Breastfeeding has many health advantages for both mother and baby. Breast milk has been shown to protect babies against gastrointestinal, urinary, respiratory and middle ear infection (Howie et al., 1990, Marild et al., 1990 and Marild et al., 2004), atopic disease if there is a family history of atopy (Burr et al., 1989, Fewtrell, 2004 and Oddy et al., 1999), and juvenile-onset insulin-dependent diabetes mellitus (Sadauskaite-Kuehne et al 2004). There is also evidence to indicate that breastfed babies have a reduced risk of becoming obese (Arenz et al., 2004 and Fewtrell, 2004). Maternal benefits include reduced risk of premenopausal breast cancer and some forms of ovarian cancer (Department of Health 1994) and a possibility of protection against hip fractures in older age (Department of Health 1998). In addition, breastfeeding provides a unique maternal–infant contact and ready availability of food for the baby.

Despite robust evidence to support the benefits of breastfeeding and international initiatives to support practice (see next section), rates in the UK are among the lowest in Europe. The 2000 Infant Feeding Survey (Hamlyn et al 2002) found that only 62% of women overall in the UK initiated breastfeeding, but noted small increases in incidence across the four countries of the UK. The Infant Feeding Survey 2005 showed an overall initiation rate of 76%, with breastfeeding rates continuing to increase across all four countries, after standardising for maternal age and social class (Bolling et al 2007). Initial breastfeeding rates were 78% in England, 70% in Scotland, 67% in Wales and 63% in Northern Ireland. Although the 2005 survey showed an increase in the prevalence of breastfeeding up to 9 months in England, Wales and Northern Ireland compared with 2000, in line with previous surveys many women reverted to artificial feeding within a few weeks of birth (Bolling et al., 2007, Foster et al., 1997 and Hamlyn et al., 2002), just over half within 6 weeks of the birth in the 2005 survey (Bolling et al 2007). The recommended minimum duration of breastfeeding is at least 6 months to ensure that the infant receives the full protective benefits of breast milk (WHO 2002).

A number of recent policy initiatives support the need to increase the uptake and duration of breastfeeding, including The Priorities and Planning Framework 2003–2006 (Department of Health 2003), the National Service Framework for Children, Young People and Maternity Services (Department of Health 2004) and the NICE postnatal care guideline (NICE 2006). Evidence to support this chapter has also been drawn from a systematic review and evidence into action briefing published by NICE as part of the agency’s work to inform public health (Dyson et al., 2006 and Renfrew et al., 2005). Problems with assessing the evidence in relation to interventions to support the uptake and duration of breastfeeding include variation in timing of follow-up, lack of clarity around definitions of exclusive and mixed breast- and bottlefeeding, and retrospective study designs.

The NICE postnatal care guideline (2006) has recommended that all maternity units in England and Wales introduce a structured training programme to support breastfeeding, using the WHO/UNICEF Baby Friendly Initiative (BFI) as a minimum standard. The BFI is a global initiative which was launched in 1992 to encourage maternity units to support breastfeeding through implementation of the 10 steps described below, and was introduced into the UK in 1994. In order to achieve full accreditation as a BFI hospital, an external review has to be undertaken to assess that all 10 steps have been implemented. Full accreditation lasts for 2 years, following which reassessment of the standards is undertaken over a period of 3 days. Intervals between subsequent assessment will depend on how well standards are being maintained. Units can also apply for a certificate of commitment, which is awarded if there is a breastfeeding policy in place, an action plan to achieve BFI and commitment to implement the plan. Currently, only around 10% of babies in England are born in units which have BFI accreditation, compared with 58% of babies born in Scotland, 46% in Wales and 39% in Northern Ireland (data from BFI website accessed 28 March 2008).

The effect of BFI implementation on breastfeeding rates in Scotland was recently evaluated in an observational study using an annual survey of progress towards achieving BFI status and annual breastfeeding rates (Broadfoot et al 2005). Breastfeeding statistics were collected when the Guthrie screening test was carried out at 7 days postpartum. Data were collected from maternity units which had over 50 births a year between 1995 and 2002 (n = 33), which provided data on 464,246 births. Women giving birth during this period were 28% (odds ratio (OR) 1.28, 95% confidence interval (CI) 1.24–1.31) more likely to be breastfeeding at 7 days if they gave birth in a hospital with the UK BFI award. Results were adjusted for mother’s age, deprivation score, hospital size and year of birth. From 1995 breastfeeding rates increased significantly faster in hospitals with BFI status (11.39% cf 7.97%). Impact on duration was not investigated as part of this study.

A more recent study (Bartington et al 2006) aimed to examine if women were more likely to commence and continue breastfeeding if they gave birth in a BFI-accredited unit in the UK, in a cohort with a high representation of women from disadvantaged and lower socio-economic groups. Women were asked to report if they had commenced breastfeeding and were still breastfeeding at 1 month after birth. Data on 17,359 singleton babies were grouped according to maternity unit BFI participation status at the time of the birth (accredited, certificated or neither award). Women who had given birth in an accredited unit were more likely to commence breastfeeding (adjusted rate ratio (ARR) 1.10, 95% CI 1.05–1.15), but were not more likely to breastfeed at 1 month (0.96, 95% CI 0.84–1.09), after adjusting for social, demographic and obstetric characteristics. Breastfeeding uptake was also independently associated with attendance at antenatal classes (1.14, 95% CI 1.11–1.17), vaginal birth (1.05, 95% CI 1.03–1.08), having a companion at the birth (1.09, 95% CI 1.04–1.1.6) and hospital postnatal stay of longer than 24 hours (1.06, 95% CI 1.04–1.09).

Much of the other evidence to support the implementation of the BFI has come from countries outside the UK, with studies from high- and low-income countries represented. A controlled, non-randomised study undertaken in Italy (Cattaneo & Buzette 2001) reported outcomes on breastfeeding rates before and after staff training for BFI. The study was undertaken in eight hospitals (no Italian hospitals at the time had achieved BFI accreditation). Data on breastfeeding rates were collected at hospital discharge and at 3 and 6 months after the birth. Hospitals were allocated to one of two groups. Following an initial period of assessment in both groups, local facilitators in group 1 were trained following an adapted programme based on BFI, which they then cascaded to the health professionals at their unit. A second assessment phase was undertaken and the intervention replicated in group 2. A third and final period of assessment was undertaken in both groups 5 months after training was completed. Data collection tools included a self-assessment tool to ascertain the extent of implementation of each of the 10 BFI steps, assessment of usual hospital practice at each point in time, a self-administered questionnaire for health professionals, interviews with mothers at discharge and a telephone interview at 3 and 6 months. All hospitals showed an improvement in compliance with each of the 10 steps. Healthcare professionals’ knowledge of breastfeeding improved, although response rates from both groups were low. Following training there was a significant increase in mothers reporting exclusive breastfeeding at discharge, full breastfeeding at 3 months, and any breastfeeding at 6 months, although the proportion of women breastfeeding at these points in time were below recommended rates, and could have reflected a lack of support for breastfeeding in the community.

A cluster randomised controlled trial (RCT) conducted in the Republic of Belarus (PROBIT study) assessed the effects of breastfeeding promotion on breastfeeding duration and exclusivity and rates of infant respiratory and gastrointestinal infection and atopic eczema during a 1-year follow-up (Kramer et al 2001). Thirty one maternity hospitals and polyclinics were randomly assigned to receive an intervention based on the BFI (n = 16) or current care (n = 15) of usual infant practices and policies. Over 17,000 mother–infant dyads were recruited, 16,491 of whom completed the 12 months of follow-up. There was a significant increase in the duration of breastfeeding to 12 months, increased exclusive breastfeeding to 6 months, and a significant reduction in the numbers of infants who had gastrointestinal tract infections and atopic eczema, but no improvement in respiratory outcomes.

DiGirolamo and colleagues (2001) carried out a longitudinal cohort study in the USA to assess the impact of the type and number of BFI practices experienced by mothers on breastfeeding outcomes. Data were collected from the 1993–94 Infant feeding practices survey, a longitudinal survey of pregnant and new mothers up to 12 months after their delivery from across the USA, conducted by the United States Food and Drug Administration. The main outcome measure was breastfeeding cessation within 6 weeks of giving birth.

Of 2610 women deemed eligible for the study, 1737 (67%) returned the antenatal and subsequent postnatal questionnaires. There were imbalances in the characteristics of responders who were more likely to be white, over 30 years of age, married and to have higher income levels and educational achievements. The focus of the study was the 1085 women who reported an intention to breastfeed for more than 2 months after their delivery. The impact of five of the 10 BFI steps was assessed: breastfeeding initiation within 1 hour of the birth; feeding only breast milk; rooming-in; breastfeeding on demand; no use of dummies. Seventeen percent of women had given up breastfeeding within 6 weeks. Only 7% of the mothers had experienced all five steps, the strongest risk factors for early cessation being late commencement of a first breastfeed and supplementation of the infant. When compared to women who experienced all five steps, mothers who experienced none were eight times more likely to stop breastfeeding within 6 weeks (OR 7.7, 95% CI 2.3–25.8).

Philipp et al (2001) compared breastfeeding initiation rates at Boston Medical Center, USA, prior to (1995), during (1998) and after (1999) the BFI was implemented. Two hundred complete medical records randomly selected by computer were reviewed for each of the 3 years. Data collected from the records enabled infants to be categorised into four groups: exclusively breastfed; mostly breast milk; mostly formula; and exclusively formula. The maternal and infant demographics for all 3 years were comparable. Full implementation of the 10 steps leading to BFI accreditation resulted in significant increases in breastfeeding initiation rates from 58% in 1995 to 77.5% in 1998 to 86.5% in 1999 (p<0.001), leading the authors to conclude that full implementation of the BFI was an effective strategy to increase breastfeeding initiation rates in the US hospital setting.

The Department of Health has estimated that the NHS could save £35 million each year in the treatment of babies with gastroenteritis alone, if all babies were breastfed, representing a saving of £300,000 for the average health district (Department of Health 1995), a saving which is likely to be significantly higher at today’s costs. The NICE postnatal care guideline included the outcome of an economic model of the likely effects of universal implementation of the BFI in English and Welsh hospitals, over a 15-year period based on breastfeeding rates, and association between breastfeeding and disease occurrence (NICE 2006). This involved estimating the likely cost of universal (and ongoing) accreditation and the effect of increased breastfeeding initiation on infant health outcomes, which included gasteroenteritis, otitis media, asthma and necrotising enterocolitis. The model showed that over a 15-year period, cost savings following implementation would outweigh the net costs of introduction of the BFI into each unit, breaking even after 6 years. An increase in breastfeeding rates attributable to the BFI of 10% over 15 years could bring an average annual saving of £8.27 million, the major part of the savings accruing from the impact of an increase in breastfeeding uptake on reductions in infant gastroenteritis. Further evidence is required of savings which would accrue from an increased duration of breastfeeding.

Breastfeeding support should be seen as an integral part of the role of the midwife providing postnatal care, and all healthcare professionals who have contact with postnatal women should have the appropriate competencies to support women to breastfeed (NICE 2006). The Royal College of Midwives (RCM) publication Successful breastfeeding (2002), which aimed to make breastfeeding advice offered by midwives consistent and evidence based, stressed that this advice should be given equal priority with other aspects of care.

A Cochrane systematic review of interventions to promote the initiation of breastfeeding included data from seven trials involving 1388 women (Dyson et al 2005). A meta-analysis based on five trials involving 582 women on low incomes in the USA showed that breastfeeding education had a significant effect on increasing initiation rates compared to routine care (relative risk (RR) 1.53, 95% CI 1.25–1.88). However, the extent to which this effect would be found based on populations from other countries would have to be considered.

The impact of placing a baby on his/her mother’s chest as soon as possible after birth or within the first 24 hours of the birth (termed ‘skin-to-skin contact’) has been assessed in relation to breastfeeding uptake and duration. A Cochrane review by Anderson et al (2003) included 17 RCTs which provided data on 806 mother–infant dyads. The methodological quality of most of the studies was poor and intervention characteristics, such as duration of skin-to-skin contact, varied greatly across the studies. Data were obtained from diverse populations from a range of countries. Fifteen studies included only women who had given birth vaginally and all but one study included only healthy, full-term babies. More skin-to-skin dyads were still breastfeeding 1–3 months (30–90 days) post birth (OR 2.15, 95% CI 1.10–4.22). The reviewers concluded that early skin-to-skin contact appeared to have some clinical benefit especially regarding breastfeeding outcomes and infant crying and had no apparent short- or long-term negative effects. Study data showed that most of the infants suckled during the skin-to-skin intervention, which may be a critical component of this intervention with regard to long-term breastfeeding success. Although further research is required, it appears women and their infants should not be separated following the birth without an unavoidable medical reason.

A recent RCT undertaken in the north of England, and not included in the above review, also examined the effects of skin-to-skin care on breastfeeding outcomes (Carfoot et al 2005). Two hundred and four women and their infants were randomised to routine care (n = 102) or immediate skin-to-skin care (n = 102). The infants in the usual care group were dried and wrapped in a towel before being handed to their mother or father. In the skin-to-skin group infants were placed prone against their mother’s skin and between her breasts as soon as possible after birth. The primary outcome was success of the first breastfeed, as assessed using the Infant Breast Feeding Assessment Tool (IBFAT). Secondary outcomes included breastfeeding at 4 months and infant body temperature at 1 hour after the birth. There were no statistically significant differences between the two groups on successful first feed or breastfeeding at 4 months, although a higher proportion of babies in the skin-to-skin group had a successful first feed (91% cf 83%) and were partially or exclusively breastfeeding at 4 months (43% cf 40%). The mean infant temperature at 1 hour after birth was also higher in the skin-to-skin group, which was statistically significant (95% CI 0.03–0.28, p = 0.02), although this difference was not considered to be clinically significant. Women who had skin-to-skin contact enjoyed the experience, and most reported they would choose to have skin-to-skin care in the future.

Almost all women who decide to breastfeed will have commenced this whilst in hospital, but most will now be discharged prior to the establishment of lactation, so that midwives in the community have an even greater role to play in breastfeeding advice. Not all babies will be ready to feed within the same time period. Babies have been shown to display a wide range of feeding behaviour following a spontaneous delivery (Henschel and Inch, 1996 and Widstrom et al., 1990).

Interventions during labour may also impact on commencement and early cessation of breastfeeding, although much of the evidence to date comes from small studies, few of which were prospective and which have produced inconclusive results. Data from a UK-wide survey of breastfeeding practices at 6 weeks postpartum found that a longer time to first feed was associated with obstetric intervention, such as induction, caesarean section and the use of pethidine during labour (Rajan 1994). A small study by Ransjö-Arvidson et al (2001) examined the impact of maternal analgesia for labour pain relief on infant behaviour. Pre-feeding infant behaviour during the first 2 hours of life was the focus of the study. A convenience sample of 28 infants whose mothers had uncomplicated pregnancies were videoed to observe for pre-feeding gestures including movement of eyes, hands and mouth, touching the nipple before suckling, rooting and licking movements. Women were grouped according to type of labour analgesia. One group had received mepivacaine via a pudendal block (n = 6), a second group had received pethidine or bupivacaine (n = 12), and a third group received no analgesia during labour (n = 10). A significantly lower proportion of babies whose mothers had received any medication touched the nipple with their hands before suckling (p<0.01), made licking movements (p<0.01) and fed (p<0.01). Small numbers limit the generalisability of the findings.

Volmanen et al (2004), in a qualitative study from Lapland, contacted 164 women to ask about problems with breastfeeding, and mixed breast and formula feeding, a median of 2.4 years after delivery. Ninety nine women returned completed questionnaires. Fifty six of these women received an epidural during labour. Although the study was limited by size and potential recall bias, those who had epidural analgesia in labour reported a higher incidence of partial breastfeeding or formula feeding (67% vs 29%; p = 0.003). The researchers speculated that infants affected by bupivacaine crossing the placental barrier (via epidural) might be less capable of stimulating lactation during the neonatal period but these findings should be treated with caution given the study limitations.

In a small well-designed cohort study, Radzyminski (2003) looked at the effect of ultra low-dose epidural analgesia on breastfeeding among 56 mother–infant dyads. Twenty eight women had epidural analgesia, with pain relief provided by low–dose fentanyl and bupivacaine, and 28 did not use any analgesia. Several methods of evaluation were employed to assess drug effects on infant feeding up to 24 hours post delivery, including a preterm infant breastfeeding behaviour scale and a neurobehaviour scale. No significant difference was demonstrated between breastfeeding behaviours of babies born to mothers who had an epidural and those who had no method of analgesia.

A prospective study from France, which also included small numbers, examined predictive factors associated with cessation of breastfeeding whilst on the postnatal ward and up to 4 months after the birth among women who gave birth in what the researchers referred to as a ‘high-tech’ maternity unit (the definition of this was not given) (Lathouwer et al 2004). Breastfeeding status was ascertained on the day of hospital discharge and at 1, 2, 3 and 4 months after birth. Data were collected on 115 women who gave birth during the first 2 months of 2002 and commenced breastfeeding (only 57% of 203 women who gave birth during the study period). At the time of hospital discharge which took place 4 days after the birth, six of the 115 (5.2%) women had given up breastfeeding. During these 4 days, 61% of breastfed infants received at least one supplementary feed with artificial baby milk. Cessation of breastfeeding occurred rapidly following discharge, particularly after 3 months. Of the 109 (95%) women still breastfeeding (which included exclusive and combined breast and artificial milk) on the day of discharge from hospital, 78% were still breastfeeding at 1 month, 66% at 2 months, 44% at 3 months and 17% at 4 months. When data for ‘currently exclusively breastfeeding’ were separated, only 69% were exclusively breastfeeding at 1 month, 47% at 2 months, 17% at 3 months and 6% at 4 months. The only factor associated with early cessation of breastfeeding during the inpatient period was defined as ‘lack of experience’ in multiparous women who decided to breastfeed for the first time (OR 35.33). There were no differences in breastfeeding outcomes by mode of delivery. The small numbers and absence of data on other obstetric, anaesthetic, demographic or maternal characteristics or levels of breastfeeding support during the postnatal period suggest that the results should be interpreted with caution.

A recent prospective cohort study from Australia examined the impact of epidural analgesia during labour on breastfeeding during the first postnatal week and breastfeeding cessation up to the first 24 weeks after the birth (Torvaldsen et al 2006). The study was undertaken in the Australian Capital Territory in 1997. The epidural solution in use at the time was bupivacaine 0.16% with fentanyl 3.3 μg/ml, administered as patient-controlled analgesia. One thousand two hundred and eighty women aged 16 years and over who had given birth to a single live infant were asked to complete questionnaires at 1, 8, 16 and 24 weeks after giving birth. Data were available on 1260 women, 416 (33%) of whom had epidurals. Infant feed data were categorised as either fully breastfeeding, partially breastfeeding or not breastfeeding at all. Following the first survey point, women who had stopped breastfeeding since the previous survey were asked when they stopped. Labour analgesia and mode of birth were associated with partial breastfeeding and problems with breastfeeding in the first postpartum week, a statistically significant finding (p<0.0001). Partial breastfeeding in the first week was associated with three intrapartum factors: use of intrapartum analgesia (p<0.0001), type of birth (p<0.0001) and onset of labour (p = 0.0003). Parity was also significant (p = 0.0006). After adjusting for parity, only epidurals and general anaesthetic were significantly associated with increased risk of partial breastfeeding in the first week; however, when analyses were restricted only to women who had vaginal births, this association was weaker and, after adjusting for parity, the strength of association between epidurals and partial breastfeeding was no longer statistically significant.

Breastfeeding difficulties in the first week after birth were associated with epidural analgesia, after adjusting for parity (OR 2.04, 95% CI 1.39–3.00), an association which remained significant when analyses were restricted to women who had vaginal births (OR 1.75, 95% CI 1.13–2.70). Breastfeeding cessation in the first 24 weeks was associated with analgesia, maternal age and level of education (p < 0.0001). Women who had received an epidural analgesia were more likely to stop breastfeeding than women who had used non-pharmacological methods of pain relief (adjusted hazard ratio 2.02, 95% CI 1.53–2.67). The fact that none of the intrapartum factors was associated with not breastfeeding at all highlights that breastfeeding uptake is more likely to be related to social factors.

The 2000 Infant feeding survey (Hamlyn et al 2002) found that 12% of women who commenced breastfeeding gave up whilst still on the postnatal ward. It also found that being given artificial milk whilst in hospital was associated with stopping breastfeeding within the first 2 weeks; 40% of mothers whose babies had been given a bottle in hospital stopped breastfeeding, compared with 13% of women whose babies had not been given a bottle (Hamlyn et al 2002). In the 2005 survey, a third of breastfed babies had received additional feeds of formula, water or glucose whilst in hospital (Bolling et al 2007).

Renfrew et al (2005), in a systematic review of the effectiveness of public health interventions to promote the duration of breastfeeding, reviewed five RCTs and two non-RCTs and concluded that supplements in the neonatal period should be given only when there are sound medical indications, which supports the WHO/UNICEF recommendation included in the BFI 10 steps. If supplements are given for medical reasons or mothers’ choice the duration of breastfeeding is less likely to be affected if only a small number of supplements are given in the first 5 days. The reasons for introducing supplementation were examined in an ethnographic study undertaken at one maternity unit in England to explore women’s and healthcare professionals’ beliefs, expectations and experiences in relation to supplementation of breastfeeding in the postnatal ward and neonatal unit (Cloherty et al 2004). Mothers who planned to breastfeed but were supplementing their babies with formula feed or expressed breast milk were invited to take part. One of the themes generated related to the conflict of the midwives’ role to support breastfeeding with their desire to protect women from tiredness or distress, a theme confirmed by the mothers’ accounts of how they felt. In some cases, the midwives suggested supplementation whilst in other cases it was requested by the mother. The importance of healthcare professionals being aware of other solutions to assist a breastfeeding mother who is tired and distressed should clearly be a priority (see Chapter 8 on Fatigue).

The length of stay on the postnatal ward has reduced considerably in the UK as a consequence of a number of drivers, including cost containment (Renfrew et al 2005). The impact of reduced inpatient care on maternal and infant health outcomes and feeding practices has been considered in several studies, although changes in service provision have tended to only include breastfeeding as a secondary outcome. The effect on the duration of breastfeeding was evaluated by Brown & Lumley (1997), in a study from Australia. The researchers assessed the impact of early discharge on maternal health outcomes, including duration of breastfeeding in women who gave birth 6–7 months previously at hospitals in Victoria. At 6 weeks, 3 months and 6 months small, non-significant differences were observed in the pattern of infant feeding of women who were discharged within 48 hours and those who stayed 5 or more days after birth. Women who left hospital on day 3 or 4 had significantly lower rates of breastfeeding at 6 weeks (OR 0.58, 95% CI 0.4–0.8), 3 months (OR 0.6, 95% CI 0.5–0.8) and 6 months (OR 0.74, 95% CI 0.6–0.96).

A Cochrane review by Brown and colleagues (2002) considered the safety, impact and effectiveness of a policy of early postnatal discharge for healthy women and their term babies, where ‘early discharge’ referred to discharge that was earlier than standard care in the setting in which the intervention was implemented. The pooled estimate from six trials which reported data on partial or exclusive breastfeeding at 1 or 2 months postpartum suggested no significant difference between women who had early hospital discharge and the control group who received standard care (pooled RR 0.96, 95% CI 0.78–01.18). However, there was significant heterogeneity between the trials included in this analysis (for example, cultural differences in duration of breastfeeding, and different measures used to assess breastfeeding).

A large longitudinal cohort study from Sweden (Waldenström & Aarts 2004) investigated the duration of breastfeeding and number of breastfeeding problems, with a focus on association with the length of postpartum stay. Women attending for their first antenatal clinic appointment were recruited from all clinics in Sweden over 3 weeks evenly spread over a 1-year period. Data on any breastfeeding were collected by postal questionnaire at 2 months and 1 year post delivery. Data on 2709 women (82% of the 3293 who originally agreed to participate) who completed questions on length of stay included in the 2-month questionnaire were presented in this paper. Women were divided into six groups according to length of postnatal stay (day 1 <24 h to day 6 ≥120 h). A number of statistical tests were undertaken to examine the effect of length of stay and potential confounding factors on duration of breastfeeding. Kaplan Meier analysis showed the unadjusted median duration of any breastfeeding was 7 months in women discharged from the postnatal ward on day 1 post delivery and 8 months in women discharged on any of the following days, a non-significant difference (p = 0.66). The difference remained non-significant after adjustment for interrelated maternal characteristics. The authors concluded that maternal characteristics and their experience of their first breastfeed may be more important predictors of breastfeeding duration than length of inpatient care.

Observational studies that have investigated infant feeding have consistently reported that younger women, those from lower socio-economic groups and of younger age when leaving full-time education are less likely to breastfeed (Bick et al., 1998 and Hally et al., 1984). Since most studies have employed large-scale quantitative techniques, subtle social and cultural influences that could affect decisions about infant feeding may not have been identified. A qualitative study of 21 women, which sought to improve understanding of how primiparae belonging to lower income groups had decided to feed their infants, suggested that breastfeeding is a skill which needs to be learnt, particularly by women with little exposure to other breastfeeding women (Hoddinott & Pill 1999). These authors suggest that women may benefit from an antenatal apprenticeship with a known breastfeeding mother. Low rates of breastfeeding over the past two decades mean that many women may not have access within their social network to an effective role model.

A cluster RCT in Mexico City assessed the effects of two different models of home-based peer counselling on the proportion of women exclusively breastfeeding at 3 months (Morrow et al 1999). One hundred and thirty women participated: 44 received six visits, two in pregnancy and at 2, 4 and 8 weeks postpartum; 52 received one visit in pregnancy and two visits at 1 and 2 weeks postpartum; and 34 received usual care (controls). The home visits were made by peer counsellors recruited from within the same community as the women, who had been trained by the La Leche League. Antenatal visits focused on the benefits of exclusive breastfeeding, the importance of positioning, and discussion of solutions to common breastfeeding problems. Postnatal visits focused on maternal concerns, information and social support. Both intervention groups were found to be significantly more likely to be exclusively breastfeeding at 3 months than the controls, which was practised by 28 (67%) of the six-visit group, 25 (50%) of the three-visit group and four (12%) of the control group. These findings were not related to any of the sociodemographic, delivery or health factors examined. However, further research is necessary in other communities since findings may not be applicable across cultures.

Additional support during the postnatal period for women who are breastfeeding does appear to be effective. Renfrew et al (2005) found high-quality evidence from two studies of the benefit of health professional and peer support on both exclusive and any breastfeeding among women from relatively advantaged communities, and evidence from another two studies of a lesser quality, both of which were targeted at African-American women from disadvantaged backgrounds, that support can be effective for any breastfeeding. Postnatal interventions which did not include specific additional breastfeeding support had no beneficial impact, and no evidence was found of effective interventions to support exclusive breastfeeding among disadvantaged women.

Tools to enable women and healthcare professionals to assess progress with breastfeeding have also been evaluated. Hamelin & McLennan (2000), in a study from Canada, assessed the relationship between the use of the LATCH Breastfeeding Charting System by nursing staff during postpartum hospitalisation and breastfeeding outcomes. A post-test control group design with non-random groups collected data from a convenience sample of 180 breastfeeding women who gave birth in an urban perinatal centre. Ninety women received the LATCH tool intervention, which assigned a numerical score of 0, 1 or 2 to five key components of breastfeeding. Six-week postpartum telephone interviews demonstrated no significant differences in breastfeeding outcomes between the two groups in exclusive breastfeeding, although women in the post-test group reported increased confidence in their assessment of infant breastfeeding and when to ask for help with breastfeeding problems. The use of the tool did not reduce the incidence of early breastfeeding problems, particularly women’s concerns about perceived insufficient breast milk, which was reported by a similar proportion in both groups (50% cf 46%).

Two studies which had considered the effectiveness of breastfeeding self-assessment tools were included in the review by Renfrew et al (2005). A study by Pollard (1998) conducted in the USA among women of different socio-economic groups evaluated the effectiveness of a breastfeeding self-monitoring daily log book in addition to a breastfeeding education session, including a video and question and answer session. A significant increase in mean duration of breastfeeding was demonstrated amongst women who adhered to the protocol, although these were likely to be married women, from higher socio-economic groups. They breastfed their infants for three times longer than women who were single who had been uncertain about breastfeeding and had not completed self-monitoring forms. The second study (Loh et al 1997) evaluated use of a simple fact sheet covering eight positive aspects of breastfeeding during the late antenatal period administered by medical students to women attending one antenatal clinic who were randomised to an intervention group and a control group. This resulted in a 38% increase in breastfeeding rates at hospital discharge in the intervention group and a similar increase in breastfeeding rates 4 weeks after the birth but this was not statistically significant, probably due to small numbers in the trial; 193 women were recruited across the two trial arms.

Renfrew and colleagues (2005) concluded that use of education interventions delivered through the use of a self-assessment tool appears to have the potential to increase the duration of breastfeeding among some groups of women, but care would be required to ensure tools were tailored for the intended population. With respect to the second study, further research was required based on larger numbers, with adequate sample size, and consideration given to issues such as method of randomisation and the most appropriate healthcare professional to deliver such an intervention. The review did not find any evidence of effectiveness of the use of written materials alone to support breastfeeding (Renfrew et al 2005).

A high proportion of early breastfeeding problems may be due to incorrect positioning of the baby on the breast. Renfrew et al (2005) considered the evidence from one Australian RCT (n = 75), in which the intervention comprised a group educational session on positioning and attachment of the baby at the breast for primiparous women on low incomes, who stated their intention to breastfeed (Duffy et al 1997). Study outcomes showed an increased duration of exclusive breastfeeding at 6 weeks postpartum, Renfrew and colleagues (2005) concluding that the trial was of high quality which should support the widespread replication of this intervention amongst similar population groups in the UK.

Despite efforts to ensure that appropriate advice and support are given, there are frequent complaints from mothers of inconsistent advice from healthcare professionals about breastfeeding (RCM 2002). It is essential that professionals who have contact with breastfeeding women are regularly updated on breastfeeding matters.

It is likely that almost all postnatal breastfeeding problems can be prevented if the baby is able to feed effectively and efficiently from the beginning (Inch & Fisher 1999). This does not always occur, however, so in addition to having knowledge about how to achieve effective, pain-free breastfeeding, all midwives need to know how to manage breastfeeding problems.