Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J Perinatol. Author manuscript; available in PMC 2010 August 27.
Published in final edited form as:
PMCID: PMC2929160

Can Early Postpartum Home Visits by Trained Community Health Workers Improve Breastfeeding of Newborns?



Whether postpartum visits by trained community health workers (CHW), reduce newborn breastfeeding problems.


CHWs made antenatal and postpartum home visits promoting newborn care practices including breastfeeding. CHWs assessed neonates for adequacy of breastfeeding and provided hands on support to mothers to establish breastfeeding. History and observation data of 3,495 neonates were analyzed to assess effects of CHW visitation on feeding problems.


Inappropriate breastfeeding position and attachment were the predominant problems (12% –15%). 6% of newborns who received home visit by CHWs within 3 days had feeding difficulties, compared to 34% of those who did not (OR: 7.66, 95% CI: 6.03–9.71, p=0.00). Latter group was 11.4 times (95% CI: (6.7–19.3, p=0.00) more likely to have feeding problems as late as day 6–7, than the former.


Counselling and hands on support on breastfeeding techniques by trained workers within first 3 days of birth, should be part of community based postpartum interventions.

Keywords: Breastfeeding, Community Health Workers, Newborn care, Postpartum visit, Bangladesh


Breastfeeding is fundamental to child survival and development. Along with the nutritional value of breast milk, it provides immunity against common infections 1, 2. The risks of death due to pneumonia and diarrhoea are, respectively, five and seven times greater among infants that are not breastfed during the first 5 months of life3. Breastfeeding also facilitates mother-baby bonds, contributing to better neurodevelopment 4. Exclusive breastfeeding in the first 6 months and continuation from 6 to 11 months has been identified as the single most effective preventive intervention in reducing child mortality, with the potential of 13% reduction of under-five child deaths worldwide 5. In addition to exclusive breastfeeding, recent evidence demonstrates that initiation of breastfeeding in the first hour of life and within day 1 has the potential of averting 22% and 16% of neonatal deaths, respectively 6. Successful long term breast feeding depends upon a successful start. Early initiation of breastfeeding is associated with greater communication between mothers and infants 7. Delayed initiation of breastfeeding and pre-lacteal feeding is common in Bangladesh8. Pre-lacteal feeding is associated with delay in establishment of breastfeeding. 9. Late initiation of breastfeeding and supplementary feeding are strong risk factors of for early breastfeeding termination. 10, 11. Early initiation, feeding of colostrum, avoiding pre-lacteal feeds and maintenance of exclusivity are the key issues for breastfeeding interventions.

Studies on Breastfeeding promotion to date have been primarily hospital based 12. Central to such efforts are the Baby Friendly Hospital Initiative13, 14. Similarly, the national breastfeeding programme in Bangladesh has also concentrated its activities more in urban hospitals and on training doctors and nurses, whereas involvement of grassroots workers at the community level is conspicuously absent15.

Large-scale improvements in breastfeeding practices at the community level have been demonstrated in Africa and Latin America 16. Yet, further information is needed on effective strategies for promoting exclusive breastfeeding and achieving high and equitable coverage 17. In the immediate postnatal period, it has been shown that if mothers are taught good breastfeeding techniques in a ‘hands-off’ style, which enables mothers to position and attach their babies for themselves based on a physiological approach, the incidence of perceived milk insufficiency decreases and breastfeeding rates are increased 18. In countries such as Bangladesh, where nearly 90% of deliveries take place at home, strategies need to be devised to empower mothers to initiate and sustain successful breastfeeding at the place they deliver and typically remain confined up to a period of at least 7 to 40 days postpartum19.

This paper presents the findings from a community-based intervention in which community health workers (CHWs) made three scheduled home visits in the first week after delivery, assessed postpartum mothers for breastfeeding techniques, provided counselling and hands-on support to establish successful breastfeeding. We describe the nature of ‘feeding problems’ and examine if early postpartum visits by trained health workers can improve rates of successful breastfeeding in neonates with no other associated neonatal morbidity.


Study design: parent trial

A cluster randomized trial to test the effectiveness of a community-based intervention package to reduce neonatal mortality was conducted from 2003 to 2005 in three rural sub-districts of Sylhet District of Bangladesh, located at the north-eastern corner of the country. Twenty four clusters with a total population of 486,351 were randomized into two intervention arms (Home Care and Community Care) and a comparison arm. This paper is based on data from the Home Care (HC) arm where trained community health workers, one per 4,000 population made two antenatal and three postpartum home visits to promote and support practices for birth and newborn care preparedness (BNCP) and newborn care including support for breastfeeding.

Intervention strategies in the Home Care intervention arm

CHWs conducted routine surveillance to identify pregnancies and made two antenatal BNCP sessions during the second and the third trimesters, involving the expectant mother and other maternal and newborn care providers from within and outside the family. The CHWs also made three postpartum neonatal care (NC) visits between days 1–3, 4–5 and 6–7 of birth. Counselling, motivation, negotiation and demonstration on maternal and newborn care, care seeking in case of complications during pregnancy, delivery and in the postpartum period and supply of a clean delivery kit and iron-folic acid tablets were the focus of BNCP sessions. Postpartum NC visits focused on immediate newborn care, promotion and support for breastfeeding, recognition of newborn danger signs, care seeking and management of newborn infections, including referral.

A female Field Supervisor supervised and supported the work of 8 CHWs; the supervisors accompanied and observed activities of each CHW at least for 2 days in a month for the entire day. Counselling and newborn assessment sessions organized by CHWs were observed using a structured supervisory checklist, scored and written feed-back provided at the end of each day's observation. Findings from such observations were discussed at monthly CHW refresher training sessions.

Training of CHWs on breastfeeding

CHWs received 21-days of classroom, hospital and community-based training on essential newborn care which included an 8-hour module on breastfeeding counselling and support, followed by a 6-hour practical session on observation and assessment of breastfeeding, and a 4-hour practical session on counselling in the community, totalling approximately 2.5 days. Topics covered included, (i) importance and basic features of breastfeeding, (ii) techniques of breastfeeding, (iii) milk expression, (iv) common problems of the breasts during breastfeeding and, (v) counselling points on breastfeeding. Training methods included lectures, hands-on demonstration, and practical exercises with real-life postpartum breastfeeding mothers and video-guided lessons.

A female trainer observed each trainee CHW while assessing a new breastfeeding mother, using a structured checklist. Refresher training was organized on areas needing improvement in assessing and supporting breastfeeding mothers.

Assessment of breastfeeding by CHWs

CHWs assessed and recorded the breastfeeding status of mothers in their visit record forms (VRF) during each postpartum NC visit. Assessments were based on information obtained from the mother and observation of breastfeeding using the algorithm presented in Table 1, which is part of the Integrated Management of Childhood Illness (IMCI) algorithm adapted in Bangladesh to include assessment of neonates in the first week of life. As shown in Table 1, CHWs asked mothers about convenience, frequency, pre-lacteal feeds, and supplementary feeding and observed feeding practices. Observations were made only if the mother had not breastfed the newborn in the past 1 hour, allowed the CHW to observe, and the baby had no clinical condition such as observed convulsion or unconsciousness that required urgent referral to a hospital. Observation of breastfeeding using the algorithm included quality of positioning (chin touching the breast, mouth wide apart, lower lip folded outwards, areola exposed more above lip than below), attachment (head and body upright, length of body lying against the mother, whole body held, faced towards breast below) and strength of sucking to determine whether or not the newborn had a 'feeding problem'.

Table 1
CHW algorithm for assessment of breastfeeding through history taking and observation

Based on findings of the assessments, CHWs were trained to help mothers to correct positioning or attachment, and to provide counselling to encourage exclusive breastfeeding. Hence, categorization of a newborn as having a 'feeding problem' implied that the mother received practical coaching and demonstration from the CHW in one or more of the three technical areas (i.e., positioning, attachment, sucking). CHWs were not explicitly aware of the study hypothesis, though this was known to them that their counselling and coaching were meant to improve the feeding practices of mothers.

Sampling and selection criteria

Records from VRFs for 13,912 women who delivered in the Home Care arm between July 2003 and December 2005 were initially considered in this analysis. In order to examine the effect of CHWs' early visitation on mothers’ competencies of breastfeeding techniques, we used the following exclusion criteria to extract the sub-sample included in the final analysis: a) those who delivered in a facility or outside the HC arm (n=1,655), b) stillbirths or abortions (n=720), c) those who did not receive any CHW home visit within 28 days (n=2,682), d) those who did not receive a late 1st week visit (6–7 days) (n=2,180) because, without an assessment of the newborn during these days it would not be possible to compare status of breastfeeding between those who received an early visit by CHW and those who did not, e) those who received a second visit around 4–5 days (n=2,836) because, a visit received before the late 1st week visit was likely to influence the status of feeding during an assessment during the late 1st week and, f) those who had any associated morbidity with or without feeding problems (n=344); newborns who were classified as 'not able to feed' as indicated in row 1 of column 3 of Table 1, were included in this last group. This led us to an effective sub-sample of 3,495 (Figure 1).

Figure 1
Exclusion criteria for selection of effective sub-sample

Protection of human subjects

The study protocol received ethical clearance from the Committee on Human Research at the Johns Hopkins University, USA, and from the Research Review Committee and Ethical Review Committee of the International Centre for Diarrhoeal Disease Research, Bangladesh. A Data Safety Monitoring Board reviewed findings in 2005 and recommended continuation of the intervention as planned. An informed consent form was completed for each of the pregnant women at the time of registration.

Data analysis

We calculated the proportion of mothers who received different antenatal and postnatal visits by the CHWs. Two key groups were compared: Group A - those who received both early (1–3 days) and late (6–7 days) 1st week postpartum visits and, Group B: those who received only a late 1st week postpartum visit. Groups were compared for selected background characteristics including household socio-economic status, maternal age and education, preceding birth interval, parity, outcome of pregnancy, sex of the newborn, gestational age, birth weight, and feeding of pre-lacteals. Chi-square for linear trend and Chi-square tests were performed for to detect statistically significant differences between the two groups

The two groups were then compared for feeding problems recorded at late visit assessments. A total of six factors were compared: mother’s report of difficulty in feeding, feeding frequency of < 8 times a day, and receiving supplementary food; and, from on-site observation: proper positioning, proper attachment, and, good strength of sucking. To assess the degree of association between early postnatal visit and breastfeeding status at a late first week visit, we calculated unadjusted odds ratios (OR) and 95% confidence intervals (for 39 cases in Group A and 7 in Group B, findings from observation of breastfeeding were missing as they fed infants in the preceding one hour). We then developed two multiple logistic regression models to estimate the net effect of early postnatal visit on breastfeeding status at a late first week visit. In Model 1, predictor variables were, no early (1–3 days) CHW visit and, background characteristics that differed significantly between groups and, in Model 2, we adjusted for all background characteristics. Both models were further adjusted for intra-cluster correlation around CHWs. Epi Info 2000 and STATA 8 packages were used for bi-variate and multivariate analyses.

Role of the funding source

The sponsors of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.


Program Coverage

Figure 2 shows average program coverage for two antenatal BNCP visits and three postpartum NC visits for the 30 months period of intervention implementation. For the first antenatal visit, the coverage was almost universal (97%) over the entire period. Coverage for the second antenatal visit was comparatively lower with an average of 80% - due primarily to migration of pregnant women out of the HC area during late pregnancy (i.e., 32–34 weeks) to deliver at their natal homes.

Figure 2
BNCP (antenatal) and NC (postpartum) visit coverage by CHWs

Coverage for the early 1st week postpartum visit (1–3 days) was 63% and for the mid-week visit within 5 days and for the late 1st week visit within 7 days was 71% and 77%, respectively.

Background characteristics of the sample

Table 2 presents the comparison of the two groups for selected background characteristics. Of those who received both early and late 1st week visits (Group A), there were more primiparous women in (20.2%) than those who received only late 1st week visits (Group A) (16.5%) (p=0.04). Among Group A women, 17.7% delivered before 37 weeks gestation, while 24.5% of the Group B delivered preterm infants (p=0.00). Groups also differed significantly in terms of proportion of low birth weight infants (< 2500 g) (p=0.01) and feeding of pre-lacteals (p=0.00).

Table 2
Selected background characteristics of respondents by status of receiving CHW visit during first week after delivery

Feeding problems in early and late visit receivers

Table 3 presents findings of breastfeeding assessments by CHWs during NC (postpartum) visits made in the first week. Findings of two assessments for Group A have been presented, (i) from the early visit assessment and, (ii) from the late visit assessment.

Table 3
Nature of feeding problems as assessed by CHWs at early and late visits by status of receiving visits

Only 6% of newborns in Group A were classified to have feeding difficulties, compared to 34% in Group B during late first week visit (OR: 7.66, 95% CI: 6.03–9.71, p=0.00). CHWs classified the newborns to have a ‘feeding problem’ primarily on the basis of their observed findings rather than the information collected from mothers through history taking. Significantly higher proportions of mothers in Group B compared to Group A were found to have each of the six factors in the assessment of breastfeeding problems (p=0.00 for all six factors).

Factors affecting breastfeeding status

Table 4 presents findings from unadjusted analysis of factors affecting breastfeeding status during late 1st week visit and of two adjusted logistic regression models. In unadjusted analysis, mothers who did not receive an early (1–3 days) CHW visit were 7.7 times more likely to have ‘feeding problems’ duringthe late 1st week (6–7 days) visit. Similarly, those who provided any kind of pre-lacteal feed to their newborn were 2.9 times more likely to have a late 1st week feeding problem than those who did not provide any pre-lacteal feed. Other considered predictors did not appear to have any significant association with the outcome. In model 1, absence of an early CHW visit (OR: 11.3, 95% CI: 6.7, 18.9, p=0.00) and feeding of pre-lacteals (OR: 2.8, 95% CI: 1.3, 5.9, p=0.01) were significantly associated with having a feeding problem at a late 1st week visit. When adjusted for all other predictors in model 2, absence of an early CHW visit (OR: 11.4, 95% CI: 6.7, 19.3, p=0.00) and feeding of pre-lacteals (OR: 2.5, 95% CI: 1.1, 5.7, p=0.03) continued to have significant association with feeding problem persisting at late visit.

Table 4
Logistic regression model for persistent feeding problem among newborns beyond 6th postpartum day


This study documented that problems related to breastfeeding are common in Bangladeshi newborns, a majority of the feeding problems were related to inappropriate positioning and attachment and an early home visit by a trained CHW to provide counselling and support can significantly reduce the prevalence of breastfeeding problems at the end of the first week of life. Mothers who were not visited by a CHW within 3 days of life were more likely to have ‘feeding problems’ at late first week visit (day 6–7) than those who were visited early by CHWs. Similarly, neonates of mothers who fed pre-lacteals were more likely to have ‘feeding problems than those did not feed pre-lacteals. These findings add to the existing body of literature on the role of trained CHWs on overcoming feeding problems in the early postpartum period.

Hospital-based promotional strategies, mostly focused around the Baby Friendly Hospital Initiative, have demonstrated success in improving breastfeeding practices in the hospital setting20, 21. However, there is growing concern that strategies of this Initiative alone cannot ensure that these practices are sustained when mothers return home. Furthermore, in settings like Bangladesh, where the vast majority of deliveries occur at home, the Baby Friendly Hospital Initiative is clearly inadequate. A randomized controlled trial comparing a hospital-based system and the same system combined with a home visits reported that although 70% of mothers breastfed at discharge from the hospital, only 30% of them exclusively breastfed at day 10 at home12. This finding reiterates the need for breastfeeding support that extends to the home 15, 21, 22, 23.

A trained worker providing hands-on support to the mother to improve breastfeeding techniques has been demonstrated and recommended in different settings18, 24, 25. Fifteen pregnancy and immediate postpartum visits by peer counsellors in Bangladesh25, and group and one-to-one peer coaching in rural Scotland24 resulted in better performance among the intervention groups in terms of early initiation and lack of pre-lacteal feeding. Our observations, though assessed through a different methodology, also show a higher success in overcoming feeding problems at days 6–7 in mothers who received an earlier visit by a trained CHW compared to those who did not.

Successful initiation of breastfeeding is a critical factor in the maintenance of exclusive breastfeeding. Even if not associated with other morbidities, it is often seen that mothers cease to breastfeed newborns at a very early stage. Anxiety over sufficiency of breast milk supply is the most common problem, in that it often results in cessation of breastfeeding in the early stages26. As a rule, however, this anxiety is unfounded as inability of the infant to obtain sufficient milk is generally due to improper mechanical technique rather than any biological insufficiency22. Our findings suggest that psychological support for breastfeeding mothers through early counselling and hands-on support for achieving proper techniques, particularly position and attachment, can aid the mother in preventing and overcoming feeding problems in the early postpartum period. A similar finding was reported by Ingram et al who utilized midwives to teach good breastfeeding techniques 18.

When controlled for other predictors of breastfeeding, CHWs’ absence in the early postpartum period was found to have a strong association with persistent feeding problems later in the first week of life. Only one other factor, feeding of pre-lacteals, had a significant association with feeding problems at the end of the first week. The process of such an association can be explained by the negative influence of pre-lacteal feeding on "coming in" of milk. It was reported that pre-lacteal feeding accounted for 44% of variations in coming in of milk 9. This observation emphasizes the need for coordinated efforts for promotion of proper infant feeding practices in our rural communities.

The findings presented have several limitations. Although we compared selected background characteristics of the groups, some degree of differences in practices may be attributed to unobserved heterogeneity that actually made one group (group A in particular) more likely to receive an earlier visit by CHWs. As CHWs scheduled their visits based on expected date of delivery, it is likely that they missed the information of deliveries occurring preterm and failed to reach families in the first week postpartum. Also, as the CHWs provided the breastfeeding intervention and served as the data collectors simultaneously, observer bias could have resulted in generalized under-reporting of improper positioning and attachment. While interpreting the findings of the study, this should also be kept into consideration that the groups compared, had disproportionate numbers.

Certain features of the data presented in this paper are unique. The intervention itself was delivered and immediately monitored by rural young female CHWs and their supervisors selected from the intervention locality. Data that we presented from the VRFs were collected through direct on-site observation of breastfeeding at the community, which has better authenticity over reported data.


The growing evidence now demands that strategies be integrated into neonatal care and postpartum maternal care interventions. Designing appropriate package and delivery strategies customized to country situations is the key to effectiveness of any such integrated approach. Our findings provide evidence of effectiveness of breastfeeding support through early postnatal visits as part of a larger newborn and maternal care intervention. We conclude that, in addition to large scale mass communication based promotional approach, one-on-one counselling and hands on support to mothers for proper breastfeeding techniques by trained workers should be part of any postpartum package and, such support should be made available at the very early days, possibly within 72 hours, to engender a successful initiation of breastfeeding.


We thank the study participants in Sylhet District Bangladesh who were generous with their time and patience with CHWs through the sessions of observation and interviews. We also thank Dr. Mahbub E Elahi Chowdhury of Reproductive Health Unit, ICDDR,B who provided advice on statistical issues.

Source of Support

Funding for the PROJAHNMO Project is provided by the United States Agency for International Development through the Family Health and Child Survival Cooperative Agreement, Global research Activity Cooperative Agreement, and the USAID Mission in Dhaka, Bangladesh, and the Saving Newborn Lives initiative of Save the Children Federation - USA through a grant from the Bill and Melinda Gates Foundation.



IM conceptualized the current analysis, SMR developed manuals and trained CHWs, AS performed the literature review. IM did the data analysis, IM and AS wrote the draft manuscript and further modifications. NB, IM and ABM were responsible for supervision of data collection and data management. AHB, GLD, SEA and HRS designed the protocol for the parent study, reviewed the study instruments and the results of the study. AHB, PW and GLD provided feedback on the analysis and made critical suggestions on the manuscript. All authors reviewed the manuscript.

Conflict of interest statement

The authors declare that they have no conflict of interest arising through participation in the current study.


1. Hanson LA. Breastfeeding provides passive and likely long-lasting active immunity. Ann Allergy Asthma Immunol. 1998;81:523–533. quiz 533-524, 537. [PubMed]
2. Oddy WH. Breastfeeding protects against illness and infection in infants and children: a review of the evidence. Breastfeed Rev. 2001;9:11–18. [PubMed]
3. Victora CG, Smith PG, Vaughan JP, Nobre LC, Lombardi C, Teixeira AM, et al. Infant feeding and deaths due to diarrhea. A case-control study. Am J Epidemiol. 1989;129:1032–1041. [PubMed]
4. Lucas A, Morley R, Cole TJ, Lister G, Leeson-Payne C. Breast milk and subsequent intelligence quotient in children born preterm. Lancet. 1992;339:261–264. [PubMed]
5. Jones G, Steketee RW, Black RE, Bhutta ZA, Morris SS. How many child deaths can we prevent this year? Lancet. 2003;362:65–71. [PubMed]
6. Edmond KM, Zandoh C, Quigley MA, Amenga-Etego S, Owusu-Agyei S, Kirkwood BR. Delayed breastfeeding initiation increases risk of neonatal mortality. Pediatrics. 2006;117:e380–e386. [PubMed]
7. Renfrew MJ, Lang S, Woolridge MW. Early versus delayed initiation of breastfeeding. Cochrane database of systematic reviews (Online) 2000 CD000043. [PubMed]
8. National Institute of Population Research and Training (NIPORT) Bangladesh Demographic and Health Survey 2004. Dhaka, Bangladesh: National Institute of Population Research and Training (NIPORT), Mitra and Associates, ORC Macro; 2005. Mitra and Associates, Macro O.
9. Ahmed FU, Rahman ME, Alam MS. Prelacteal feeding: influencing factors and relation to establishment of lactation. Bangladesh Medical Research Council bulletin. 1996;22:60–64. [PubMed]
10. DiGirolamo AM, Grummer-Strawn LM, Fein S. Maternity care practices: implications for breastfeeding. Birth (Berkeley, Calif. 2001;28:94–100. [PubMed]
11. Marandi A, Afzali HM, Hossaini AF. The reasons for early weaning among mothers in Teheran. Bulletin of the World Health Organization. 1993;71:561–569. [PubMed]
12. Coutinho SB, de Lira PI, de Carvalho Lima M, Ashworth A. Comparison of the effect of two systems for the promotion of exclusive breastfeeding. Lancet. 2005;366:1094–1100. [PubMed]
13. WHO/Unicef. Protecting, promoting and supporting breastfeeding: the special role of maternity services: A Joint WHO/Unicef Statement. Geneva: WHO; 1989. [PubMed]
14. UNICEF. Take the baby friendly initiative. New York: UNICEF; 1991.
15. Ahmed S, Parveen SD, Islam A. Infant feeding practices in rural Bangladesh: policy implications. Journal of tropical pediatrics. 1999;45:37–41. [PubMed]
16. Quinn VJ, Guyon AB, Schubert JW, Stone-Jimenez M, Hainsworth MD, Martin LH. Improving breastfeeding practices on a broad scale at the community level: success stories from Africa and Latin America. J Hum Lact. 2005;21:345–354. [PubMed]
17. Bryce J, el Arifeen S, Pariyo G, Lanata C, Gwatkin D, Habicht JP. Reducing child mortality: can public health deliver? Lancet. 2003;362:159–164. [PubMed]
18. Ingram J, Johnson D, Greenwood R. Breastfeeding in Bristol: teaching good positioning, and support from fathers and families. Midwifery. 2002;18:87–101. [PubMed]
19. Winch PJ, Alam MA, Akther A, Afroz D, Ali NA, Ellis AA, et al. Local understandings of vulnerability and protection during the neonatal period in Sylhet District, Bangladesh: a qualitative study. Lancet. 2005;366:478–485. [PubMed]
20. Philipp BL, Merewood A, Miller LW, Chawla N, Murphy-Smith MM, Gomes JS, et al. Baby-friendly hospital initiative improves breastfeeding initiation rates in a US hospital setting. Pediatrics. 2001;108:677–681. [PubMed]
21. Haider R. The Baby-Friendly Hospital Initiative and Activities in Bangladesh. Bangladesh J Child Health. 1993;12:72–74.
22. World Health Organization. Essential Newborn Care and Breastfeeding. Geneva, Switzerland: World Health Organization; 2002.
23. McKeever P, Stevens B, Miller KL, MacDonell JW, Gibbins S, Guerriere D, et al. Home versus hospital breastfeeding support for newborns: a randomized controlled trial. Birth (Berkeley, Calif. 2002;29:258–265. [PubMed]
24. Hoddinott P, Lee AJ, Pill R. Effectiveness of a breastfeeding peer coaching intervention in rural Scotland. Birth Berkeley, Calif. 2006;33:27–36. [PubMed]
25. Haider R, Ashworth A, Kabir I, Huttly SR. Effect of community-based peer counsellors on exclusive breastfeeding practices in Dhaka, Bangladesh: a randomised controlled trial [see commments] Lancet. 2000;356:1643–1647. [PubMed]
26. Colin WB, Scott JA. Breastfeeding: reasons for starting, reasons for stopping and problems along the way. Breastfeed Rev. 2002;10:13–19. [PubMed]