The target population consisted of adult males (≥ 18 years) who attended the outpatient department of a district hospital, Free State, South Africa. The hospital is a public hospital in a community of about 181 712 people.¹⁵ The hospital serves as the referral centre for 13 primary healthcare clinics. An average of 70 patients are seen daily at the outpatient department, 21% (n = 15) of whom are adult males.

A convenience sampling method was adopted for this study, and data were collected twice a week. At the time of this study, the estimated number of men ≥ 18 years old living in Botshabelo was N = 26 000 (sample frame).¹⁵ The estimated minimum sample size was calculated as n = 268 (at 90% confidence level [CL] and margin of error [MOE]: 0.05).¹⁶ The study population consisted of consenting men who were at least 18 years old, irrespective of whether they were married or had children. A colour-coded sticker system was used to ensure that no eligible male participant could participate in the study more than once (i.e. a colour-coded sticker was placed on a participant’s folder once the participant had completed the questionnaire).

The development of the structured questionnaire that was used in this study was informed by findings from prior studies,^17,18 after a thorough literature review and content analysis. Search terms, such as EBF, adult males and exclusive breast-feeding and knowledge of and attitudes of men towards EBF, were used to search for relevant articles during the literature review process. The databases used to access articles were Google Scholar, MEDLINE, PubMed, CINAHL, SABINET, Science Direct and Directory of Open Access Journals. Concepts were identified to formulate closed-ended questions, which were answered using an adapted Likert-scale ranking. The questionnaire was self-administered, made available in three languages commonly spoken in the local community (i.e. English, Sesotho and Setswana) and distributed manually. The self-administered questionnaire comprised five distinct sections. Section A of the questionnaire obtained data on participants’ eligibility to participate in the study. The demographic and socio-economic details of the participants were captured in Section B of the questionnaire. Questions in Section C assessed participants’ knowledge of aspects of EBF, whilst data on participants’ attitudes and practices regarding EBF were captured in Sections D and E, using a five-point Likert scale, that is, strongly agree, agree, not sure, strongly disagree and disagree.

Prospective participants were briefed about the study by one of the researchers (O.M.M.). Participants were informed that they could choose to participate in the study, or not, and that failure to participate would not compromise their treatment at the hospital. A detailed explanation about the study (also contained in an information leaflet) was given to participants who gave verbal consent. Participants were also informed that implied consent was being given by agreeing to complete the questionnaire. One of the researchers (O.M.M.) and/or a research assistant assisted participants who were incapable of completing the questionnaire on their own. Participants were instructed to drop the completed questionnaire in a sealed box situated in the front desk at the outpatient department. O.M.M. emptied the box daily and all completed questionnaires were locked in a safe. Data from the questionnaires were inputted on a Microsoft Excel sheet on a computer with password protection. The questionnaire had no trace of identification, to ensure that data were collected anonymously. The study was conducted over a period of two months (August 2020 – September 2020).

We performed a pilot study before the official start of data collection to test the suitability of the study design and methods, the chosen data collection method and the overall structure of the questionnaire. The pilot study consisted of 10 eligible adult men who were selected using convenient sampling. No changes to the structured question resulted from the pilot study. The estimated time needed to complete the questionnaire was 30 min.

Questionnaire items were scored to determine the percentages of correct or expected and incorrect or unexpected responses. The scoring range of the knowledge questions was 10 (maximum) to 0 (minimum). A score of ≥ 70% (7/10) was considered ‘adequate knowledge’, scores between 50% (5/10) and 60% (6/10) were considered ‘average knowledge’ and a score of ≤ 40% (4/10) was regarded as ‘poor knowledge’. Attitude towards EBF was assessed by obtaining participant responses to 16 items using a five-point Likert scale. For the purpose of this study, ‘strongly agree’ and ‘agree’ responses were summed as ‘agree’, whilst the ‘strongly disagree’ and ‘disagree’ responses were summed as ‘disagree’. The expected responses to the attitude items 1–8 were ‘disagree’, whilst the expected responses for items 9–16 were ‘agree’. There were 10 items in the practice section, and participants’ responses were obtained using a five-point Likert scale. The five-point Likert scale was collapsed to three, that is, ‘Agree’, ‘Disagree’ and ‘Not sure’. Participants who scored ≥ 50% (≥ 5/10) were regarded as having ‘good’ EBF practices, and those who scored < 50% (< 5/10) were considered as having ‘poor’ EBF practices. The expected response for practice items 3 and 4 in the practice section was ‘disagree’, whilst the expected response for the other items in this section was ‘agree’.

The data were analysed using Statistical Analysis System (SAS) version 9. Descriptive statistics (e.g. medians) was used for continuous variables, whilst frequencies and percentages were computed for categorical data. The associations between demographic data and knowledge, attitudes and practice scores were assessed using chi-square or Fisher’s exact tests. A p-value of < 0.05 was taken to be significant.

The validity of the structured questionnaire was examined by comparing the questionnaire elements with those of previous and similar studies, as well as by conducting the pilot study. A departmental evaluation committee, consisting of consultant family physicians and a biostatistician, subjected the questionnaire to review and approval.¹⁹

The Health Sciences Research Ethics Committee of the University of the Free State (UFS-HSD2020/0324/2508) granted approval for the study. Further approval was obtained from the Head of the Free State Department of Health.

Two hundred adult males participated, out of an approximate 240 participants who presented during the study period, giving a response rate of 83%. One third (n = 64; 32%) of the study participants were in the age category 31–40 years, with a median age of 37 years (minimum = 18 years; maximum = 65 years). In total, 56 (28%) participants completed high school, whilst 54 participants (27%) had some secondary-level education and 51 participants (25.5%) had undergone tertiary-level education. The majority (n = 188; 94%) were of the Christian faith. Sixty-eight (34%) of the participants were married, 45 (22.5%) of them were single and 42 (21%) were living with their partners. The majority (n = 158, 79%) were of Southern Sotho ethnicity. Regarding employment, 106 participants (53%) had jobs and the majority were skilled/highly skilled. About 50% of them received an annual income ≥ R10 000.00. Of the 200 participants, 148 (74%) had children – the majority had more than one child – and 84 (56.8%) of these fathers had attended antenatal care with the mothers of their children (Table 1).

Table 2 describes the responses of the participants to the EBF knowledge questions. Knowledge was assessed by asking questions about aspects of EBF, such as duration of EBF and the effect of EBF on a baby’s immunity. It was found that 65 (32.5%) participants had adequate knowledge, whilst 52 (26%) and 83 (41.5%) participants had average and poor knowledge of EBF, respectively (Tables 2 and 5).

Participants who scored ≥ 50% (≥ 8/16) of the 16 attitude items were regarded as having positive attitudes towards EBF, whilst those who scored < 50% (< 8/16) were regarded as having negative attitudes towards EBF. Our findings reveal that the majority (n = 153; 76.5%) of the participants had positive attitudes towards EBF (i.e. had scores of ≥ 8/16), whilst 47 (23.5%) had negative attitudes (i.e. scores of < 8/16; Tables 3 and 5).

In total, 151 (75.5%) participants reported good practices in relation to EBF, whilst 49 (24.5%) participants reported having poor practices (Tables 4 and 5).

Table 5 provides a summary of outcomes of the assessment of knowledge, attitudes and practice in relation to EBF.

After having obtained participants’ responses on aspects of knowledge, attitudes and practices regarding EBF, we investigated the association between participants’ sociodemographic characteristics and their knowledge, attitudes and practices regarding EBF through bivariate analysis using chi-square or Fisher’s exact tests. The findings, as presented in Table 6, show that adult men in the age group 18–30 years were more knowledgeable of EBF than the other age groups (p = 0.001). In addition, adult males who had completed Grade 12 and tertiary education had significantly more knowledge than participants with lower levels of education (p = 0.001). We found a significant association between marital status and practice of EBF, as participants who were married had better practices of EBF (p = 0.027). Furthermore, the results show that adult men who had attended antenatal visits with their partners were significantly more knowledgeable and had a more positive attitude towards EBF (p = 0.000 and 0.001, respectively).

As far as we know, this is the first study executed in Botshabelo that investigated the knowledge, attitudes and practice regarding EBF of men. Several limitations of this study should be noted. Firstly, only health-seeking adult men were included in the study. It is possible that there are other adult men in the Botshabelo community with varying levels of knowledge and different attitudes and practices regarding EBF, thus, limiting the generalisability of this study. Secondly, because of time constraints and the advent of the COVID-19 pandemic, only 200 participants were surveyed; the involvement of a greater number of participants could, perhaps, have made the findings more representative of the community. The fact that non-random sampling was used could also be considered as a limitation of this study.