The effect of a combined intervention on exclusive breastfeeding in primiparas: A randomised controlled trial

Abstract An antenatal/postnatal intervention involving proactive telephone support and written materials was conducted among primiparas. Four hundred women, from the Split‐Dalmatia County, Croatia, were randomized between November 2013 and December 2016 into three groups: intervention (IG), active control (ACG) and standard care (SCG). Primary outcome was exclusive breastfeeding (EBF) at 3 months. Secondary outcomes included breastfeeding difficulties, attitudes towards infant feeding, breastfeeding self‐efficacy and social support. Practice staff were blinded to group allocation. Of 400 women, 45 (11%) were lost to follow‐up, and final analyses were conducted on 129 (IG), 103 (ACG) and 123 (SCG) participants. EBF rates at 3 months were significantly higher for the IG (odds ratio [OR] 4.6, 95% confidence interval [CI], 2.7 to 8.1; EBF 81%) as well as at 6 months (OR 15.7, 95% CI, 9.1 to 27.1; EBF 64%) compared with SCG (EBF 47% at 3 months and 3% at 6 months). Higher rates were also observed for the ACG at 3 months (OR 2.2, 95% CI, 1.3 to 3.8, EBF 68%) and 6 months (OR 2.3, 95% CI, 1.4 to 3.9, EBF 16%). Participants in the IG had the highest increase in positive attitudes towards infant feeding, in comparison to baseline, and significantly higher breastfeeding self‐efficacy. Participants in SCG experienced significantly more breastfeeding difficulties, both at 3 and 6 months, in comparison to AC and IGs. Written breastfeeding materials and proactive telephone support among primiparas are an effective means of increasing breastfeeding rates, decreasing breastfeeding difficulties and improving self‐efficacy and attitudes towards infant feeding.


| INTRODUCTION
Breastfeeding is associated with numerous established health benefits for both the infant and mother  as well as considerable potential savings to health services . These outcomes are even greater for infants that are exclusively breastfed.
Hence, the World Health Organisation (WHO) recommends exclusive breastfeeding (EBF) for the first 6 months of an infants' life followed by ongoing breastfeeding, along with timely and appropriate complementary foods, for at least 2 years (World Health Organisation [WHO], 2001).
Despite this, EBF rates in Croatia are far from recommended.
A cohort study conducted in 2011 found that 96% of women initiated breastfeeding, of which 39% were EBF at 3 months, and only 8% were still practicing this form of infant nutrition at 6 months Cochrane review of telephone support for postpartum women (Lavender, Richens, Milan, Smyth, & Dowswell, 2013) identified eight intervention trials aiming at improving breastfeeding outcomes at 6 weeks, 3 and 6 months. Results were inconsistent for any or EBF at 6 weeks, whereas at 3 and 6 months, women receiving telephone support were more likely to be EBF.
Based on these findings, we aimed to test (a) the effect of an educational intervention in the form of a breastfeeding booklet distributed during pregnancy and (b) the effect of four proactive telephone calls provided by a health professional during the prenatal and postnatal period, on EBF rates at 3 and 6 months among first time mothers in a setting with high initiation rates. In addition, we assessed the effect of these interventions on "any breastfeeding" rates and breastfeeding difficulties at 3 and 6 months, as well as breastfeeding self-efficacy, infant feeding attitudes and social support at 3 months.

| Trial design
This was a single-centre, controlled, randomized, three-arm, superiority study, with blind-outcome assessment.

| Setting
The

| Study participants' eligibility criteria
The study population was primigravidae, with a singleton pregnancy, who attended their primary care obstetrician between 20 to 32 weeks of pregnancy. In addition, they were required to speak Croatian and reside within the territory of the Republic of Croatia for at least a year. Those who were unable to communicate in Croatian by phone, planning to leave the country within a year or had a severe medical or psychiatric problem that could be aggravated by participating in the study, were excluded.

| Interventions
Patients were randomly assigned to one of three groups: • Intervention group (IG)received a breastfeeding booklet and a general, pregnancy booklet, followed by four proactive telephone callsone in pregnancy and three after delivery, at 2, 6 and 10 weeks.

Key messages
The provision of a combined antenatal and postnatal intervention, involving proactive telephone support and a breastfeeding booklet, among first time mothers: • increases exclusive breastfeeding at 3 and 6 months • increases breastfeeding self-efficacy • improves attitudes towards infant feeding • decreases the occurrence of breastfeeding difficulties • Active control group (ACG)received a general, pregnancy booklet, followed by four proactive telephone callsone in pregnancy and three after delivery, at 2, 6 and 10 weeks.
• Standard care group (SCG)received standard care, that is, did not receive any written materials or phone calls before or after birth.  (Michie et al., 2013). All interventions were conducted by DP, a registered nurse with 15 years of clinical experience, of which 2 years were spent working in a primary care obstetric practice, and who completed additional breastfeeding training.

| Outcome measures
The primary outcome of this study was the proportion of mothers EBF at 3 months, measured using a postal, self-completed infant feeding survey. In addition, we measured the proportion of EBF infants at 6 months, so as to enable comparison with other studies. Secondary outcomes included any breastfeeding, prevalence of childhood illnesses (mother-reported), prevalence of breastfeeding difficulties and whether help was sought, maternal BMI, infant weight gain and reasons for stopping breastfeeding at 3 and 6 months, whereas attitudes towards infant feeding, breastfeeding self-efficacy and social support were measured at 3 months only. In our study, "EBF" refers to infants who received breast milk only, as well as those who received water in addition to breast milk (known as "predominant breastfeeding," EU, 2008). We combined them into one group given the known similar clinical outcomes for these infants and the small sample size of the latter group in our study (n=11). Disaggregated data are available in

| Randomization and blinding
Two obstetric practices were added atop those planned in the protocol, to aid recruitment, bringing the total to eight. Practice staff recruited eligible women to the study and, once consented, forwarded their details to the lead investigator who randomized each participant to one of three arms of the study, using a computer random number list pregenerated by a member of the research team. Stratified randomization was performed based on known predictors of EBF, namely, smoking status (no, yes or stopped during pregnancy) and educational level (elementary school, high school or college), to ensure balance in each group. Practice staff were blinded to group allocation, and study participants were not told of the existence and differences between the study groups; they were only informed that a study on infant feeding was being held.

| Statistical analysis
Intention to treat analysis was conducted using the "worst-case" scenario: all initially randomized participants in the IG were considered to have stopped breastfeeding, and those in the active control and SCGs as if they continued with EBF. As the primary outcome analysis (breastfeeding rates and associated odds ratios) in intention to treat analysis showed the same direction of results as analysis based only on the patients that were fully followed (Table A1), all primary and secondary analyses below are reported on the participants that were fully followed. Additionally, although we did not per protocol plan to collect data on predominant breastfeeding, participants provided information for it, so we also present those results in Table A2.
Main outcome data (the rates of breastfeeding), as well as participants' nominal data, are reported using the number and percentages calculated based on total number of participants in each group, and the differences between the groups tested using chi-square tests and ordinal regression for the primary outcome (standard group serving as a comparison). Numerical outcomes of secondary outcomes and participant characteristics due to their nonnormal distribution were reported as medians and interquartile ranges, and group differences were tested with Kruskall-Wallis test. Possible association of participants' characteristics and the type of breastfeeding at 3 and 6 months was tested using ordinal regression analyses. For the regression analyses categorical variables were dichotomized wherever possible (exception being education which we grouped into primary or secondary school, college and university).

| Evaluation of telephone records
As per protocol, a 10% (n = 40) random sample of phone calls was selected and assessed for fidelity by a trained psychologist, independent from the research team. The lead investigator was found to consistently adhere to the study protocol, applying a tailored personcentred approach.

| RESULTS
Of the 518 eligible women approached during their antenatal visit, 118 declined to participate, with the rest randomized to one of three groups. One-hundred and thirty-six women were allocated to the IG, 128 to the ACG and 136 to the SCG. All participants who underwent randomization (n = 400) received the intervention as allocated.

No. of household members
In the first 3 months postpartum most participants relied on breastfeeding help from community nurses (n=241, 68%), their relatives (n=237, 67%) and friends (n=171, 48%, Table A3), with their relatives and friends continuing to be the biggest help in the 3-to 6-month period (Table A4).

| DISCUSSION
This study shows that written breastfeeding information and telephone support during pregnancy and at 2, 6 and 10 weeks postpartum leads to a significant increase in EBF rates among first time mothers.
Additionally, mothers who received nonbreastfeeding-focused written information and telephone support showed a slight increase in EBF rates compared with the SCG, which is in line with McFadden's systematic review finding of any support leading to increased breastfeeding rates. In addition, participants who received the breastfeeding-focused intervention had a significantly more positive attitude toward infant feeding and higher breastfeeding self-efficacy at three months postpartum, compared with control groups, as well as significantly fewer breastfeeding difficulties during the first 6-month period.
Our positive results may have been achieved partly due to the type of written materials provided to participants during the study.
Firstly, plain language was used in all written materials and pretested by a pilot sample of 40 expectant women. Secondly, the pregnancy booklet was written by a Croatian parenting organization, that is, for parents, by parents. In addition, the breastfeeding booklet used evidence-based information from the chapter "Promoting breastfeeding during pregnancy," part of the UNICEF/WHO 20-hr course for maternity staff. Behaviour change technique intervention  components, such as "health consequences," "social and environmental consequences" and "emotional consequences" were also highlighted as part of the antenatal written information (Michie et al., 2013) An easy-to-read A4 format was chosen, with plenty of illustra- respectively. These figures should be interpreted with caution given that intended duration of EBF was assessed as "for at least 4-6 months," which is not the same as intended duration of EBF for 6 months. Still, the discrepancy between intention and realization is great, prompting us to ask ourselves why this occurred and how can we better support women to realize their breastfeeding goals.
Breastfeeding is not a single woman's taskit is a collective responsibility, in which the whole community plays an important role.
Obstetricians play a key role, given their ready access to expectant mothers and influence. They are ideally suited to provide educational materials to expectant couples, including conflict-free information on breastfeeding, and yet despite this, expectant mothers in the Split-Dalmatia County did not routinely receive any educational materials.
We did not directly ask participants whether they were satisfied with the proactive breastfeeding support they received, but indirect comments made by participants were all positive. As all of our phone support was conducted by one person, interventions aiming to replicate this support may be influenced by the skill and warmth of the persons who conduct them. It can be assumed that the research nurse was highly motivated to provide phone support to participating mothers, which is less likely to be achieved in real life (i.e., having all professionals equally motivated for providing support and ensuring continuity of support). Hence, further studies on the intervention are needed to evaluate outcomes when support is provided by other health professionals. Although we did not recruit all of the patients we planned according to our protocol, due to time constraints of the main investigator, our sample size calculation was based on data which greatly differed from the findings we observed in our study. This is most likely due to the 10-year time lapse during which breastfeeding awareness has been influenced by numerous promotional campaigns. Our final differences were greater than predicted, allowing our sample to have sufficient power to demonstrate the effectiveness of the intervention. Initially, we also planned to collect infant birth demographics, but it proved more difficult to do so, and hence, these were omitted. Despite these limitations, methodological rigour was adhered to, resulting in comparable groups at baseline, minimal attrition, avoidance of the Hawthorne effect and fidelity in implementation of planned intervention. In addition, participants were followed-up for a relatively long period of time, during which woman-centred care was consistently provided in a pro-active manner.

| CONCLUSIONS
Written breastfeeding materials and proactive telephone support among first time mothers are an effective means of increasing breastfeeding rates, decreasing breastfeeding difficulties and improving breastfeeding self-efficacy and attitudes towards infant feeding during the first six months postpartum.