported an increase of 30 percentage points in their repeated cross-sectional study.
Indicators and study design varied considerably among studies, making consolidation of findings from studies not included in previous literature reviews difficult.
Media
We did not identify any reviews documenting the effectiveness of media-related interventions on breastfeeding outcomes, but did identify 19 primary research studies reporting significant outcomes associated with interventions that included media-based SBCC approaches. Nine reported a significant effect with respect to breastfeeding initiation, 13 reported a significant effect with respect to uptake of EBF, and three reported a significant effect with respect to continued breastfeeding practices.
Only one of these studies, Gupta et al. (2004), presented findings regarding the effect of media-based approaches independent of other SBCC approaches. Researchers measured the impact of exposure to messaging through four different media channels: radio, television, posters, and print materials. They reported a greater likelihood of EBF for six months among those exposed to messages through these forms of media two or more times.
Community/Social mobilization
Lassi et al., (2010) included findings related to community/ social mobilization. The review, however, looked at both community/social mobilization and IPC, and did not disaggregate results between the two approaches. It is therefore not possible to assess the effect of IPC independent of other SBCC approaches.
Eight primary research studies reported a significant effect of interventions including community/social mobilization with respect to breastfeeding initiation, six reported a significant effect with respect to EBF, and one reported a significant effect with respect to continued breastfeeding.
None of these studies reported on the effect of community/ social mobilization in isolation from other SBCC approaches, so it is difficult to derive conclusions regarding the singular effectiveness of community/social mobilization in changing breastfeeding practices.
Multiple SBCC approaches
One review documented the effect of multiple SBCC approaches on the uptake of breastfeeding practices. Lassi et al. (2010) reported an RR of 1.94 among studies employing IPC and/or community/social mobilization to improve uptake of early initiation of breastfeeding. They did not, however, distinguish among studies employing IPC, community/ social mobilization, or both.
In addition to the review by Lassi et al. (2010), we identified twenty-three primary research studies reporting positive outcomes associated with interventions that included more than one SBCC approach. Eleven reported a positive effect with respect to early initiation of breastfeeding, 15 report a positive effect with respect to EBF, and three reported a positive effect with respect to continued breastfeeding at six months or older. Among the most notable are three RCTs not included in the Lassi et al. (2010) review: Bang et al. (2005); Morrow and Guerrero (2001); and Omer et al. (2008).
None of these studies measured the effect of the utilization of more than one SBCC approach compared with the utilization of just one approach. One must use caution, therefore, in drawing conclusions about the effectiveness of interventions employing multiple approaches versus those employing just one approach.
Implementation processes followed
Few of the studies reviewed in this chapter assessed the effect of implementation processes (e.g. intensity and timing of communications; type and training of person communicating messages, or target audience) on the effectiveness of the SBCC approach on breastfeeding practices, nor did they consistently report all aspects of the implementation processes.
Little work has been done to compare the effect of timing or frequency of interventions. Imdad et al. (2011) conducted a sub-analysis which showed that “prenatal counselling had greater impacts on breastfeeding rates at 4-6 weeks, while combined prenatal and postnatal promotion were important for breastfeeding rates at 6 months.” Renfrew et al. (2012) found no effect of the timing of the support (during the antenatal or postnatal visit), while Jolly et al. (2012) found that “combined antenatal and postnatal peer support was not associated with a significant improvement in not breast feeding at last study follow-up, whereas postnatal only interventions did significantly reduce not breast feeding.” Likewise, Aidam et al. (2005) found that women who received two prenatal counselling sessions, one counseling session 48 hours postpartum, and six postpartum home visits were just as likely to EBF at six months as those who received only postpartum counselling.
The intensity of communications ranged between one and twenty visits or sessions, and several meta-analyses and studies did make comparisons between the effectiveness of interventions with varied intensities. Jolly et al. (2012) found that women in the more intensive interventions (≥5 contacts planned) were more likely to report breastfeeding at last follow-up compared with usual care. Similarly, Renfrew et al. (2012) found that “studies with four to eight visits seemed to be associated with a more pronounced treatment effect.” However, the authors warned that “care is needed in the interpretation of this finding as there is inconsistent reporting due to variations in the timing of outcome assessments, and the settings of studies and the population groups included in studies with more faceto- face visits also varied. It is likely that support will be most effective when it reflects the local needs of the population.” Renfrew et al. also found a greater treatment effect on EBF at six months when the “intervention was delivered by nonprofessionals (average RR=0.74) compared with professionals (average RR=0.93) or both (average RR=0.76).” Likewise, women who received face-to-face support were 19% less likely to have given up EBF at six months compared with those who did not (average RR=0.81).
Piwoz et al. (2007) conducted an RCT comparing those who received no visits, one visit, two visits or three visits when counselling on HIV and infant feeding was provided. They found that EBF at six weeks and three months was incrementally higher with each additional visit. Statistical significance of the differences between the groups were not reported, but the authors did conclude that “frequency of intervention contact was associated with safer breastfeeding knowledge and exclusive breastfeeding practices.”
Finally, Morrow and Guerrero (2001) conducted an RCT in which they compared EBF at three months between those who received six home visits and those who received three home visits. While those receiving six home visits appeared to be more likely to EBF at three months than their counterparts who received only three home visits (a 15% difference), the statistical significance of the difference was not reported. Both were effective when compared with a control group.
In terms of target audiences, the majority of the studies reviewed focused attention on the mother or caregiver. Among the 48 studies (literature reviews and primary research) reporting statistically significant results, all targeted the mother of the child herself. Twenty-six targeted only the mother of the child herself. Fourteen also targeted direct influencers, 18 targeted local community actors, and three targeted actors of the enabling environment in addition to pregnant or lactating women.
Only Lassi et al. (2010) compared the effect of targeting one vs. multiple audiences or of targeting different audiences. The authors found that “the most successful packages were those that emphasized involving family members through community support and advocacy groups and community/ social mobilization and education strategies, provision of care through trained CHWs via home visitation, and strengthened proper referrals for sick mothers and newborns.”
Conclusions
The body of literature focusing on SBCC approaches being used to improve breastfeeding practices is strong and broad, supporting the claim that SBCC approaches can and do succeed in improving uptake of the behaviors promoted. While the literature may reflect a bias to publish positive results, it also underscores the important role of SBCC approaches in improving nutrition practices – practices which have been shown to have an impact on the nutritional status of women, infants, and children.
There is more consistency in how breastfeeding practices are defined and measured than in the other chapters of this review. This is particularly so in the case of early initiation of breastfeeding. However, even with globally-recognized indicators and measurement guidance from the WHO, considerable variation remains. For example, while fourteen studies reported statistically significant findings related to EBF children six months old, 19 reported on EBF at a range of other ages.
There is considerable variation in the SBCC interventions designed to improve breastfeeding practices – in the interactions or combinations
with other interventions, target groups, content, messages, scale and coverage, length and intensity, as well as context.
| SBCC approah | Initiating breastfeeding early | Exlusive breastfeeding | Continued breastfeeding |
| Interpersonal Communication |
20, 16 reporting statistical
|
40, 31 reporting statistical significance:
|
16, 10 reporting statistical
|
| Media |
10, 9 reporting statistical
|
17, 13 reporting statistical significance:
|
6, 3 reporting statistical
|
| Community/social mobilization |
11, 8 reporting statistical
|
8, 6 reporting statistical
|
2, 1 reporting statistical
|
Interpersonal communication was the most prevalent SBCC approach found to be effective at improving breastfeeding practices in both implementation and research. While media and community/social mobilization were used, they were almost always used with at least one other communication approach. The evidence included in the review also reaffirms the importance of peer support for improving breastfeeding practices, whether one-on-one or in groups. The question remains as to whether many projects use IPC because it is more effective than other approaches at improving breastfeeding practices, or whether it has become the default approach. Given the complex and personal nature of the suite of behaviors making up optimal breastfeeding, it is likely the former. Unfortunately it is both difficult and expensive to disaggregate the contribution of single channels or specific activities within a multi-channel intervention, and an important question for SBCC practitioners and researchers is whether that line of research is useful.
SBCC interventions to promote breastfeeding practices are suited to iterative programming, because even if standards for behaviors aren’t met, there can be significant movement toward the standard. In a hypothetical example, the target number of women exclusively breastfeeding at six months might not be met, but many women may have continued EBF for a month or two more than they would have had the intervention not happened. In this example, the intervention may not have succeeded in achieving the standardized indicator, but moved the EBF duration in the right direction. Iterative interventions can learn what supported that move in the right direction, and build on them.
All of the studies reporting statistically significant results targeted the woman herself. Nearly half targeted other audiences or influencers as well. Little has been done to compare the effect of targeting one vs. multiple audiences or of targeting different audiences, but what was done suggested that targeting multiple contacts has a greater effect than targeting only the woman herself, given the important role of husbands, mothers-in-law, and community leaders. With regard to timing, few comparisons were made between the effect of different timings and what little was done presents contradictory evidence.
There is some evidence that intensity (number of contacts) influences the effectiveness of SBCC interventions in promoting adoption of optimal breastfeeding practices. The intensity of communications ranged between one and twenty visits or sessions. While some studies reported statistically significant improvements in breastfeeding practices after only one or a few contacts, evidence from several metaanalyses and primary research studies strongly suggests that increasing the number of contacts increases the positive effect of SBCC interventions on breastfeeding practices.
The majority of studies were implemented on a relatively small scale, within a few health facilities or communities and typically with fewer than 500 study participants per group. Because of these limitation and due to the lack of standardization in the way research related to SBCC is designed and described, it is challenging to make conclusions beyond the fact that projects with SBCC will result in uptake of promoted practices.
The evidence for SBCC and breastfeeding suggests the effect of context (including social norms, culture, and environmental factors) as well as existing national breastfeeding rates. This underscores the importance of proper context assessments, formative research and/or ethnographic study prior to SBCC implementation.
While there is much to be learned from this large body of literature to aid us in developing future programs, a number of questions remain. These include questions related to:
- the positive (or negative) effect of using multiple SBCC approaches compared with focusing on only one;
- the positive (or negative) effect of targeting multiple audiences or influencers of the behaviors being promoted, rather than focusing on just one target population;
- the role of context, in other words, the effect of the same SBCC intervention implemented in different contexts;
- the effectiveness of different approaches (including intensity and targeting) for different behaviors;
- the cost and cost effectiveness of various SBCC approaches (particularly as it relates to scalability); and
- the effectiveness and sustainability of these approaches when implemented at scale.
