Psychometrics of the breastfeeding self-efficacy scale and short form: a systematic review

Abstract

Background

The Breastfeeding Self-Efficacy Scale and its short-form were developed in Canada and have been used internationally among numerous maternal populations. However, the psychometric properties of the scales have not been reviewed to confirm their appropriateness in measuring breastfeeding self-efficacy in culturally diverse populations. The purpose of this research was to critically appraise and synthesize the psychometric properties of the scales via systematic review.

Methods

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed. Three databases (EMBASE, MEDLINE, and PsycINFO) were searched from 1999 (original publication of the Scale) until April 27, 2022. The search was updated on April 1, 2023. Studies that assessed the psychometric properties of the BSES or BSES-SF were included. Two researchers independently extracted data and completed the quality appraisals.

Results

Forty-one studies evaluated the psychometrics of the BSES (n = 5 studies) or BSES-SF (n = 36 studies) among demographically or culturally diverse populations. All versions of the instrument demonstrated good reliability, with Cronbach's alphas ranging from .72 to .97. Construct validity was supported by statistically significant differences in mean scores among women with and without previous breastfeeding experience and by correlations between the scales and theoretically related constructs. Predictive validity was demonstrated by statistically significant lower scores among women who ultimately discontinued breastfeeding compared to those who did not.

Conclusion

The BSES and BSES-SF appear to be valid and reliable measures of breastfeeding self-efficacy that may be used globally to (1) assess women who may be at risk of negative breastfeeding outcomes (e.g., initiation, duration and exclusivity), (2) individualize breastfeeding support, and (3) evaluate the effectiveness of breastfeeding interventions.

Article summary

The BSES and BSES-SF appear to be valid and reliable measures of breastfeeding self-efficacy that can be used globally to identify women at-risk for poor breastfeeding outcomes.

Background

The benefits of breastfeeding for disease prevention and health promotion are undisputed. If breastfeeding exclusivity occurred at a near universal level among young infants, it is estimated that 823,000 deaths in children under the age of five could be prevented annually [1]. Due to the beneficial effects, exclusive breastfeeding for the first six months postpartum is recommended internationally. While the overall global rate of exclusive breastfeeding for infants less than six months of age is currently 44% [2], the World Health Organization has set a goal to achieve at least a 50% exclusivity rate by 2025 [3]. One potential highly effective strategy to improve exclusive breastfeeding rates is to tailor supportive resources among women at risk of poor breastfeeding outcomes [4]. Breastfeeding self-efficacy is one possible modifiable variable that has been consistently associated with positive breastfeeding outcomes, including exclusivity [5, 6].

Breastfeeding self-efficacy is defined as a mother’s confidence in ability to breastfeed [7] and predicts “whether a mother chooses to breastfeed, how much effort she will expend, whether she will persevere in her attempts until mastery is achieved, whether she will have self-enhancing or self-defeating thought patterns, and how she will emotionally respond to breastfeeding difficulties” (p. 736). Consistent with Bandura’s Social Learning Theory [8], Dennis’ breastfeeding self-efficacy theory [7] hypothesizes that maternal breastfeeding self-efficacy may be affected by four primary sources (e.g., antecedents) including [1] performance accomplishments (e.g., past breastfeeding experiences), [2] vicarious experiences (e.g., watching other women breastfeed), [3] verbal persuasion (e.g., encouragement from influential others such as friends, family, and lactation consultants), and [4] physiological responses (e.g., pain, fatigue, stress, depression, anxiety). Thus, an individual’s self-efficacy may be enhanced by altering the sources of information.

The 33 item five-point Likert Scale Breastfeeding Self-Efficacy Scale (BSES) was developed by Dennis [9]. Each item was preceded by the phrase I can always…, with responses ranging from not at all confident to always confident. A psychometric evaluation of the BSES were initially conducted with a sample of 130 Canadian women, resulting in a Cronbach's alpha of 0.96 and 73% of all item-total correlations falling within the 0.30-0.70 range [9]. Factor analysis revealed two distinct factors: (1) Breastfeeding Technique Subscale, and (2) Intrapersonal Thoughts Subscale. In the initial sample, BSES scores were predictive of breastfeeding duration at 6 weeks postpartum. Internal consistency data, however, suggested the presence of redundant items and so the scale was retested in a larger Canadian sample [10]. After a detailed item analysis, 19 items were deleted, culminating in the 14-item BSES-Short Form (SF) [10]. Total scores range from 14–70 with lower scores indicating lower breastfeeding self-efficacy. Today, the BSES-SF is widely used internationally to identify women who may be at-risk for prematurely discontinuing breastfeeding.

In a meta-analysis of 11 trials evaluating breastfeeding self-efficacy interventions among women of term infants, researchers found that intervention groups participants were 1.56 times more likely to be breastfeeding at 1 month increasing to 1.66 times more likely at 2 months postpartum [5]. The researchers concluded that interventions that began in the postpartum period that used combined delivery settings (e.g., hospital and community) or were theoretically derived, had the largest effect on breastfeeding self-efficacy and rates of breastfeeding. Further, meta-regression analysis suggested the odds of exclusive breastfeeding increased by 10% among intervention participants for each 1-point increase in mean BSES scores between the intervention and control groups. Similarly, another systematic review and meta-analysis of 12 trials [11] revealed that women receiving breastfeeding support interventions had significantly improved breastfeeding self-efficacy scores during the first 4 to 6 weeks postpartum (SMD = 0.40, p = 0.006, 95% CI [0.11, 0.69]) and decreased perceptions of insufficient milk supply (median, 3.3, p < 0.001).

While a general review of the BSES-SF was completed [12], no systematic review has been undertaken to examine the application of both the BSES and BSES-SF in culturally and demographically diverse populations. The cross-cultural adaptation of the scale, as well as the validation among mothers with specific demographic or clinical characteristics, is crucial for enabling the instrument to serve as a useful international measure of breastfeeding self-efficacy. Furthermore, the development of modified versions of the instrument with sound psychometric properties facilitates an awareness of—and sensitivity to—the needs and perceptions of mothers with varied cultural, socioeconomic, and medical histories. Thus, the objective of this review was to appraise the validated BSES and BSES-SF in culturally and demographically diverse contexts.

Methods

Design

We performed a systematic review of quantitative studies using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [13]. A protocol was not registered.

Sample: Defining the articles reviewed

Quantitative studies (e.g., methodological, cross-sectional, cohort) were included if they used the BSES or BSES-SF and provided data concerning both the reliability and validity of the scale. Studies that aimed to determine predictive cut-off scores for the scale were also included if they analyzed the sensitivity and specificity of the scale at the identified cut-off. A total of 41 studies were included in the review. All studies were published in English; thus, there were no exclusions for language. Figure 1 displays a flowchart of the search strategy and the study selection process.

Fig.1

PRISMA Flow diagram

Data collection: The search strategy and process

The initial search was conducted for published studies between 1999 (the publication year of the original BSES) and December 2021. Additional studies were identified via reference list searches and doing a ‘cited by’ search on Google Scholar and Medline as of April 1, 2023. We also contacted experts in the field to retrieve any data from recently completed studies not yet published. Searches of the literature were conducted using EMBASE, MEDLINE, and PsycINFO. The term “Breastfeeding Self-Efficacy Scale” was searched and Medical Subject Heading (MeSH) terms for “breastfeeding”, “self-efficacy”, “cultural adaptation”, “psychometric”, and “validated” were broadened to capture relevant literature. The search strategy is presented in Table S1.

Measurement

A structured data extraction form was developed to organize data from the studies by publication year, study location, objectives, population, validation of the BSES or BSES-SF, language used for the scale, time of administering the scale, and means and standard deviations (SD) of BSES or BSES-SF scores. Study quality was assessed using criteria suggested by Shrestha et al. [14] and Mirza and Jenkins [15] and included: (1) clarity of the study aim, (2) sample size justification, (3) sample representativeness, (4) clear inclusion and exclusion criteria, (5) description of maternal demographic data, (6) reporting of a response rate, (7) appropriate statistical analyses, and (8) evidence of participant informed consent. Possible scores of 1 (e.g., sample size justified) or 0 (e.g., sample size not justified) were used and combined to give a possible total score of 8 for each study. Quality criteria specific to the translation and validation of psychometric tools such as the BSES and the BSES-SF were adapted from Shrestha et al.’s [14] additional quality assessment parameters, which included an assessment of the translation methods used, cultural adaptation, as well as any modifications made to items in the scale. Two authors independently extracted data and completed the quality appraisals. When information was unclear authors reviewed the data and/or had a third author review to achieve consensus.

Data analysis

Data pertaining to the psychometrics of the scales were summarized and included internal consistency, factor analyses, known groups analyses, and predictive validity of the scales, as well as the correlation of the scores with other theoretically related constructs, negative and positive predictive value, and sensitivity and specificity. As no other measure of breastfeeding self-efficacy was used in these studies for comparison, instruments to which BSES or BSES-SF scores were compared differed slightly among the analyzed studies. Additional scales administered in several studies included the Edinburgh Postnatal Depression Scale (EPDS) [16] and the General Self-Efficacy Scale (GSES) [17]. Since self-efficacy has been shown to be negatively influenced by psychological disorders such as depression; a negative correlation between BSES and EPDS scores was hypothesized. Based on the rationale that breastfeeding self-efficacy should be enhanced among mothers with higher overall self-efficacy, a positive correlation between BSES and GSES scores was also hypothesized.

Results

Characteristics of included studies

The search yielded 305 studies, with 194 screened after duplicates were removed through automation. Of these screened studies, 171 did not meet the inclusion criteria, leaving 21 eligible studies after Dennis’ two original BSES [9] and BSES-SF [10] studies were excluded. Sixteen additional studies were identified via reference searches and reverse Google Scholar searches and five studies were identified pre-publication through author correspondence, with one not meeting the inclusion criteria. In total, 41 studies were included in the review with five focused on the original BSES and 36 presenting data on the BSES-SF. Of the BSES-SF studies, three reported on a modified scales for fathers [18,19,20] and three reported on a modified scale for mothers of preterm infants [21,22,23]. All included studies were of high quality with scores ranging from four to 8 eight (see Table S2).

Characteristics of the included studies are presented in Table 1. For the five studies validating the BSES, sample sizes ranged from 100 to 276 participants, with 60% falling within the recommended sample size for psychometric assessments of 5–10 subjects per item (Nunnally and Bernstein: Psychometric theory, unpublished). The 36 studies assessing the BSES-SF had sample sizes ranging from 18–1,524, all except two [24, 25] surpassing the minimum recommended sample size. Participants were primarily recruited from maternity wards of urban hospitals and most sample characteristics were representative of the regional population. The most common time to assess breastfeeding self-efficacy was in-hospital during the postpartum period (n = 17; 41%). However, the timing of assessment varied among some studies with assessment of breastfeeding self-efficacy conducted during the third trimester of pregnancy [26, 27], postnatally (time not reported) [28, 29], at 1 week [27], or open time periods such as two to six weeks [20].

Table 1 Language of validation, objective, sample population and mean BSES/BSES-SF scores for all included studies

The study samples included in the review were diverse in both cultural and demographic characteristics (see Table 1). Six studies assessed the scale in a clinically specific maternal sample including adolescents [39, 41], Canadian Indigenous women [48], black women in the United States [47], and low-income women [25, 40]. In two studies the authors assessed the English version in diverse populations in Australia [26] and the UK [43]. In most studies researchers reported that the scale was used with mothers of term infants. However, three studies reported on the use of the BSES-SF specifically in mothers of ill and preterm infants (< 36 weeks) in Canada [22], Turkey [21], and China [23]. Three studies reported on the use of the BSES-SF for fathers in Canada [19], Malawi [18], and Turkey [20]. Modifications for the BSES-SF scale for mothers of ill and preterm infants included, the addition of four new items, some items were changes to be applicable to mothers of preterm infants, and the item stem was changed from I can to I think I can. Similarly, for the BSES version for fathers, some items were changed to reflect the partner’s experience in assisting the breastfeeding mother.

The scale was translated into 24 different languages, including but not limited to Spanish, Turkish, Swedish, Croatian, Acholi and Langi, Japanese, Portuguese, Italian, Polish, and Malay (Table 2). In 20 studies, there were minor item word modifications. The translation methods were comparable across all studies with most completing back-translation (n = 28; 84.8%) to ensure semantic equivalency. However, there were variations noted in the quality of the translation process. Among the studies translated into English (n = 33), only 13 (approximately 40%) had at least two qualified personnel doing the forward and back translation with some not specifying if they were independent groups of translators. Eight studies were classified as not applicable regarding translation as they were psychometric studies conducted in English or the tool had been previously translated to English. Pilot testing was conducted in most of the translated studies (n = 27; 77.1%) with samples sizes ranging from 5 to 31. A few studies did not report data on pilot testing or translation procedures.

Table 2 Translation methods and modifications for all included studies

Psychometric assessments of BSES and BSES-SF

Reliability

Five studies that evaluated the BSES reported Cronbach's alphas ranging from 0.88 to 0.97 while the remaining 36 studies that assessed the BSES-SF reported Cronbach's alphas ranging from 0.72 to 0.96. Studies validating non-English versions (n = 34) had a wider range in Cronbach's alpha’s (0.72—0.95) than the seven studies analyzing English versions (0.88—0.96). The original BSES reported a Cronbach’s alpha of 0.96 [9] and the BSES-SF reported a Cronbach's alpha of 0.94 [10], indicating that the translated and modified versions are comparable. The reliability of all modified tools was further supported by the fact that deletion of any single item did not lead to an increase in Cronbach's alpha of more than 0.10, except one [37]. For studies validating the BSES-SF, the majority of item-total correlations were above the recommended 0.30 criterion.

Construct validity

Five studies validated the construct validity of the BSES using factor analysis [26, 30,31,32,33], in addition to the original development by Dennis and Faux [9]. The five studies reported a 2-factor solution with eigenvalues ranging from 4.75 to 11.99 and the combined two factors contributing to 29.2% to 57% of the variance. In the original development of the BSES, the 2-factor solution had eigenvalues of 2.75 and 16.87 and the combined two factors contributed to 45.6% of the variance [9]. The 2-factor solution in all cases was consistent with the original BSES factor analysis and congruent with the theorized breastfeeding technique and intrapersonal thoughts subscales [9]. In the studies that completed a factor analysis of the BSES-SF, a single-factor solution was frequently reported, consistent with the original BSES-SF’s unidimensional structure [10].

Construct validity was further assessed in 32 studies using known-groups analysis. It was hypothesized that women who have successfully breastfed in the past would have higher self-efficacy scores than those with no prior breastfeeding experience. All studies (n = 18) except two [21, 41] reported significant differences in mean BSES or BSES-SF scores among women who had previously breastfed compared to those with no previous breastfeeding experience (Table 3). Nineteen studies compared mean breastfeeding self-efficacy scores between primiparous and multiparous women, with only half reporting a statistically significant difference based on parity. This finding is consistent with the breastfeeding self-efficacy theory [7] and supports the importance of the information source of performance accomplishment and that previous breastfeeding experience, not parity, is an important indicator of breastfeeding self-efficacy.

Table 3 Psychometric properties of the modified versions of the BSES AND BSES-SF for all included studies

Mean breastfeeding self-efficacy scores were compared across the other known-group variables including: (1) accurate versus inaccurate breastfeeding knowledge [37]; (2) intended breastfeeding duration more than 6 months versus less than 6 months [38, 47]; (3) positive versus negative previous breastfeeding experience [33, 42, 49]; (4) timing of decision to breastfeeding comparing early versus late pregnancy [39]; and, (5) exclusive versus partial breastfeeding (Maurer, et al.: The breastfeeding self-efficacy scale - short form (BSES): German translation and psychometric assessment, unpublished). In all studies, researchers reported significant group differences in mean breastfeeding self-efficacy scores and the scale’s ability to accurately predict group membership. When in-hospital BSES-SF scores were examined between maternal-paternal pairs, a significant correlation (r = 0.53, p < 0.001) was found [19].

Correlations between breastfeeding self-efficacy scores and other theoretically related constructs were examined in several studies. Constructs for which a hypothesized positive correlation was found included the General Self-Efficacy Scale (GSES) [49]; the Sense of Coherence (SOC) subscales of comprehensibility, manageability and meaningfulness [51]; the Stress Management Self-Efficacy Scale (SMSE) [49]; the Rosenberg Self-Esteem Scale (RSES) [58]; the Hill and Humenick Lactation Scale (HHLS) [22, 26]; the Breastfeeding Attitudes Questionnaire (BAQ) [39]; the Iowa Infant Feeding Attitude Scale (IIFAS) [19]; the Exclusive Breastfeeding Social Support scale (EBFSS) [59]; the Multidimensional Scale of Perceived Social Support [36]; the Network Support for Breastfeeding Tool [47]; the WHO Quality of Life (QoL)-BREF [18]; maternal perceptions of breastfeeding progress [19]; and paternal perceptions of breastfeeding importance [19] (Table 3). Researchers in eleven studies reported a significant negative association between EPDS and BSES scores [28,29,30, 36, 38, 44, 48, 52, 53, 55, 57]. In particular, women with higher scores on the EPDS had significantly lower BSES scores. Using other measurements of depression, Boateng et al. [37] reported a negative correlation between Centre for Epidemiological Studies Depression Scale (CES-D) [60] and BSES-SF scores and two researchers [29, 54] reported a negative correlation between the HADS depression subscale and BSES-SF scores. All three studies that examined the correlation between anxiety and BSES-SF, researchers found a negative correlation between anxiety and BSES-SF scores, using different measurement tools [21, 29, 38]. Lastly, Otsuka et al. [50] reported a negative correlation between BSES-SF scores and maternal perceptions of insufficient milk supply as measured by the PIM tool. Notably, translated scales also displayed significant construct validity and convergent validity with translated versions of other theoretically related constructs. Thus, evidence of construct and convergent validity in these studies conducted with non-English speaking participants adds to the strength of the psychometric analysis and provides further support for the use of the translated version in the tested population.

Predictive validity

The utility of BSES and BSES-SF scores as a means of predicting actual breastfeeding outcomes was assessed in most of the studies, wherein prior mean scores of mothers who breastfed were compared to those who discontinued breastfeeding. In the original BSES validation assessment, mothers still exclusively breastfeeding at six weeks postpartum had significantly higher mean scores (M = 173.5, SD = 20.9) than those engaging in combination feeding (M = 161.9, SD = 37.1) or exclusively bottle-feeding (M = 145.3, SD = 22.4) [9]. Two studies assessed predictive validity of the BSES-SF with regard to breastfeeding initiation outcomes [10, 39] (Table 3). Thirteen studies evaluated the utility of the BSES (n = 2) and BSES-SF (n = 11) as predictors of breastfeeding duration, and 25 studies reported results on the predictive validity of BSES (n = 4) or BSES-SF (n = 21) scores as indicators of breastfeeding exclusivity. Statistical significance was reached in all 25 studies except four.

Sensitivity and specificity

In two studies [35, 45], researchers evaluated the sensitivity and specificity of the BSES-SF at particular cut-off points using receiver operating characteristic (ROC) curve analysis. Sensitivity provides an estimate of the tool’s accuracy in identifying mothers at risk of premature breastfeeding cessation (i.e., the proportion of the sample that discontinued breastfeeding prior to the time period in question and scored below the identified cut-off). In contrast, specificity refers to the proportion of the sample that did not discontinue breastfeeding before the assessed time period and scored above the identified cut-off. In the Hong Kong Chinese version of the BSES-SF, Ip et al. [45] identified the optimal cut-off for predicting early breastfeeding cessation (before 6 months) to be a score of 45.5 out of 70. Sensitivity at this cut-off score was 77%, and specificity was 73%. The negative predictive value of the scale indicated that 92% of mothers who scored below 45.5 discontinued breastfeeding prior to 6 months postpartum. In the Spanish version of the BSES-SF, Balaguer-Martinez et al. [35] found that the area under the curve was above the threshold for good predictive power for mothers who were exclusively breastfeeding at 1 and 2 months. To achieve 80% sensitivity the BSES-SF cut-off score was 59 at 1 month and 58 at 2 months. These findings demonstrate the utility of the BSES-SF as a tool to identify mothers at risk of prematurely discontinuing breastfeeding.

Discussion

In this systematic review, an evaluation of the reliability and validity of the BSES and BSES-SF scales in multiple languages, as well as their psychometric assessments in specific perinatal populations, including fathers [18,19,20] and parents with an ill or preterm infant [21,22,23], was conducted to appraise their effectiveness in identifying women and partners at risk of poor breastfeeding outcomes. This review is the first to evaluate the psychometrics of translated and adapted versions of the BSES and BSES-SF among demographically and culturally diverse populations. Additionally, this is the first review to assess the rigour and quality of the studies that have adapted and applied the BSES or BSES-SF for measuring breastfeeding self-efficacy. The BSES and BSES-SF have been psychometrically tested in 41 studies and translated into 26 languages other than English. Psychometric properties of the BSES and BSES-SF reported in the included studies were comparable to the original studies completed by Dennis [9, 10] indicating the utility of the instrument as an adaptable and reliable tool for measuring breastfeeding self-efficacy in diverse populations and settings. However, our review also found that some studies had deficiencies in their translation and cultural adaptation processes. Thus, while the BSES and BSES-SF appear to have sound psychometric properties across studies, caution in the interpretation of the findings should be considered as cultural aspects may not have been captured by the instruments [14]. Future studies should use established methodological approaches [14, 61] for translating and adapting the BSES-SF for use in cross-cultural research to ensure important cultural nuances are included in translated and culturally adapted tools.

Reliability was indicated with all Cronbach alphas coefficients exceeding the recommended 0.70 for established instruments (Nunnally and Bernstein: Psychometric theory, unpublished). There was variability between studies with some translated studies having wider Cronbach alpha coefficients than the non-translated studies. This could be due to the sample size, slight modifications made during the translation or adaptation process, cultural nuances that might not be captured in the translated version or varying sample characteristics of the included studies. We anticipated some variability and do not believe that this is an appreciable difference as all studies exceeded the recommendation. As such, this does not impact the overall reliability of the BSES-SF to be used in diverse samples effectively to guide interventions.

The modified versions of the BSES and the short form were shown to be conceptually valid with the majority of the studies reporting expected correlations with other widely used and theoretically related constructs such as the depression, anxiety, general self-efficacy, breastfeeding attitude, and sense of coherence. Moreover, considering the main results from across all studies, the BSES and BSES-SF and their adapted versions demonstrated significant predictive validity at various time points in the postpartum period. This finding supports the notion that even after translation or modification of these scales, they remain useful tools for identifying mothers at risk of negative breastfeeding outcomes in terms of all three breastfeeding parameters: initiation, duration, and exclusivity. This has significant implications for applying the BSES and BSES-SF in future studies across various cultural and demographic contexts.

Conceptual equivalence was another key indicator in measuring the comparability of the original English BSES and BSES-SF with its adapted, translated versions. While most of the studies identified used forward and back-translation, it is important to note that some studies used lower quality translation processes with only one individual performing the translation and in other studies limited details were provided. Back-translation to English may have been influenced by translators possessing knowledge of the original English version of the scale, especially if forward and back-translations were conducted by the same group. This method may result in conceptual disparities between the original English BSES and its short form and their translated versions. Thus, it is important that rigorous translation and adaptation processes be used to enhance the validity and reliability of the instrument (e.g., BSES-SF) for use among individuals of diverse cultures and languages [61].

While the BSES and BSES-SF were widely used globally, we found that middle to higher income countries have predominantly adapted and validated the tool. This is important to acknowledge as cultural and demographic influences may have led to higher breastfeeding self-efficacy in some studies. In countries with higher rates of breastfeeding, women are typically more often exposed to the primary sources of self-efficacy (e.g., antecedents) [7] such as vicarious experience (seeing others breastfeed) and verbal persuasion (receiving positive reinforcement). Furthermore, women may have received more assistance (e.g., education and support) [5] with breastfeeding in some settings thereby enhancing others sources of information such as performance accomplishment.

In many low and middle-income countries, there is a growing burden of breastfeeding attrition and increased reliance on formula use [62]. Major contributing factors to this trend are the influence of the private sector in promoting formula as an alternative to breastfeeding, lack of access to health care professionals and support from caregivers, limited education, and poor awareness stemming from broader political and economic disadvantages [4, 61, 63]. These contributing factors highlight the need for continued analyses of the factors associated with low breastfeeding initiation, duration and exclusivity, and underline the importance of developing reliable and valid instruments for identifying mothers most at risk of developing suboptimal breastfeeding practices. The psychometrics of such tools should be assessed in a wide range of languages, demographic contexts, and cultural settings. While our review found that the BSES and BSES-SF are adaptable, reliable, and validated tools globally, the benefits of the tool have not been tested and evaluated in resources poor settings.

Overall, these findings have important implications globally for clinical practice. The BSES and BSES-SF appear to be reliable and valid tools that may be used to assess mothers’ breastfeeding self-efficacy and plan interventions that are based on maternal need [7]. While our review included studies that assessed breastfeeding self-efficacy at various time periods, the antenatal period and the early postpartum period, soon after delivery, have the most clinical utility for assessing breastfeeding self-efficacy. As the main purpose in administering the BSES-SF is to identify women who may be at risk for early breastfeeding discontinuation, assessment of breastfeeding self-efficacy later in the postpartum period is not as clinically relevant as infant feeding has typically been established.

The total BSES score (14 – 70) may be used to identify mothers with low breastfeeding self-efficacy who may be at risk for negative breastfeeding outcomes (e.g., initiation, duration and exclusivity) and may benefit from additional supportive interventions [64]. Conversely, mothers who have high breastfeeding self-efficacy may be recognized as being more likely to succeed with breastfeeding; however, additional assistance may still be required, particularly when experiencing breastfeeding difficulties [7]. It is noteworthy that we do not recommend a specific cut-off score (e.g., total score) to delineate high versus low breastfeeding self-efficacy as breastfeeding self-efficacy scores can be culturally specific and vary. The single item BSES scores (1 – 5) can also be used to assess maternal perceptions of self-efficacy regarding specific components of breastfeeding (e.g., determine the baby is getting enough, proper latch, exclusive breastfeeding, etc.). Individual items scores (1 – 5) can be used to identify perceptions of low self-efficacy (item score ≤ 3) and items where the mother feels efficacious (item score ≥ 4) so that support specific to individual maternal needs can be provided [64]. The BSES has also been utilized to develop and/or evaluate the efficacy of various types of supportive interventions [5, 6]. Finally, the BSES assessment may provide health care professionals with a better of understanding of where mothers lack breastfeeding confidence and why they may be unsuccessful despite additional support [7].

Limitations

Our systematic review has several limitations. First, we found that only two studies included sensitivity and specificity data and reported negative and positive predictive values. Hence, we were not able to conduct a comprehensive analysis of these parameters. Second, the quality of the translation and cultural adaptation among several studies was lacking. Some studies also had missing data (e.g., not reported) affecting the assessment of reliability and validity. Finally, while most studies in this review employed very similar validation methods, the timing of assessment varied among studies as did the characteristics and geographic location of the study participants. The heterogeneity of the samples can make comparison across studies difficult; however, the consistency of the findings between studies also suggests the versatility of the tool among diverse groups.

Conclusion

Breastfeeding is a practice that is approached and perceived differently among different cultures, and premature discontinuation of breastfeeding is a global public health concern. Continued efforts are needed in the cross-cultural adaptation of the BSES and BSES-SF to effectively serve diverse populations and provide contextually appropriate measures of breastfeeding self-efficacy. We recommend that future studies validating translated or adapted versions of the BSES and BSES-SF adopt more systematic approaches to empirical validation, cultural adaption, and translation of the scales that are consistent with those used in the original analysis of the psychometric properties of the BSES and BSES-SF. Considering the extent to which the BSES and BSES-SF take into account the needs and perceptions of non-English-speaking mothers’ self-efficacy, further efforts should be made to translate the BSES into other languages.