Predictors of breastfeeding self‐efficacy among women attending an urban postnatal clinic, Uganda

Abstract Aim The aim of this study was to determine the factors associated with breastfeeding self‐efficacy among postnatal women in Kampala, Uganda. Methods This was a descriptive cross‐sectional study that was conducted among women attending a postnatal clinic at a teaching hospital in Kampala. Three hundred and eighty‐four postnatal women were randomly selected to respond to an interviewer‐administered questionnaire. We used the Breastfeeding Self‐Efficacy scale (BFSES) to assesses breastfeeding self‐efficacy (BFSE). Descriptive statistics and percentages were used to summarize the findings. Bivariate and multivariate logistic regressions were used to determine predictors of BFSE. Results Participants had a mean BFSE score of 48.65. The 14 item BFSES consistently measured breastfeeding confidence with a Cronbach's alpha of 0.89. About six in 10 women (60.2%) had high BFSE, the rest (39.8%) had low BFSE. Having a partner (adjusted odds ratio (aOR): 13, 95% CI 3.46–15) and receiving breastfeeding support from health workers (aOR: 4.45, 95% CI: 1.95–6.12) were significantly associated with BFSE. Conclusion A notable number of mothers had a low BFSE. Health workers should support breastfeeding mothers to achieve the desired exclusive breastfeeding levels. Relevance to clinical practice The findings of the study provide a direction for midwives in maternity care in educating and supporting women about breastfeeding for the improvement of exclusive breastfeeding rates thus realization of benefits of exclusive breastfeeding.


| Study design and setting
This cross-sectional study was conducted to determine factors associated with BFSE of postnatal mothers. The study was carried out on postnatal units in a teaching hospital in Kampala, Uganda. The facility is one of the largest hospitals in the country offering maternal and child health services including antenatal, postnatal and special care for new born babies. The hospital has an official bed capacity of 1,790 beds. The postnatal section of this Hospital admits an average 100 mothers per day.

| Study population and sample size
The study population included women who had delivered either normally or by caesarean birth. The study included women who voluntarily consented to participate in the study. Women who had birth complications and those who were too sick to participate in the study were excluded. We used Epi-info version 7 to calculate the samples size based on a consideration of a two-tailed significance level, an (alpha) of 0.05, an expected frequency of low BFSE of 50% and a 95% confidence level. Based on these considerations, we enrolled a total of 384 postnatal mothers.

| Breastfeeding self-efficacy
We conceptualized BFSE as the mother's confidence in her ability to breastfeed her baby. This conceptualization was based on the theorization of Dennis in his BFSE theory (Dennis, 1999). Based on this theory, Dennis and Faux developed the BFSE (Dennis & Faux, 1999) which was tested among 130 Canadian women. The original 43 item version of BFSE has since been reduced to 14 items through rigorous methodological studies and found to be valid and reliable in China (Dai and Dennis, 2003), in Poland (Wutke & Dennis, 2007) and in Australia (Creedy et al., 2003). In the present study, BFSE was operationalized as a summated score on the 14-item version. The scale is anchored on a 5-point liker scale where 1 indicates not at all confident and 5 indicates very confident. The total score on this scale ranges from 14-70 and higher scores correspond to higher levels of breastfeeding confidence. The internal consistency of the BFSES in this study was 0.89, with a Median score of 52. Participants with scores equal to or more than 50 were considered to have higher BFSE, and those with scores below 50 were considered to have low BFSE (Wutke & Dennis, 2007. In addition, we solicited information about; age, marital status, tribe, level of education, employment status, average monthly income, parity, social support, previous breastfeeding experience and health facility related variables form each participant.

| Sampling and data collection procedures
We randomly selected the participants from the postnatal clinic by writing numbers corresponding to total number of mothers admitted on the ward on pieces of paper. The pieces of paper were then placed in the containers and mixed thoroughly. Twenty pieces of paper were then picked without replacement; mothers corresponding to the picked pieces of paper were then approached and asked to participate in the study. Mother who accepted to participate in the study were then provided with detailed information, those who consented were then asked to respond to questions about their breastfeeding practices, social demographics and BFSE. Data collection was conducted in one of the rooms at the ward by two trained research assistants.

| Data management and analysis
We conducted data analysis using SPSS version 23 statistical software. The levels of BFSE were computed by summing the 14 items of the BFSES. As proposed by (Wutke & Dennis, 2007), we categorized mothers with scores of 50 or more as having high BFSE and those with scores <50 as having low BFSE. We then conducted bivariate analysis and generated crude odds ratios with corresponding 95% confidence intervals (CI). Following bivariate analysis, variables with p-values of 0.2 or more and those that we deemed plausible were considered for the multivariate binary logistic regression model. The significance level was set at 0.05 and all p-values were two tailed.

| Ethical considerations
Research Ethics committee approval was sought and secured from

Makerere University School of Health Sciences Ethics Review
Committee (SHSREC REF 2015-047) and administrative approval from the hospital. Permission was also obtained from the postnatal unit head nurses. Written informed consent was also obtained from the study participants after provision of information about the study. Participants were assured of confidentiality and privacy by assigning them with numbers instead of names, and it was indeed maintained throughout the study. Participants had the liberty to withdraw from the study at any time, and this did not affect their access to the services at the clinic.

| Characteristics of the study participants
The mean age of the participants was 26.3 (SD: 6.16) years with a range of 16-42 years. Nearly one half of the participants 183 (47.7%) had obtained secondary education. Slightly more than one half of the participants 207 (53.9) had delivered their babies by spontaneous vaginal delivery. The mean age of the babies was 2 (SD: 1.26) days. Most of the mothers 245 (63.8%) had delivered baby boys. The mean weight of the babies was 3.29 ± 0.72 kg (Table 1).

| Breastfeeding self-efficacy of the postnatal mothers
The BFSE of the postnatal mothers was determined using the 14item short form of the BFSES. The items of this scale in the present study were highly consistent with a Cronbach's α coefficient of 0.89.
The item mean scores ranged from 3.12 for item number two "I can breastfeed my baby without formula" to 3.93 for item number one "I can determine if my baby is getting enough milk." The mean scores and standard deviations for the other items are presented in Table 2.
Overall, the mean BFSES of postnatal mothers was 48.65 ± 8.36 with a median score of 52. The scores for the other items are presented in Table 2. Based on the cut-off score of 50 which was proposed by Wutke and Dennis (2007), most postnatal mothers 231 (60.2%) had high BFSE and the rest 153 (39.8%) had low BFSE.

| D ISCUSS I ON
The study set out to determine the predictors of BFSE using the 14-item short form of self-efficacy scale. The scale was developed to measure the mother's confidence in her ability to breastfeed her baby. This was established among women at the postpartum unit before discharge after delivery. The BFSE determined how much effort mothers place on breastfeeding, how she will respond to any challenges and how long she will persevere breastfeeding in the face of obstacles. In this study the Cronbach's alpha of the BFSES is (0.89), this implies that the scale items reliably measured the intended factor which was the BFSE. From the studies that have been done in western countries like Iran, Australia, China and Turkey, the recent publications show strong Cronbach's alpha scores ranging from (0.83-0.93) (Oliver-Roig et al., 2012;Otsuka et al., 2014;Zhu et al., 2014). The overall mean score of the participants on the scale was 48.65. The BFSE levels of mothers are lower compared with studies done elsewhere. In a study that was carried out among primiparas mothers in Australia that showed the mean score to be 58.32 (Keemer, 2013). Similarly, in another study conducted in Turkey among pregnant and postnatal women the median scale score was 55.8 (Tokat et al., 2010). The difference in the scores could be speculated that the Uganda mothers do not receive adequate information regarding breastfeeding from healthcare providers. Most of the mothers (61.7%) reported not to have received any teaching or demonstration on breastfeeding. In this regard, when mothers are not supported by health professionals to breastfeed, their BFSE will be low. This implies that they will not breastfeed their babies for the recommended time and the benefits from such will not be maximized.
On the other hand, this study found out that 60.2% of the women were above the cut-off of the BFSES. The results of this study indicated that a greater number of women had a high BFSE based on the cut-off. This implies that a greater number of women are confident in breastfeeding their babies. On the other hand, a statistically significant number of women (39.8%) had a low BFSE. This may not be surprising given the fact that women who do not exclusive breastfeed their children in Uganda is approximated to 40%. This implies that such women cannot cope with the challenges hence giving up on exclusive breast feeding. On an important note, identification of such mothers allows healthcare professionals to provide appropriate intervention to modify their attitudes and enhance their confidence to breastfeed (Zhu et al., 2014).

| Predictors of breastfeeding self-efficacy
At the bivariate level, maternal BFSE was associated with the mode of delivery, the relationship with the partner or presence of a spouse, employment status and breastfeeding social support.