Determinants of High Breastfeeding Self-Efficacy among Nursing Mothers in Najran, Saudi Arabia

Many factors have been found to correlate with satisfactory Exclusive Breastfeeding (EBF) practices. The relationships between EBF practices and associated factors are complex and multidimensional; Breastfeeding Self-Efficacy (BSE) is the most important psychological factor that may help the mother to overcome any expected barriers. This study investigates the determinants of high breastfeeding self-efficacy among Saudi nursing mothers. Methods: This is a descriptive cross-sectional study investigating the determinant of BSE among 1577 nursing mothers in primary health centers in Najran City, Saudi Arabia. The study uses a cluster random sampling technique. Data collection was performed from June 2022 to January 2023 using a self-reported questionnaire that encompasses the Breastfeeding Self-Efficacy Scale—Short Form (BSES-SF), Gender Friendly Breastfeeding Knowledge Scale (GFBKS), Iowa Infant Feeding Attitude Scale (IIFAS), and a basic data questionnaire to assess women’s demographic factors and obstetric history. Results: The mean score for all BSES-SF items was between 3.23–3.41, the highest mean score was in mothers who felt comfortable breastfeeding with family members present (3.41 ± 1.06), and the lowest mean was in mothers who could breastfeed their baby without using formula as a supplement (3.23 ± 0.94). The overall BSE score was high among 67% of the study participants. Binary logistic regression showed that being a housewife, being highly educated, having breastfeeding experience, and being multiparous are positive predictors for high BSE (p ≤ 0.001). In addition, having adequate breastfeeding knowledge and positive breastfeeding attitudes were positively associated with higher BSE (p = 0.000). Conclusion: BSE can be predicted by modifiable predictors such as mothers’ education, working status, parity, breastfeeding experience, adequate breastfeeding knowledge, and positive attitudes toward breastfeeding. If such predictors are considered during breastfeeding-related educational interventions, it could lead to more effective and sustainable effects in community awareness regarding breastfeeding.


Introduction
Breastfeeding is the first and most important step toward healthy infants and communities. The World Health Organization recommends exclusive breastfeeding for the first 6 months of life, followed by the appropriate introduction of complementary foods with continued breastfeeding to two years and beyond [1]. Breastfeeding has evidencebased and well-known short-and long-term benefits for mothers and infants. Breast-milk composition is continually changed from one feed to another to meet the infant's body requirements. Human milk contains easily digested proteins such as glycoproteins, enzymes, and endogenous peptides, which help enhance immunity, cognitive development,

Study Design and Setting
A descriptive cross-sectional study was conducted in four primary healthcare centers in Najran City, Saudi Arabia. Najran City is the largest city and capital of the Najran region. Najran region lies in southwestern the border with Yamen. The Najran population has numerous traditions and health beliefs related to breastfeeding, which may positively or negatively influence infant feeding practices. Najran City contains thirteen major primary health centers affiliated to the Ministry of Health [22].

Study Participants
The inclusion criteria were nursing mothers, with a child aged one day to less than six months, an absence of any breastfeeding contraindications such as HIV, or any condition that may hinder the continuity of EBF or affect milk production such as breast augmentation, lift, or reduction, or nipple surgery, aged 18 years or over, literate, and willing to participate in the study. where n = sample size, df = design effect of cluster sampling, t = the parameter related to the precision of obtaining the largest sample size (1.96 for an error risk of 5%) where the normal curve cuts off an area at the tails (the desired confidence level is 95%), p = expected prevalence of high BSE, q = 1 − p the expected proportion of moderate or low BSE, d = maximum tolerable error (the desired level of precision). The sample size was 1689, after adding 10% to compensate the anticipated sample loss or non-response rate. Najran City contains thirteen primary health centers; the researchers randomly selected 30% of the primary health centers (4 centers). The total sample size of 1689 was divided equally among the four primary health centers (422 participants from each center). In each center, all clinics dealing with children aged from 1 day to less than six months were included in the study. A proportional sample based on the follow-up rate was selected in each clinic, using the convenience sampling technique. The clinic nurse acted as a facilitator during the sampling process. Eight data collectors were distributed in the four centers (two in each center). Data collection took place from June 2022 to January 2023, three days per week from 9 a.m. until 2 p.m.
The study participants were allocated according to Figure 1.
The inclusion criteria were nursing mothers, with a child aged one day to less than six months, an absence of any breastfeeding contraindications such as HIV, or any condition that may hinder the continuity of EBF or affect milk production such as breast augmentation, lift, or reduction, or nipple surgery, aged 18 years or over, literate, and willing to participate in the study.

Sample Size Determinations and Sampling Procedures
The sample size was calculated based on the following formula: where n = sample size, df = design effect of cluster sampling, t = the parameter related to the precision of obtaining the largest sample size (1.96 for an error risk of 5%) where the normal curve cuts off an area at the tails (the desired confidence level is 95%), p = expected prevalence of high BSE, q = 1 -p the expected proportion of moderate or low BSE, d = maximum tolerable error (the desired level of precision). The sample size was 1689, after adding 10% to compensate the anticipated sample loss or non-response rate. Najran City contains thirteen primary health centers; the researchers randomly selected 30% of the primary health centers (4 centers). The total sample size of 1689 was divided equally among the four primary health centers (422 participants from each center). In each center, all clinics dealing with children aged from 1 day to less than six months were included in the study. A proportional sample based on the follow-up rate was selected in each clinic, using the convenience sampling technique. The clinic nurse acted as a facilitator during the sampling process. Eight data collectors were distributed in the four centers (two in each center). Data collection took place from June 2022 to January 2023, three days per week from 9 a.m. until 2 p.m.
The study participants were allocated according to Figure 1.

Study Variables
The study's dependent variable was BSE, and the independent variables were women's demographic characteristics, obstetric history, previous breastfeeding experience, breastfeeding knowledge, and attitude.

Study Measurement Tools
BSE was evaluated using BSES-SF, developed by Dennis to assess breastfeeding confidence among puerperal women. The BSES-SF was rated on a 5-point Likert scale ranging from 1 (not at all confident) to 5 (always confident). The total scale score ranged from 14-70, classified as low self-efficacy from 14-42 and high self-efficacy from 43-70. Based on the study by Amini et al., the BSES-SF had good internal consistency (r = 0.910) [23]. The psychometric analysis of the BSES-SF Arabic version was tested in the United Arab Emirates and revealed a highly reliability measure (α = 0.95) [24].
The maternal breastfeeding knowledge was evaluated using the Gender Friendly Breastfeeding Knowledge Scale (GFBKS), created by Gupta et al. It comprised 18 statements rated on a 5-point Likert scale (scored as 1 = false, 2 = maybe false, 3 = don't know, 4 = maybe true, 5 = true). The GFBKS content validity was more than 0.80. The total scale score ranged from 18 to 90; the score for inadequate knowledge was 18-54, and adequate knowledge was from 55-90 [25]. The GFBKS Arabic version was validated by Tamim et al. and showed acceptable internal consistency (0.652) [26].
The Iowa Infant Feeding Attitude Scale (IIFAS) was utilized to evaluate women's attitudes to breastfeeding. IIFAS was created by De la Mora et al., and contains 17 items ranked on a 5-point Likert scale, from 1 (strongly disagree) to 5 (strongly agree). Overall IIFAS scores ranged from 17 to 85; each woman was considered to have a negative  or positive attitude (52-85) based on her score. IIFAS items' internal consistency ranged from 0.85 to 0.86 [27]. Charafeddine et al. assessed the psychometric properties of the IIFAS Arabic version with a sample of Lebanese women and found acceptable internal consistency (0.640) [28]. The scale was also validated in a study on Saudi pregnant women; the Cronbach alpha was 0.595 [29].
The researchers developed a basic data questionnaire to evaluate the women's demographic characteristics and obstetric history. Demographic data include age, occupation, residency, monthly income, and women's and their husbands' education. Obstetric history included gravidity, parity, number of living children, duration of pregnancy, complications during the most recent pregnancy and delivery, mode of delivery for the most recent child, and breastfeeding experience.

Data Collection Procedures and Technique
Data collection w from June 2022 to January 2023, three days weekly from 9 am until 2 pm. The data collection team comprised eight trained data collectors with previous experience in data collection. The data collectors were present in the waiting room in each clinic to identify the eligible participants based on the inclusion criteria. For each eligible participant, the data collector explained the study purpose and participant's role and obtained informed consent. A self-administered questionnaire was completed in the presence of the data collector. The data collector's role was to answer queries, clarify any concerns, and ensure data completeness in the questionnaire.

Data Quality Control
The data collectors have bachelor's degrees in nursing and previous data collection experience. Before data collection, the research team held two training sessions for the data collectors. The first session explained the research proposal, procedure, and ethics. At the end of the session, a copy of the data collection instrument was given to the data collectors to read before the next meeting. In the next session, one of the researchers provided a complete explanation of the questionnaire and clarified any queries. After data collection and during data entry, 27 questionnaires containing missing data were excluded from the analysis.

Ethical Approval
After research proposal approval from the deanship of scientific research, the proposal and data collection tools were assessed and approved by the Najran health affairs ethical committee; ethical approval 2023-02 E. Permission to start data collection was also obtained from the MCH administration. Informed consent was taken from each mother before starting data collection. Participant anonymity was applied, and data were utilized only for research purposes. The participants were informed about their rights to decline participation without any penalties or consequences for the care provided.

Data Analysis
Data analysis was performed using Statistical IBM, version 23 '(IBM Corp., Armonk, NY, USA)'. Data were described using numbers, percentages, means, and standard deviation. The total BSE, knowledge, and attitude scores were obtained by summing items. Chi-square (X 2 ) and Fisher exact tests (FET) were used to test group differences. Predictors of high BSE were examined through binary logistic regression. All the independent variables were categorical, and the first category was considered a reference. All factors were analyzed for multicollinearity in the regression model. The final model was checked with the Cox and Snell R-square goodness-of-fit test. Results were judged statistically significant at p < 0.05.

Results
The frequency distribution of nursing mothers by demographic characteristics and overall breastfeeding self-efficacy scores are illustrated in Table 1. More than three-quarters (78.8%) of the nursing mothers were between the ages of 20-35, and the majority (93.6%) were urban residents. More than half (55.9%) were housewives, and 51.4% reported insufficient family income. Regarding mothers' and their husbands' education, 55.0% and 66.6% had a university education, respectively. BSE was significantly high among housewives and university-educated mothers (p = 0.000). Approximately one-third (32.1%) of the nursing mothers breastfed exclusively, and 32.3% were primiparas. Around two-thirds (66.9%) delivered vaginally, and 19.7% reported complications during the most recent delivery. Concerning the participants' overall knowledge and attitudes toward BF, more than half (60.2%) had adequate breastfeeding knowledge and 56.9% had a positive breastfeeding attitude. Moreover, exclusive breastfeeding, parity, adequate knowledge, and positive attitude were associated with higher BSE (p = 0.000) ( Table 2). The mean scores and standard deviation of BSES-SF items among nursing mothers are represented in Table 3. The overall mean BSES-SF score was 51.31 ± 10.79, with the mean score ranging from 3.23-3.41 for all BSES-SF items. As shown in the table, the highest mean score was for mothers who felt comfortable breastfeeding with family members present (3.41 ± 1.06), and the lowest mean was in mothers who could breastfeed their baby without using formula as a supplement (3.23 ± 0.94). Table 3. Mean scores and standard deviation of BSES-SF items among nursing mothers.

BSES-SF Items
Mean SD  Table 4 presents the binary logistic regression analysis of high BSE predictors. Occupational status was significantly associated with high BSE. Housewives had a 1.6 times higher probability of having high BSE (AOR 1.686; 95% CI 1.23-2.30, p = 0.001) when taking employed mothers as a reference. Educational level was also a significant predictor of high BSE. Higher odds ratios were found in mothers who had received university (OR 69.474; 95% CI 39.52-122.11, p = 0.000) and secondary education (OR 45.140; 95% CI 27.95-72.88, p = 0.000) compared with mothers who only could read and write. Furthermore, breastfeeding experience was significantly associated with high BSE. A mother who breastfed exclusively had a 5.9 times higher probability of having high BSE (OR 5.949; 95% CI 1.35-26.10, p = 0.000) compared with non-exclusively breastfeeding mothers. In addition, multiparous mothers had a 3.1 times higher probability of having high BSE (OR 3.170; 95% CI 1.96-5.120, p = 0.000) when compared with their primiparous counterparts. Finally, having adequate breastfeeding knowledge and positive attitudes were positively associated with higher BSE (OR 2.769; 95% CI 1.88-4.064, p = 0.000; and OR 4.803; 95% CI 2.60-8.85, p = 0.000, respectively).   AOR: Adjusted odds ratio, CI: confidence interval, * significant at p < 0.05, ** significant at p < 0.001.

Discussion
The present study findings reveal that approximately two-thirds of the nursing mothers had a high BSE. The significant determinants of high BSE included high educational level, being a housewife, having adequate knowledge, and having positive attitudes regarding breastfeeding. Multiparous mothers and mothers who had previously exclusively breastfed their babies also had a high BSE. The current study findings can help health professionals and decision-makers to design and implement supportive interventions to improve maternal BSE. Increasing BSE will facilitate and accelerate the improvement of breastfeeding practices for the benefit of nursing mothers and their infants.
In the present study, the overall mean BSES-SF scores of 1577 Saudi nursing mothers were moderate, with a mean score of 51.31 ± 10.79 out of 70. The results indicate that Saudi nursing mothers are confident in breastfeeding their infants. Along the same lines, a cross-sectional study was conducted by Khresheh and Ahmad to evaluate the associated demographic variables of BSE among pregnant participants in Saudi Arabia. Their results indicated moderate to relatively high averages for a prenatal BSE scale of 70 out of 100. They also added that BSE is a significant variable affecting breastfeeding practice [21].
Similar mean scores were also reported among other participants in international studies using the same instrument: 50.80 ± 8.91 in Iran [23] and 49.7 in Cyprus [30]. In contrast, the BSE mean was much higher in Turkey (55.13 ± 8.39) than was documented in the current study [31], while a lower BSE mean (47.3 ± 10.50) was reported in China [32]. Among the aforementioned studies, the research by Mercan et al. [31] evaluated BSE in the first 42 days of the postpartum period. The research by Ip et al. [32] conducted a longitudinal cohort study and followed up to six months after delivery. The differences between the current study and the aforementioned studies may be attributed to the differences between the studies' designs and the data collection period.
This study revealed that a higher educational level significantly affected the BSE scale score. There is no doubt that educational level greatly affects breastfeeding knowledge and awareness. A high educational level may enhance BSE, empowering the mother in terms of health-seeking behavior and health education. Prior studies emphasized the important role of a high educational level in raising women's BSE and consequently enhancing successful breastfeeding practices [33,34]. In addition, cross-sectional surveys conducted in Taiwan suggested that educational level was positively associated with an increased likelihood of BSE, particularly among university-educated mothers [35]. According to Bandura's self-efficacy theory, if a person has good experience, their expectations will be higher [36]. In the current study, multiparous mothers with experience of exclusive breastfeeding had a higher probability of high BSE. According to Elgzar et al., previous experience of motherhood may initiate internal confidence in infant care and consequently shape maternal ideation behaviors regarding breastfeeding [37].
In several studies, higher BSE was related to previous positive breastfeeding experiences [37][38][39]. In addition, multi-parity was associated with successful breastfeeding among Saudi mothers in Al Hassa City [40]. The mediating role of BSE in successfully initiating and continuing breastfeeding practices has been documented in national and international studies [21,[41][42][43].
Another important finding was the significant association between employment situation and the BSE score. Our findings revealed that housewives were 1.6 times more likely to have a high BSE compared with working mothers. This result was similar to other studies where working mothers outside their homes had low BSE and high barriers to optimal breastfeeding practices [44,45]. In addition, a recent qualitative study in Saudi Arabia showed that workplace policies and short periods of maternity leave (45-70 days) were connected with working mothers' early introduction of supplementary feeding or early weaning [46]. Therefore, it is necessary to establish supportive rules to ensure that breastfeeding breaks for working nursing mothers are available in the workplace. An Ethiopian study reported that increasing the period of breastfeeding leave from work and establishing childcare centers near workplaces significantly improved breastfeeding practices among employed mothers [47].
The present study found that adequate breastfeeding knowledge and positive attitudes predict higher BSE. At a national level, a recent Saudi study showed that mothers with gestational diabetes mellitus who had good breastfeeding knowledge were more likely to have higher BSE [33]. The significant role of knowledge in the current study indicates the importance of promoting strategies to improve nursing mothers' breastfeeding awareness and self-efficacy. Previous international studies have reported the benefits of educational interventions for enhancing mothers' BSE and breastfeeding rates [13,48]. Consistent with prior studies [18,49], positive attitudes toward breastfeeding among nursing mothers were also significantly linked with higher BSE in the present study. Attitude toward breastfeeding has also been recognized as a predictor of breastfeeding behavior among women in Western Saudi Arabia [50]. Thus, educating pregnant and postpartum mothers to improve their knowledge and develop positive attitudes toward breastfeeding may enhance BSE and improve breastfeeding practices among women in Saudi Arabia.

Strengths and Limitations
Our study has numerous strengths. A large sample size acquired using a random cluster sampling technique provides sufficient power to analyze the role of various predictors of BSE. Furthermore, this is the first study in Saudi Arabia to investigate the determinants of BSE in nursing mothers of babies under six months of age. Some limitations are also worthy to be mentioned. The current data were collected using a self-reported questionnaire, which may be susceptible to recall bias. In addition, it was not effectively possible to apply a random sampling technique to select the participants from each primary health center; therefore, we used a convenience sample.

Conclusions
The current study found that mothers' education, working status, parity, breastfeeding experience, adequate breastfeeding knowledge, and positive attitudes were predictors of BSE. If such predictors are considered during breastfeeding-related educational interventions, it could lead to more effective and sustainable effects in community awareness regarding breastfeeding. Despite the great efforts made by the Saudi Ministry of Health to raise community awareness regarding breastfeeding, multifaceted breastfeeding educational interventions, counseling, and support are needed to improve mothers' BSE and thereby enhance proper breastfeeding practices.