The prevalence of breastfeeding (BF) in Spain has been increasing progressively in recent years. The frequency of exclusive BF at 6 months of life has increased from 28.5% in 2011 to 39% in 2017.1 However, we are still far from fulfilling the recommendation of the World Health Organization to maintain exclusive BF through age 6 months.2

Multiple factors have been found to be associated with early discontinuation of BF, including maternal age, educational attainment or psychosocial status.3 Although these are not modifiable factors, it is important to be aware of them in order to identify mothers at risk of discontinuing BF. On the other hand, there are modifiable factors on which it is possible to act to attempt to minimise this risk.4

One of the modifiable risk factors for early BF discontinuation is the lack of support to mothers from health care professionals.3,5 Providers in maternity wards play an important role in the successful establishment of BF, and primary care providers in its successful maintenance over time.6,7 To achieve this goal, it is essential that validated tools be available to objectively assess BF. At present, there are several scales that assess BF technique through the direct observation of a feed by a health care professional.8,9 The use of scales facilitates the standardization of assessment criteria and objective identification of aspects in need of improvement.6 We recommend observation of at least one full feed in the maternity ward and another one when the infant enters the primary care system.6

There is another important factor at play in the maintenance and exclusivity of BF: self-efficacy. The level of this trait in mothers is also a good predictor of early BF discontinuation.10,11 Self-efficacy refers to how competent and confident women feel about their ability to breastfeed. The main aspects that contribute to this factor are previous personal experience BF, past vicarious experience of BF in others, the physical and psychological condition of the mother and verbal persuasion by others.12 Previous studies have demonstrated that self-efficacy is a modifiable factor and that interventions that improve self-efficacy achieve an increase in BF duration and exclusive BF.4,13,14 Thus, instruments for assessment of self-efficacy can be useful and help identify mothers at risk of premature discontinuation of BF.13

The scale used most frequently for this purpose is the Breastfeeding Self-Efficacy scale-short form (BSES-SF), developed in Canada in 2003.12 The BSES-SF has been translated to different languages and adapted to different cultures and exhibited an adequate validity and reliability.15–21 There is also a validated version of the scale in Spanish.22 In addition, the BSES-SF has been found to be a good predictor of early BF discontinuation, with lower scores in the scale associated with an increased risk of discontinuation.18,19,23,24

This motivated us to assess the potential usefulness of the BSES-SF as a predictor of BF practices at the primary care level. The BSES-SF is a simple and feasible scale that could be very helpful to health care professionals. In addition, we sought to establish the optimal cut-off point to identify mothers at risk of BF discontinuation.

The data used in this study were obtained in the context of the study Lactancia y Desarrollo Infantil (Breastfeeding and Infant Development, LAyDI), so the results presented here are from a secondary analysis of the data obtained in said project.

A total of 210 paediatricians of the network participated in the LAyDI study. Fig. 1 presents a flowchart of the sample under study. We found the following rates of BF in the initial primary care visit: exclusive BF, 67.2% (65.1%–69.2%), mixed feeding, 22.1% (20.4%–24.0%) and formula feeding 10.7% (9.4%–12.1%). The mean age of the infants in the initial visit was 10.8 days (standard deviation [SD], 4.0). Table 1 summarises the main characteristics of the participants at the beginning of the study.

Figure 1.

Flowchart of study participants.

BF, breastfeeding.

(0.15MB).

The mean BSES-SF score in the total sample was 59.76 points (SD, 9.95). The mean score in each of the groups was 62.54 points (SD, 7.91) in the exclusive BF group compared to 51.79 points (SD, 10.88) in the mixed feeding group. This corresponds to a difference between the 2 groups of 10.42 points (P < .001).

Fig. 2 shows the ROC curves obtained from the BSES-SF scores of mothers that exclusively breastfed their infants. We ought to note that only the curves corresponding to the visits at 1 and 2 months had an AUC above the threshold established to define good predictive power (0.75).

Figure 2.

ROC curves of BSES-SF scores in the exclusive breastfeeding group at different infant ages.

AUC, area under the curve (with 95% confidence interval); BSES-SF, breastfeeding self-efficacy scale-short form.

(0.13MB).

Fig. 3 presents the ROC curves of mixed feeding group. In this instance, there were no AUCs greater than 0.75 in any infant age group.

Figure 3.

ROC curves of BSES-SF scores in the mixed feeding group at different infant ages.

AUC, area under the curve (with 95% confidence interval); BSES-SF, breastfeeding self-efficacy scale-short form.

(0.13MB).

Since the only curves with AUCs above the threshold for good predictive power were those in the exclusive BF group at 1 and 2 months, these were the age groups for which we calculated the sensitivity and specificity of BSES-SF scores ranging from 30 to 70 points. The results are presented in Table 2. As can be seen in the table, a sensitivity of 80% was achieved with a score of 59 points at 1 month and 58 points at 2 months.

The BSES-SF score found in our study (59.76 points) was above the scores reported in other studies conducted in Spain. In the validation study conducted by Oliver-Roig et al., the mean score in the BSES-SF was 51.94 points,22 while Marco-Alegría et al. reported a mean of 50.03 points.25 In these studies, the BSES-SF was administered in the maternity ward before day 5 post birth. This may explain the observed differences, as in the LAyDI study, mothers completed the BSES-SF on day 10 post birth, when they had been breastfeeding longer and presumably felt more competent. Ideally, the initial newborn evaluation in primary care centres should take place 48–72 h after discharge.26 The results of our study suggest that the time elapsed between both care settings (maternity ward and primary care centre) is longer than recommended, which poses a risk to infant health and may contribute to early BF discontinuation.26

Studies conducted in other countries have produced heterogeneous results, with the lowest scores found in Asian countries (where the prevalence of BF is substantially lower): 47.3 points China19 and 42.39 points in Japan.27 In other European countries, scores are similar to those in our study: 54.8 points in Italy,18 57.4 in Sweden15 and 63.6 in Portugal.21 Therefore, the scores obtained in the scale seem to be associated with the prevalence of BF in the region where self-efficacy is being assessed.

The BSES-SF performed adequately for prediction of BF maintenance. Some studies have found that the scale offers a good predictive power through 3 months post birth,17,18,23,27 compared to 1 month in others.19,22 In our study, the BSES-SF exhibited a good predictive power through 2 months post birth in mothers that breastfed exclusively. We did not find any studies that analysed the predictive power of the scale differentiating between mothers that practiced exclusive BF and mothers that practiced mixed feeding. Given the substantial differences we observed in the BSES-SF scores between these 2 groups of mothers, we thought it would be prudent to make a separate analysis for each. Based on the results, the scale does not seem to be a good predictor in the case of mixed feeding.

As for the optimal cut-off point in the BSES-SF, a similar study estimated a cut-off score of 50 points.27 However, this study was conducted in the Asian population, in which BF trends, as we mentioned above, are very different compared to Western countries. Considering that for the purpose of prediction the cut-off point should be selected with the aim of detecting the highest possible number of mothers at risk of BF discontinuation, we decided to use the sensitivity as the parameter used to select the optimal cut-off value. The ultimate purpose of identifying mothers at risk of discontinuing BF is to offer increased BF support.4 We arbitrarily selected a minimum sensitivity of 80% to define adequate predictive power, although any other of the values presented in Table 2 could be applied. Thus, we propose using a cut-off score of 58 points, as it offers a sensitivity greater than 80% in the 2 age groups in which we calculated the threshold.

Most studies on scales used for assessment of BF have been made in the hospital setting and in the first days post birth.5,18–22,27 We only found one study conducted in the primary care setting that used the BSES-SF to assess factors that affect BF.24 Our analysis revealed that the application of scales in the primary care setting to predict BF outcomes can be useful. The BSES-SF is completed in a short time and can be administered in the first visit of newborn infants that are exclusively breastfed. However, the best approach would be to implement follow-up plans developed and implemented by hospital maternity departments in collaboration with primary care services.6 Breastfeeding assessment data collected in hospital should reach the primary care system, ensuring continuity of care. The best way to convey this information would be through standardised instruments6 like the BSES-SF or scales used to assess observed feeds.8

One of the limitations of this study was the loss to follow-up of some patients, a common limitation of longitudinal studies that also affected the LAyDI project. Although 17.4% of the initial sample did not complete the follow-up, the number of participants that continued in the study at 6 months post birth was greater than the estimated minimum necessary sample size. Thus, although there were losses to follow-up, the results are still valid. On the other hand, since we analysed secondary data, we did not perform a sample size calculation specifically for our study. As a result, there were considerable differences between the participants in the 2 feeding modality groups that we analysed separately (exclusive BF and mixed feeding). Nevertheless, due to its substantial size, the sample yielded BSES-SF scores representative of the population despite the disparities between the feeding modality groups.

Due to all of the above, we conclude that the BSES-SF is a useful instrument at the primary care level to identify mothers at risk of discontinuing BF before 2 months post birth. Mothers with scores of 58 points or lower should receive additional BF support.

The LAyDI study was awarded one of the research grants of the Asociación Española de Pediatría de Atención Primaria-Fundación Pediatría y Salud (AEPap-FPS) in 2017.

The authors have no conflicts of interest to declare.