Background The Baby-Friendly Hospital Initiative (BFHI) has been implemented by WHO and Unicef with a view to protect, promote and support breast feeding. This paper aims to assess the influence of the BFHI on breastfeeding indicators in Brazil, using data from the 2nd Survey of Breastfeeding Prevalence, conducted in 2008.
Methods Data on 64 municipalities were analysed: a total of 65 936 infants under the age of 1 year who were covered by the 2008 immunisation campaign. The outcomes of interest were breast feeding in the first hour of life in infants under 1 year of age; exclusive breast feeding on the first day after hospital discharge in infants under 4 months of age; exclusive breast feeding in infants under 2, 3 and 6 months of age; and pacifier use in infants under 6 months of age. The influence of birth in baby-friendly hospitals (BFHs) on these end points was analysed by means of Poisson regression with robust variance for complex samples.
Findings Infants born in BFHs were 9% more likely to be breast fed in the first hour of life and 6% more likely to be breast fed on the first day at home. Exclusive breast feeding was 13%, 8% and 6% more likely in infants under the ages of 2, 3 and 6 months, respectively, born in BFHs. Birth in a BFH also correlated with significant less pacifier use.
Conclusions The BFHI has had an impact on several indicators of breast feeding. The authors hope the results of this study will make policy makers and health professionals aware of the importance and potential of this strategy.
- Breast feeding
- health surveys
- healthcare policy
- primary care
- primary healthcare
Statistics from Altmetric.com
Breast feeding is the most effective strategy for prevention of child mortality and promotes physical, mental and psychological well-being in children.1 On the basis of scientific evidence, WHO recommends that children be exclusively breast fed for 6 months and that breast feeding be continued, with use of complementary foods, up to the second year of life or beyond.2
Despite widespread recognition of the superiority of breast feeding, exclusive breastfeeding (EBF) rates are still low, and the duration of breast feeding is still unsatisfactory in several countries.3
Studies conducted in Brazil have shown that breastfeeding rates are steadily increasing. Data are available from two sources: representative population-wide household surveys, which show that median duration of breast feeding has increased from 2.5 months in 1975 to 5.5 months in 1989,4 7 months in 19965 and 14 months in 2006,6 and surveys conducted in Brazilian capitals and the Brazilian Federal District during the nationwide immunisation campaigns of 1999 and 2008, which showed a 1-month increase in median duration of EBF (up to 54.1 from 23.4 days) and overall breast feeding (from 10 to 11.2 months).7 ,8
Progress is undeniable, but the reality revealed by recent surveys is still far from ideal when compared with WHO criteria.8 In this context, several strategies have been implemented in the country with a view to protect, promote and support breast feeding. One notable such strategy is the Baby-Friendly Hospital Initiative (BFHI), launched by WHO and the Unicef in 19919 and is part of WHO/Unicef Global Strategy for Infant and Young Child Feeding.10
The BFHI aims to ensure that all facilities providing maternity care follow the Ten Steps to Successful Breastfeeding and apply the principles and purposes of the International Code of Marketing of Breast-Milk Substitutes.9
Using data obtained from periodic reviews of the number of maternity care facilities given baby-friendly hospital (BFH) status, Unicef has reported that, as of late 2005, nearly 20 000 maternities in 156 countries had been designated as such at some point in time.11
When it was implemented in Brazil in 1992, the BFHI was expanded with the provision of financial incentives to hospitals achieving BFH status and the inclusion of further certification criteria, which include proof of compliance with the Brazilian Standard for Marketing of Infant and Young Child Foods (Norma Brasileira de Comercialização de Alimentos para Lactantes e Crianças de Primeira Infância, NBCAL), availability of a trained professional for maternal and newborn care during delivery, allowing presence of a companion during rooming-in and having a Cesarean section rate in line with thresholds established by state- or municipal-level health administrators or proving that Cesarean section rates are declining towards these thresholds.12 Development of these criteria sought to establish connections and synergy among the various Brazilian strategies and policies for maternal and child health.
At the time of writing, Brazil had approximately 3750 maternity care facilities, 335 (9%) of which are BFHs, providing baby-friendly coverage to approximately 21% of deliveries in the country every year.12
The present study seeks to assess the influence of the BFHI on breastfeeding indicators in Brazil, using data collected during the 2nd Survey of Breastfeeding Prevalence, conducted by the Brazilian Ministry of Health in 2008.
The 2nd Survey of Breastfeeding Prevalence was conducted during the second stage of a broad immunisation campaign carried out in 2008, with the purpose of characterising feeding practices among infants under the age of 1 year. As Brazilian annual immunisation campaigns achieve coverage of practically 100% of the population, surveys conducted as part of these campaigns can provide precise information on various child health end points. Besides, this strategy has been widely recommended and used in Brazil because it enables to collect data within a short period of time and at relatively costs.15
In addition to the 27 Brazilian capitals (26 states and Federal District), as originally planned in the survey design, various municipalities in several states took part in the survey. This sample includes infants living in the 27 Brazilian capitals, representing the situation of all states of the country, and in other 37 large municipalities that participated in the survey if >4000 infants under 1 year were included in the vaccination campaign.
Details on the methods of the survey have been described elsewhere.8 Workers in all municipalities received training on how to conduct the survey and were given instructional handbooks for standardisation of data collection procedures. Two-stage probability proportional to size cluster sampling was chosen as the sampling methodology. Vaccination stations and children vaccinated at each station were randomly selected in the first and second stages of sampling, respectively. Sample size took into account the prevalence of EBF in Brazilian capitals and the Brazilian Federal District as reported by the 1st Survey of Breastfeeding Prevalence (1999).
The survey instrument contained closed-ended questions on intake of breast milk, other milk types and other foods (including water, tea and other fluids) on the day prior to administration of the survey. The choice of a ‘current status’ approach sought to minimise the possibility of recall bias. In accordance with WHO recommendations, infants were considered exclusively breast fed when they received nothing but breast milk—only Oral Rehydratation Solucion (ORS), drops, syrups (vitamins, minerals, medicines) and nothing else.16 Data were also collected on breast feeding in the first hour of life and on foods provided to the newborn on the first day at home after hospital discharge (in order to mitigate the possibility of recall bias, this information was obtained only for infants under the age of 4 months). Information on place of birth was collected by means of an open-ended question; respondents were asked to provide the name of the facility in which the child had been born. Responses were then checked against a list of baby-friendly facilities and classified by BFHI status (Yes/No). Data were also collected on the infants (gender, age, birth weight, delivery method and pacifier use) and their mothers (educational attainment, age, parity and employment status).
The instrument was administered by trained interviewers to the guardians of all infants under the age of 1 year throughout the second stage of the 2008 immunisation campaign. Data were entered by municipal workers into an online application developed for this specific purpose. The resulting database was exported into the SPSS V.16.0 (SPSS) and Stata V.10.0 (StataCorp LP) software packages.
Specific procedures for analysis of data from complex probability-sample surveys were used to obtain descriptive statistics.
As we worked with current status data, instead of asking how long the child was exclusively breast fed, we obtained data on infant feeding in the 24 h preceding the survey. For this reason, in the statistical analysis, we adopted probit regression, which allows us to obtain estimates of breast feeding at different ages and their median duration.
Bivariate analysis was performed between each independent variable and each outcome of interest, with ‘born in BFH’ (yes/no) as the main independent variable and the following exposure variables: ‘gender’ (male/female), ‘infant age’ (0–2, 3–5, 6–8 and 9–12 months), ‘low birth weight’ (yes/no), ‘mode of delivery’ (vaginal, forceps assisted/Cesarean), ‘pacifier use’ (yes/no), ‘firstborn’ (yes/no), ‘maternal age’ (<20 years/≥20 years), ‘educational attainment’ (<8 years/≥8 years) and ‘maternal employment status’ (not employed outside the home/employed outside the home but on maternity leave at the time of interview).
The outcomes analysed were breast feeding in the first hour of life in infants under 1 year of age; EBF on the first day after hospital discharge in infants under 4 months of age; EBF in infants under 2, 3 and 6 months of age; and pacifier use in infants under 6 months of age. The influence of birth in a BFH on the various outcomes was analysed by means of Poisson regression with robust variance for complex samples, as recommended for cross-sectional data with non-rare outcomes.17 The independent effect of birth in a BFH was adjusted for the various exposure variables that, on bivariate analysis, were associated with each outcome at a significance level of ≤20% on χ2 testing.
The study design was approved by the São Paulo Institute of Health Research Ethics Committee (protocol number 001/2008, 6 May 2008) after discussion with the National Bioethics Commission of Brazil (CONEP), which agreed about application of an oral consent for mother and care takers, after explaining the objectives of the survey, possibility of refusing or interrupting the interview, confidentiality of information and after giving them written instructions about contacting the research coordinator.
The study sample comprised 65 936 infants under the age of 1 year (35 941 under the age of 6 months), and responses were obtained for 91% of population included in the survey. Male gender was slightly predominant, as was Caesarian birth and non-firstborn status. The prevalence of low birth weight was 8.2%, and less than one-third of deliveries took place in BFHs. Most mothers had ≥8 years of formal education and were either on maternity leave or not employed outside the home at the time of the survey (table 1).
Analysis of the outcomes of interest showed that little over two-thirds of infants were breast fed in the first hour of life; most infants under the age of 4 months were exclusively breast fed on the first day at home after hospital discharge and 63.7% of those under 60 days, 59.6% of those under 90 days and 46.3% of those under 180 days were exclusively breast fed.
The median duration of EBF among infants born in BFHs was 60.2 days (95 CI% 56.5 to 64.2) versus 48.1 days (95% CI 45.3 to 50.8) for those born elsewhere.
Bivariate analyses for each of the study outcomes are shown in tables 2 and 3 shows the independent effect of birth at a BFH on the various study end points, adjusted for the other variables included in multivariate analysis. Infants born in the BFHI-accredited facilities were 9% more likely to be breast fed in the first hour of life (prevalence ratio (PR) 1.09, 95% CI 1.06 to 1.11) and 6% more likely to be breast fed in the first day at home after hospital discharge (PR 1.06, 95% CI 1.04 to 1.09). Those under 2 months of age were 13% more likely to receive EBF when born in BFHs (PR 1.13, 95% CI 1.07 to 1.20) and 8% (PR 1.08, 95% CI 1.03 to 1.13) and 6% (PR 1.06, 95% CI 1.01 to 1.11) more likely to receive EBF at ages <3 and <6 months, respectively. Birth at a BFH also provided significant less pacifier use.
The present study showed that the BFHI had a positive effect on breastfeeding indicators in a probability sample of infants under the age of 1 year from 64 Brazilian cities.
The finding of a higher prevalence of infants breast fed in the first hour of life, as recommended by ‘Step Four’, at the BFHI-accredited maternity hospitals shows that these facilities are fully compliant with this initiative recommendation. Furthermore, in light of the evidence suggesting positive impact of breast feeding in the first hour of life on child health, one may infer that the BFHI is helping to decrease child mortality in Brazil.18 ,19
The median duration of EBF in infants born at BFHs (60.2 days; 95% CI 56.5 to 64.2) was statistically greater than in infants born elsewhere (48.1 days; 95% CI 45.3 to 50.8) and was also longer than the overall duration of EBF in infants born in all Brazilian capitals and the Federal District in 2008 (54.1 days; 95% CI 50.3 to 57.7).8
Some Brazilian studies have shown that the BFHI has a positive effect on breastfeeding practices. Venancio et al20 analysed data for 11 481 infants under the age of 4 months from 84 municipalities in the state of São Paulo, which conducted studies during the immunisation campaigns, and found that infants born in non-baby-friendly maternity hospitals were at greater risk of EBF discontinuation than those born in BFHs.
In a before-and-after observational study conducted at a university hospital in Southern Brazil, Braun et al14 found that infants born at the hospital before it received the BFHI accreditation were at a higher risk of discontinuing EBF in the first month of life and discontinuing breast feeding before the age of 4 months.
Caldeira and Gonçalves13 in a comparative analysis of breastfeeding indicators reported by two cross-sectional studies enrolling randomly selected children under the age of 2, before and after implementation of the BFHI in the public maternity care network of a Brazilian city, concluded that implementing the initiative across all public maternity hospitals led to a significant increase in breastfeeding rates in the municipality.
Using a quasi-experimental nested design in a cohort of infants from Southern Brazil, Silva et al21 found that infants born in BFHs had higher EBF rates at 1 month of age.
Studies conducted in other countries have also showed beneficial effects of the BFHI implementation. In a randomised controlled trial carried out in Belarus, Kramer et al22 found that Ten Steps to Successful Breastfeeding-based interventions increased BF and EBF rates.
Two Swiss studies reported a positive impact of the BFHI on breastfeeding rates. Philip et al23 found significantly longer durations of exclusive and non-EBF when delivery occurred at a BFH featuring a high level of compliance to Unicef guidelines. Merten et al24 ascribed the increased prevalence of EBF in the country since 1994 to a greater number of accredited baby-friendly facilities.
Abrahams and Labbok25 assessed the influence of the BFHI in 14 countries and concluded that implementation of the Initiative was associated with a significant increase in yearly EBF rates among infants under the ages of 2 and 6 months.
This study also showed that the BFHI in Brazil decreased the odds of pacifier use by infants in the sample. As many studies have revealed a negative impact of pacifier use on breast feeding,26–28 this finding could be one of the mechanisms whereby the BFHI has had a positive impact on EBF rates.
The favourable results of this study suggest that the BFHI has been playing a role in increasing breastfeeding and EBF rates in Brazil. Studies conducted in Brazil during the immunisation campaigns in 1999 and 2008, using the same methodology procedures, showed that the median duration of EBF increased from 23.4 to 54.1 days, and, in the same period, a substantial number of maternity care facilities (255) received the BFHI accreditation in the country.12
Analysis of the influence of the BFHI on EBF in different age ranges (infants under the ages of 2, 3 and 6 months) showed that the positive impact of the initiative on EBF prevalence declines over time, providing further evidence of the need for constant support to encourage women to maintain EBF until the sixth month of infant life.
We believe that the BFHI could have an even greater impact if all accredited hospitals were fully compliant with the 10 steps. Unfortunately, during the accredited hospital reassessment process, of the facilities reassessed in the past 3 years (278/335), fewer than half were found to follow all 10 steps, which may have interfered with evaluation of the impact of the initiative on breast feeding. This led the Brazilian Ministry of Health to establish a strategy for monitoring BFHs, launched in 2010, in which all hospitals with the BFHI status must conduct a yearly self-assessment, feed data into an online application and plan all improvements required for full compliance with the 10 steps.
We hope that the results of this study will make policy makers and health professionals aware of the importance and potential of this strategy and urge them to strengthen the BFHI, improving existing facilities and increasing the number of accredited hospitals.
What is already known on this subject
Several strategies have been implemented with a view to protect, promote and support breast feeding. One notable such strategy is the Baby-Friendly Hospital Initiative, launched by WHO and the Unicef in 1991.
What this study adds
The present study showed that the Baby-Friendly Hospital Initiative had a protective effect on exclusive breast feeding in a large probability sample of children under the age of 1 year from 64 municipalities, including all Brazilian capitals.
Our thanks to teams of states and municipalities who participated in the II Prevalence of Breastfeeding in Brazilian Capitals and Federal District.
Funding This study was supported by Brazilian Ministry of Health.
Competing interests None.
Ethics approval Ethics approval was provided by São Paulo Institute of Health Research Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.