Health literacy of pregnant women and duration of breastfeeding maintenance: A feasibility study

J Adv Nurs. 2021;77:703–714. wileyonlinelibrary.com/journal/jan | 703© 2020 John Wiley & Sons Ltd Received: 16 July 2020 | Revised: 23 September 2020 | Accepted: 27 October 2020 DOI: 10.1111/jan.14625 ORIGINAL RESEARCH: EMPIRICAL RESEARCH – QUANTITATIVE Health literacy of pregnant women and duration of breastfeeding maintenance: A feasibility study Rafa Vila-Candel1,2 | Francisco Javier Soriano-Vidal2,3,4,5 | Desirée Mena-Tudela6 | José Antonio Quesada7 | Enrique Castro-Sánchez8,9 1Department of Obstetrics and Gynaecology, Hospital Universitario de la Ribera, Valencia, Spain 2Department of Nursing, Universitat de València, Valencia, Spain 3Department of Nursing, University of Alicante, Alicante, Spain 4Department of Obstetrics and Gynaecology, Hospital Luis Alcanyis, Xàtiva, Spain 5Foundation for the Promotion of Health and Biomedical Research in the Valencian Region (FISABIO), Valencia, Spain 6Department of Nursing, Universitat Jaume I. Av de Vicent Sos Baynat, Castelló, Spain 7Department of Clinical Medicine, Universidad Miguel Hernández, Elche, Alicante, Spain 8National Institute for Health Research Health Protection Research Unit (NIHR HPRU), Healthcare Associated Infection and Antimicrobial Resistance at Imperial College London, London, UK 9School of Health Sciences, City, University of London, London, UK Correspondence Desirée Mena-Tudela, Department of Nursing, University Jaume I. Avda, Sos I Baynat s/n 12071, Castellón de la Plana, Spain. Email: [email protected] Funding information This project has been funded by the Conselleria de Educación, Investigación, Cultura y Deporte of the Generalitat Valenciana in its call for grants for the conduction of R&D&I projects developed by emerging research groups in 2018 (Reference GV/2018/036). ECS is affiliated with the National Institute for Health Research Health Protection Research Unit (NIHR HPRU) [grant number HPRU2012–10047] in Healthcare Associated Infections and Antimicrobial Resistance at Imperial College London in partnership with Public Health England (PHE), he is an NIHR Senior Nurse and Midwife Research Leader, and acknowledges the support of the NIHR Biomedical Research Centre. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR, the Department of Health or Public Health England. Abstract Aims: Research the association between health literacy (HL) and exclusive breastfeeding at 4-months postpartum. Background: Despite the benefits of breastfeeding (BF), its rates are low worldwide. Among the reasons for abandonment is the level of maternal education. Maternal education has been associated with HL, but evidence between HL and BF maintenance is limited. Design: A cross-sectional study. Methods: The sample compromised 229 nursing mothers recruited from January 2018 to the end of December 2018 at Spain by systematic sampling method. Women were interviewed postpartum on parameters associated with the start and continuation of BF up to 4 months postpartum. Multivariate logistic regression models to explain exposure variables and exclusive BF cessation at 4 months. Results: Approximately 10% of the participants had inadequate HL. Factors associated with early cessation of exclusive BF at 4 months in the multivariate model adjusted using a stepwise variable selection process based on a likelihood ratio test were civil status, risk of pregnancy, type of delivery, limited or inadequate level of HL, and LATCH score at discharge, with an 85.6% area under the ROC curve. Conclusions: Our study offers preliminary evidence regarding the hitherto inconsistent relation between HL and early cessation exclusive BF at 4 months, supporting the conduct of further studies with larger sample sizes and greater statistical 704 | VILA-CANDEL Et AL. 1 | INTRODUCTION Breastfeeding (BF) has long-term clinical benefits, it lowers the risk of infections and sudden infant death syndrome, childhood leukaemia, or obesity (Chan et al., 2016). In turn, maternal benefits comprise improved bonding with the newborn infant, body weight stabilization, lessened risk of postpartum depression, and a lower incidence of breast or ovarian malignancies (Abou-Dakn, 2018; Chowdhury et al., 2015). Lastly, BF affords both shortand long-term economic and environmental societal advantages (Rollins et al., 2016). 2 | BACKGROUND The World Health Organization (WHO) recommend BF as the optimal feeding for all infants and exclusive breastfeeding (EBF) for the first 6 months of life; however, the rate of EBF in 4 months after delivery globally is still low (Victora et al., 2016). Despite these apparent benefits only about 36% of all infants are exclusively breastfed worldwide until approximately the age of 6 months (Victora et al., 2016; Vila-Candel et al., 2019). Many regions have different trends in the percentage of infants age 0–5 months exclusively breastfed, Eastern and Southern Africa 56%, South Asia 57%, North America 35%, and Eastern Europe and Central Asia 42% (UNICEF, 2019). Several reasons for this early cessation have been identified, including hypogalactia, newborn weight gain below the recommendations, lower maternal age, low socioeconomic status, level of maternal education, BF knowledge, intention, and self-efficacy BF, an unsupportive working environment, and negative maternal expectations or experience with BF (Oribe et al., 2015; Ramiro González et al., 2018; Vila-Candel et al., 2019). The influence of health literacy (HL) on BF practices are yet to be explored by researchers (Tsai et al., 2015). The WHO defines HL as “the cognitive and social skills which determine the motivation and ability of individuals to gain access to, understand and use information in ways which promote and maintain good health” (World Health Organization, 1998). A range of HL measurement tools are available (Sørensen et al., 2012). However, most tools do not reflect the multidimensional nature of HL, as they are predominantly focused on reading comprehension, pronunciation, and numeracy (Haun et al., 2014). Low or inadequate HL is linked to poor health and clinical outcomes, including increased hospitalisations, emergency department use, poor overall health status, and higher levels of mortality (Batterham et al., 2016). Given the significant burden of low HL on citizens and users of the health and social care system (Castro-Sanchez et al., 2018), on identifying and addressing this modifiable risk factor has been emphasised to improve health outcomes (Kilfoyle et al., 2016). A factor that conditions behaviours that afford positive benefits towards BF is HL (Batterham et al., 2016; Castro-Sánchez et al., 2016). Level of education has been associated with HL (Garcia-Codina et al., 2019; Sørensen et al., 2015), with clear evidence of the close and direct association between education level and health outcomes (Castro-Sanchez et al., 2018; Van Der Heide et al., 2013). However, low levels of HL were closely related to poorer health status, even in populations with high levels of education. Even though HL and its relationship with educational levels have not been clarified in previous literature (Van Der Heide et al., 2016), have strongly recommended that HL should be part of the equation to evaluate the manner in which health information is to be handled in the target population. In view of the importance of maternal HL and BF, the definition of predictors and risk factors could make an essential contribution to improved maternal and infant health (Khorasani et al., 2017). A limited number of studies have evaluated the relationship between maternal HL and BF behaviours. Few studies have explored the relationship between maternal HL and BF behaviours, suggesting a positive association between the HL of the mother and the duration of BF (Barnes et al., 2018; Kaufman et al., 2001). 3 | THE STUDY 3.1 | Aims The present study was designed to investigate the association between HL and exclusive BF at 4-months postpartum. The rate of exclusive BF at 6 months in Spain is very low, making it difficult to obtain a sufficient number of mothers for a robust analysis power. Such studies are warranted before endorsing HL-based interventions aiming to mitigate early cessation exclusive BF. Impact: Low or inadequate HL is linked to multiple poor health and clinical outcomes. We investigated the prevalence of exclusive BF at 4 months postpartum, and the impact of HL in maintaining optimal exclusive BF practices. Limited or inadequate HL was one of the factors associated with early cessation of exclusive BF in the multivariate regression model, although further research is needed. KEYWORDS breastfeeding, breastfeeding abandonment, breastfeeding duration, early cessation, exclusive breastfeeding, health literacy, nursing, women | 705 VILA-CANDEL Et AL. (Cabedo-Ferreiro et al., 2019). Furthermore, to assess the prevalence of EBF at 4 months postpartum and identify the variables associated with low and high HL. 3.2 | Design This study used a descriptive, cross-sectional, questionnaire-based design. 3.3 | Participants The catchment area of the Health Department comprises of 251,000 inhabitants, with about 1,700 births per year. La Ribera University Hospital is a state-funded tertiary hospital with 300 beds. The portfolio of health services offered to citizens in Spain depends on each of the 17 Autonomous Regions, and their ability to independently tailor health resources to the needs of the local population. The population was defined as women in their postpartum period (24–48 hr after delivery) who were followed-up until their infants reached 4 months of age. We selected women in their postpartum period using a probability sampling with systematic monitoring. The research team selected to interview two women at random per week from those who were inpatient. Convenience sampling was performed, with the following inclusion criteria: (a) nursing mothers in their immediate postpartum period, (b) women without cognitive problems, language difficulties or those who were illiterate, and (c) single births at term to facilitate comparison with previous studies (Vila-Candel et al., 2017). Women could not be contacted in the telephone interview after three attempts were excluded during follow-up visits from the study. During the first interview at the hospital, consent forms and personal contact information for further follow-up appointments were obtained from candidates who were willing to participate. The sample size was estimated by EPIDAT 3.1 with the following settings: assuming a risk of attrition of EBF in the group of patients with adequate HL of 40% (Rius et al., 2014), versus 60% among patients with inadequate HL, with equally sized groups, a power of 85% and a 95% level of confidence. The calculated sample size required was 222 participants. 3.4 | Data collection Data were collected from participants using a paper questionnaire. The researchers recruited convenience samples of women from hospital from January 2018 to the end of December 2018. The HL screening tool was self-administrated and took approximately 10 min to complete. Feeding type was classified into exclusive BF including expressed milk or milk from a donor (Oribe et al., 2015; Winkvist et al., 2015) or formula feed, either alone or mixed (if feeding combined BF and formula feed). Breastfeeding status was recorded at the time of hospital discharge (48–72 hr), and at 1, 2, and 4 months postpartum. Early cessation of EBF was classified as cessation before 4 months postpartum (yes/no). Subsequently, the community midwife at the healthcare centre, who already had two established postpartum control appointments with the mothers (at 1 & 2 months postpartum), gathered information regarding their BF status. At 4 months postpartum, a telephone interview was conducted, chosen to minimize the risk of missing data and maximize data accuracy, as this was the period where women scheduled an appointment with their paediatrician and the paediatric nurse responsible for the care of the newborn for routine immunization. 3.5 | Ethical considerations The study adhered with the recommendations of the Declaration of Helsinki and was approved by the Clinical Research Ethics Committee of La Ribera University Hospital in November 2017 (HULR_11/2017/#43). All participants were fully informed about the study, the voluntary nature of participation, and confidentiality. The research team recruited the women at the hospital and written informed consent was obtained in all cases, and guaranteed anonymity and confidentiality. 3.6 | Data analysis Descriptive statistics were reported as means and standard deviations for continuous variables (age, age at first pregnancy, gestational age, birth weight, LATCH score, duration of BF), and as ranges and percentages for categorical variables. The exposure variable of the literacy level measured by NVS was analysed categorically. The categories ‘Inadequate’ and ‘Limited’ were collapsed into a single category. Regarding BF status from hospital discharge, the response variable “EBF cessation at 4 months” was divided into two categories: “EBF at 4 months” and “gives up BF at 4 months”, which included either formula feeding or mixed feeding. We analysed the level of HL with the NVS screening tools and the characteristics of the women who gave up BF at 4 months via 2 × 2 tables and the chi-squared test (χ2) for qualitative variables. Moreover, the comparison of averages was made with the application of the Student t-test for quantitative variables. Likewise, follow-up dropouts were analysed over the 4 months to determine any differences between groups. To analyse the magnitude of the association between variables, we generated multivariate logistic regression models to explain exposure variables and EBF cessation at 4 months. These models were adjusted using a stepwise variable selection process based on a likelihood ratio (LR) test. The results present adjusted odds ratios (OR), confidence intervals (CIs) calculated for a confidence level 706 | VILA-CANDEL Et AL. of 95%, and associated p-values. Data analysis was performed on SPSS v.25.0 statistical package (IBM Corp. Released 2018. IBM SPSS Statistics for Windows, Armonk, NY, USA) and R (R project 2019, Version 3.5.1). Statistical significance was considered for p < 0.05. 3.7 | Validity, reliability, and rigour Data collection included demographic characteristics and, HL evaluation tools. LATCH and, other gynaecological data were obtained from electronic medical records. 3.7.1 | Early cessation of exclusive breastfeeding In all routine visits, the mothers were asked about the early cessation of EBF and duration of EBF since their discharge from hospital (1. Are you BF your baby without the help of any formula feeding? 2. If not, for how long have you been exclusively BF your baby?). Finally, the information related to the type of feeding and duration was registered in the personal electronic medical health record. 3.7.2 | Breastfeeding LATCH score The BF LATCH score serves to predict EBF success up to 6 weeks postpartum (Sowjanya & Venugopalan, 2018) and provides a systematic method to gather information about individual BF sessions. LATCH score assigns a numerical score of 0, 1, or 2 to five key components of BF for a possible total score of 10 points: “L” is for how well the infant latches onto the breast, “A” is for the amount of audible swallowing noted, “T” is for the mother's nipple type/condition, “C” is for the mother's level of comfort, and “H” is for the amount of help the mother needs to hold her infant to the breast (Jensen et al., 1994). LATCH score showed that correlations were strong and positive for each item and total LATCH score. The Spearman correlation coefficients ranged from 0.65– 0.91 (Altuntas et al., 2014). We assessed the LATCH score in the maternity unit daily during admission and included it in our data upon discharge (Tornese et al., 2012). 3.7.3 | Health literacy screening tool The HL was explored through an interview at discharge using the Newest Vital Sign (NVS) (Weiss et al., 2005). This tool explores reading and numeracy using a set of six questions based on an icecream's nutritional information label to the respondent. Participants were not briefed about the type of answer expected nor was extra time provided. The total score (0–6 points) categorizes individuals as having a strong probability of limited literacy (score: 0–1 points), possible limited literacy (score: 2–3 points) or adequate literacy (score: 4–6 points) as per authors instructions. NVS scale has also been validated for the Spanish speaking population, with moderate reliability (Cronbach α = 0.69). 4 | RESULTS 4.1 | Demographic characteristics Regarding the sociodemographic characteristics of the participants, the mean (SD) age was 32.5 (5.2) years. At Table 1 can be found details of the distribution according to the HL screening tool used. The mean (SD) LATCH score on the day of discharge in women with EBF and mixed BF was 8.95 (0.95). A third (38.8%, N = 89) of the women planned to use formula feeding immediately after delivery, as the only continued type of feeding. A total of 278 women were initially selected through the inclusion criteria for the study. Two hundred and twenty-nine mothers completed the interviews at discharge (82.3%) and 187 mothers at infant age 4 months (65.1%), as presented in Figure 1. We analysed the homogeneity among those who agreed, and those who declined to participate. There were no significant differences in age, age at first pregnancy, gestational age at delivery, or country of origin between the group that accepted and the group that declined to participate. Regarding the women lost to follow-up, statistically significant reductions in the sample population were recorded at one, two, and 4 months among the participants with lower educational status (p = 0.011; p = 0.017; p = 0.019, respectively) and lower HL levels as measured by the NVS (p = 0.019; p = 0.035; p = 0.016, respectively), presented in Figure 2. 4.2 | Prevalence of exclusive breastfeeding The time course of EBF over follow-up is presented in Figure 3. The prevalence of EBF at the time of hospital discharge was 55.0%, 95% CI (48.5–61.4) versus 46.0%, 95% CI (26.5–38.6) at 1 month, 39.0%, 95% CI (20.8–32.2) at 2 months, and 25.6%, 95% CI (16.1–26.6) at 4 months. In turn, 48.3%, 95% CI (14.1–24.2) of the women showed early cessation of EBF. 4.3 | Variables related to health literacy The NVS screening tool yielded scores between 0–6. These scores were categorized as indicative of inadequate (9.6%, N = 22), limited (45.4%, N = 104), and adequate HL (45%, N = 103). Table 1 presents the results obtained with the HL levels analysed in relation to the NVS tool (Adequate/Inadequate or Limited). We identified statistically significant differences in terms of the level of education (p < 0.001) and maternal employment (p < 0.001). The multivariate logistic regression model for inadequate or limited HL suggested that level of education and maternal employment were associated | 707 VILA-CANDEL Et AL. TABLE 1 Characteristics of the study sample according to the health literacy screening tool (NVS) (N = 229) Characteristics NVS Total row Adequate Limited or inadequate χ2, dfa p-valueN N (%) N (%) Civil status (N = 228) With partner/separated/ divorced 74 32 (43.2) 42 (56.8) 2.9, 1 0.087 Married 154 89 (57.8) 65 (42.2) Level of education Primary school 84 19 (22.6) 65 (77.4) 41.4, 2 < 0.001 Secondary school 90 41 (45.6) 49 (54.4) University 55 43 (78.2) 12 (21.8) Employment status Self-employment/Higher professional/Managerial employment 22 13 (59.1) 9 (40.9) 22.5, 2 <0.001 Employee 132 73 (55.3) 59 (44.7) Student/Unemployed 75 17 (22.7) 58 (77.3) Country of origin Spain 211 98 (46.4) 113 (53.6) 2.3, 1 0.126 Foreign 18 5 (27.8) 13 (72.2) Partner employment status Self-employment/Higher professional/Managerial employment 32 17 (53.1) 15 (46.9) 2.4, 2 0.301 Employee 175 79 (45.1) 96 (54.9) Student/Unemployed 22 7 (31.8) 15 (68.2) Parity Nulliparous 124 58 (46.8) 66 (53.2) 0.3, 1 0.553 Multiparous 105 45 (42.9) 60 (57.1) Medical risk factors during pregnancy None/Low risk 181 87 (48.1) 94 (51.9) 5.1, 2 0.075 Pre-gestational/ gestational diabetes/ Pre-eclampsia 22 5 (22.7) 17 (77.3) Thyroid pathology/ARTb 26 11 (42.3) 15 (57.7) Type of delivery (N = 168) Eutocic 91 44 (48.4) 47 (51.6) 2.3, 2 0.310 Instrumented 37 23 (62.2) 14 (37.8) Urgent cesarean section 40 23 (57.5) 17 (42.5) Type of breastfeeding at discharge (N = 229) EBFc 126 62 (49.2) 64 (50.8) 2.7, 2 0.259 Formula feeding 89 34 (38.2) 55 (61.8) Mixed feeding 14 7 (50.0) 7 (50.0) Type of breastfeeding at 1 month (N = 198)d EBFc 74 47 (63.5) 27 (36.5) 10.7, 2 0.005 Formula feeding 100 39 (39.0) 61 (61.0) Mixed feeding 24 10 (41.7) 14 (58.3) (Continues) 708 | VILA-CANDEL Et AL. with limited or inadequate HL. In addition, maternal age was identified as a confounding variable. Women who were unemployed or currently studying were more likely to have inadequate HL as compared with self-employed/higher professional/managerial employed women, all presented in Table 2. 4.4 | Variables related to early cessation of exclusive breastfeeding at 4 months On considering the characteristics of the women and the referred early cessation of EBF before 4 months, we observed statistically significant differences in the multivariate logistic regression model performed (N = 88). The area under the ROC curve for early cessation of BF was 0.856 (95% CI [0.777–0.935]; p < 0.001). Regarding the quantitative variables, only the BF LATCH score at discharge was associated to early cessation of EBF (p = 0.002), with a higher average score 9.3, (SD 0.64) versus 8.7 (SD 1.01) among the women who stopped EBF. Civil status (p = 0.001), risk of pregnancy (p = 0.002), and type of delivery (p = 0.018) were associated with cessation. The HL level as per the NVS tool was not statistically significative with the early cessation of EBF but the association was high (OR = 2.6; 95% CI [0.837–8.553]; p = 0.097) and clinically relevant, as presented in Table 3. 5 | DISCUSSION This is one of the most comprehensive studies published to date in Spain on BF initiation, maintenance, and predisposing factors for EBF at 4 months postpartum, and on the overall influence of HL upon these behaviours. Based on the results of the present study, even though there was no significant relationship between the levels of HL of nursing mothers with the EBF at 4 months, although their association could be plausible. The prevalence of EBF at 4 months was slightly higher than what was reported previously in Spain (Rodríguez-Pérez et al., 2017), although lower than in other cohorts in northern Spain (51.4%–62.5%) (Oribe et al., 2015; Ramiro González et al., 2018). These differences could be attributable to the methodologies used for data collection, or the scope of BF policies of the different organizations (DíazGómez et al., 2016). Additionally, we need to consider the impact of follow-up losses on the true prevalence of EBF. In our study, over one-half of all the women were using formula feeding at 4 months. It could, therefore, be inferred that a 13.5% of the women had opted for formula feeding due to EBF failure, and this was in line with our previous study (Vila-Candel et al., 2018). While BF initiation rate was high, the continuation rate declined subsequently. The EBF rate was still sub-optimal and do not meet the global public health recommendation set by WHO, and this may indicate that additional measures are needed (De Roza et al., 2019). Health literacy has been identified as a critical and modifiable factor to improve health outcomes and reduce health disparities (Hoffman et al., 2017). We found that while the NVS tool classified 45% of the women as having adequate HL, this figure was higher than previously reported across the board (Cheong et al., 2018; Gazmararian et al., 2012; Komenaka et al., 2015; Poorman et al., 2014). Factors associated with HL were level of education and employment status. Since education is a crucial HL predictor, Characteristics NVS Total row Adequate Limited or inadequate χ2, dfa p-valueN N (%) N (%) Type of breastfeeding at 2 months (N = 191)e EBFc 60 41 (68.3) 19 (31.7) 13.6, 2 0.001 Formula feeding 110 43 (39.1) 67 (60.9) Mixed feeding 21 9 (42.9) 12 (57.1) Type of breastfeeding at 4 months (N = 187)f EBFc 48 32 (66.7) 16 (33.3) 9.7, 2 0.007 Formula feeding 120 49 (40.8) 71 (59.2) Mixed feeding 19 11 (57.9) 8 (42.1) EBFc cessation at 4 months No 46 32 (69.6) 14 (30.4) 3.9, 1 0.046 Yes 43 21 (48.8) 22 (51.2) aχ2, df: chi-square test, degrees of freedom. bART: assisted reproductive technology. cEBF: exclusive breastfeeding. dLost to follow-up at 1 month = 31. eLost to follow-up at 2 months = 38. fLost to follow-up at 4 months = 42. TABLE 1 (Continued) | 709 VILA-CANDEL Et AL. this could explain the relatively weak discriminating capacity of the NVS scale (Delanoë et al., 2016). In general, low HL was also correlated with lower socioeconomic or employment status (Sørensen et al., 2013). Accordingly, mothers with a higher education level tended to initiate BF more often and were likely to breastfeed for a more extended period of time, compared with their less-educated FIGURE 1 Flowchart of patient selection and study follow up 278 women were recruited 229 women were included Women at 1 month. 198 Women at 2 months. 191 Women at 4 months. 187 31 women did not answer to the 1 month follow up call. 7 women did not answer to the 2 months follow up call. 4 women did not answer to the 4 months follow up call. 49 women were excluded: 20 declined to participate 4 were absent in the room 6 declared to be busy 3 feel too unwell 16 declared language barrier FIGURE 2 Distribution of dropouts during the follow-up period of EBF between HL levels according to the NVS tools (N = 42) 35% 30% 25% 20% Percentage 15% 10% 5% 0% 1 month 2 months 4 months Adequate HL Limited HL Inadequate HL 710 | VILA-CANDEL Et AL. counterparts (Ramiro González et al., 2018; Whipps, 2017). Thus, our findings confirm the existence of relationships between educational level and HL, as per other studies (Van Der Heide et al., 2013). Among the women opting for early cessation of EBF, about half had limited or inadequate HL according to the NVS. Some authors have reported that 30% of women with inadequate HL have never breastfed, as compared with 13% of women with adequate HL (Poorman et al., 2014). Other studies based on the REALM (Rapid Estimate of Adult Literacy in Medicine) tool have concluded that a high HL is associated with EBF at two months (Kaufman et al., 2001), and with a significantly high likelihood of BF continuation at postpartum (Stafford et al., 2016). Other authors have observed an association between maternal HL and BF self-efficacy (Khorasani et al., 2017). Furthermore, no significant relationship between HL levels and BF pattern has been reported as per other studies (Mirjalili et al., 2018). In this work, single, separated, divorced, or widowed women, at risk in pregnancy, with limited or inadequate HL, with spontaneous delivery and with low LATCH values at discharge were associated with the early cessation of EBF at 4 months. In our study, a stable relationship was associated with more extended periods of BF and was attributed to partner support of BF. This support has been previously associated with increased rates of initiation and duration of BF (Leng et al., 2019; Rempel et al., 2017). Besides, women at increased antenatal obstetric risk were less likely to opt for EBF at 4 months. The underlying reasons could be clinical or psychological, as seen in limited EBF and assisted reproduction (Barrera et al., 2019), or hypertensive pregnancy disorder (Groer et al., 2013). Nevertheless, the observation could be related to increased levels of anxiety, resulting in a shortened EBF period (Ystrom, 2012). Several studies have confirmed that birth by caesarean section significantly hinders BF in the first hour of life (Hobbs et al., 2016). However, mothers who successfully start BF after a caesarean section procedure are as likely to continue EBF at 6 months, as compared with mothers who give birth via vaginal delivery (Kiani et al., 2018; Lau et al., 2015). Regarding HL levels, women with higher scores could have increased awareness of the importance of BF, have better access to resources, or FIGURE 3 Distribution of type of breastfeeding between discharge to the fourth month (N = 229) 60% 50% 40% 30% 20% 10% 0% EBF Formula feeding Type of feeding. Mixed feeding At discharge 1 month 2 months 4 months Percentage TABLE 2 Odds ratios and 95% confidence intervals of the multivariate logistic regression analysis of factors related to inadequate health literacy (NVS) Factors NVSa OR (95% CI) p-value Age 1.037 (0.974–1.105) 0.252 Civil status With partner/ separated/divorced – Married Level of education Primary school 1 Secondary schoolb 0.393 (0.194−0.797) 0.01 Universityb 0.085 (0.035−0.207) <0.001 Employment status Self-employment/ Higher professional/ Managerial employment/ Employee 1 Student/Unemployed 3,723 (1.854–7.475) <0.001 Country of origin Spain – Foreignb Note: Adjusted odds ratio calculated using the stepwise selection method based on the likelihood ratio test. aArea under the ROC curve for inadequate health literacy according to NVS = 0.775, 95% CI (0.715–0.834); χ2 = 58.774; p < 0.001; N = 126. bVariable differs significantly between type of health literacy at p < 0.05. | 711 VILA-CANDEL Et AL. be surrounded by other women who would help them address difficulties (Díaz-Gómez et al., 2016; Vila-Candel et al., 2017). This seems to support our hypothesis, as women with low HL have lower EBF rate at 4 months postpartum than women with adequate HL. We analysed the HL and sociodemographic characteristics of women who were lost to follow-up. Women with lower educational and HL levels were at a higher risk of dropping out of abandoning the study. Healthcare professionals should thus tailor local resources to support these women. The use of tools including the LATCH score at discharge or HL screening questionnaires before delivery (Tornese et al., 2012), could identify women who may be more hesitant to access health services or seek specialized professional help (SorianoVidal et al., 2018). 5.1 | Limitations Our results related to EBF, while far from excellent, improve upon other comparable studies (Oribe et al., 2015; Rius et al., 2014). Nevertheless, our study has limitations. Firstly, regarding the multivariate model constructed to determine the factors that explain the early cessation EBF, only 88 subjects could be included in the model. This size to build a model that identifies effects between variables is scarce. Secondly, the magnitude of the OR obtained should be considered in the light of the limited sample size, corresponding to women with low HL and early cessation of EBF. This could have led to overestimating the likelihood of EBF cessation. However, factors associated with cessation would not be influenced by this limitation. On the other hand, we pragmatically selected and used NVS tool based on their soundness and experience. While there are more than 40 tools available, there is still a paucity of HL research conducted in Spain (Karnoe & Kayser, 2015), with no tools validated in this country. The only validated tools to date are referred to Spanish-speaking Hispanics (Lee et al., 2010; Weiss et al., 2005). In these cases, some authors recommended adjusted linguistic validation procedures. Unfortunately, such validation has been made for Spanish spoken in Spain, with no transcultural validation or adaptation to the pregnant women (Acquadro et al., 2008). 6 | CONCLUSION Multiple well-known bio-psycho-social and economic factors influence decisions about BF, such determinants should already be considered by BF support and promotion programs. Our study offers preliminary evidence regarding the hitherto inconsistent relation between HL and early cessation EBF at 4 months, supporting the conduct of further studies with larger sample sizes and greater statistical power. Such studies are warranted before endorsing HL-based interventions aiming to mitigate early cessation EBF. On the other hand, health literacy would contribute to and promote person-centred maternity health care and would result in multiple other benefits for women and infants. ACKNOWLEDGEMENTS The authors acknowledge the support of all the nurses and midwives of the Obstetrics Department at La Ribera University Hospital who helped to screen suitable participants for the study, and all the participant women. CONFLICT OF INTEREST No conflict of interest has been declared by the authors. AUTHOR CONTRIBUTIONS VILA-CANDEL, Rafa and MENA-TUDELA, Desirée designed the work. VILA-CANDEL, Rafa; SORIANO-VIDAL, Francisco Javier and, QUESADA, José Antonio acquired, analysed, and interpreted the data. VILA-CANDEL, Rafa and MENA-TUDELA, Desirée wrote a draft. All authors reviewed the content of the paper with significant intellectual contributions. All authors approved the final version for publication. TABLE 3 Odds ratios and 95% confidence intervals of the multivariate logistic regression analysis of factors related to early cessation of breastfeeding at 4 months (N = 88) Factors OR (95% CI) p-value LATCHa scoreb 0.321 (0.155−0.666) 0.002 Civil status With partner/ separated/divorced 1 Marriedb 0.105 (0.028−0.392) 0.001 Medical risk factors during pregnancy None/Low risk 1 High riskb 19.017 (3.056–118.342) 0.002 NVS categories Adequate 1 Limited or Inadequate 2.675 (0.837–8.553) 0.097 Type of delivery Eutocic 1 Instrumented 0.397 (0.106–1.487) 0.171 Urgent cesarean section 0.154 (0.033−0.724) 0.018 Note: Adjusted odds ratio calculated using the stepwise selection method based on the likelihood ratio test. Area under the ROc curve for early cessation of breastfeeding = 0.856, 95% CI (0.777–0.935); χ2 = 40.650; p < 0.001; N = 88. Analysis carried out on 88 participants with information on all variables. One participant excluded from analysis due to incomplete data civil status. aLATCH: LATCH assessment tool. bVariable differs significantly between type of feeding at 4 months at p < 0.05.