Study setting and design
This research took place in three health centres located in Mekelle, Ethiopia. Mekelle is a large city located in north Ethiopia. Mekelle has nine health centres, a tertiary hospital, and three general hospitals. This paper reports on baseline data from a quasi-experimental study, designed to test the effectiveness of a SMS-based breastfeeding education intervention targeting expectant couples attending antenatal care (ANC) in public health centres in Mekelle city, Ethiopia.
Source and study population
Three health centres with the highest number of mothers attending ANC follow-up were purposively used as recruitment sites. A total of 293 mothers attending their antennal care (ANC) in these three public health centres were approached by nurses either in person at the health centre during the ANC appointment or via a phone call to check eligibility criteria. Eligible couples satisfied the following inclusion criteria: able to read and understand the local language (Tigrigna), living in a union, mother had no known medical issues that could hinder breastfeeding, there were no known issues with the foetus or pregnancy and both members of the couple were able to provide a written informed consent. Based on these criteria, 128 expectant couples were included in this study.
Data were collected through a face-to-face interview between September and October 2018 by trained nurses working in the health centre. A modified version of the cross-cultural adaption process was utilised, in which the questionnaires were translated to Tigrigna, after which the Tigrigna versions of the questionnaires were back-translated to English by two public health nutrition experts from Mekelle University . Finally, face validity to check for understanding and language was conducted for all questionnaires with fathers and mothers who had children under two years of age but not reliability and validity assessment.
Variables and measurement
Questions on father and mother characteristics such as age, educational status, income, and employment status; and pregnancy and childbirth-related variables such as parity, ANC provider, number of ANC appointments, breastfeeding information during ANC, breastfeeding experience, maternity leave provision, and breastfeeding intention were developed based on the literature [12, 13, 17, 22, 38].
Breastfeeding knowledge/awareness was assessed using a questionnaire adopted from the Food and Agricultural Organization (FAO) of the United Nations (UN) . This questionnaire has ten open questions, which were later coded into Yes or No responses, based on the protocols. Each correct answer was scored, responses were totalled, and the percentage of correct responses recorded. Breastfeeding attitudes were measured using the Iowa Infant Feeding Attitude Scale (IIFAS). This tool has 17 questions, and uses a five-point Likert scale ranging from 1= strong disagreement to 5= strong agreement . Out of the 17 questions nine were reverse scored, thus, these responses were recoded before calculating the total attitude score. The total score was calculated out of 85, with a minimum of 17 and maximum of 85. The questions are non-gendered and can be asked to men and women without modification.
The Partner Breastfeeding Influence Scale (PBIS) was used to measure perceived breastfeeding support and was assessed using five dimensions of partner breastfeeding support (Savvy (Cronbach alpha: men =0.87, women=0.82), helping (Cronbach alpha: men =0.79, women=0.82), appreciation(Cronbach alpha: men =0.86, women=0.84), breastfeeding presence (Cronbach alpha: men =0.88, women=0.82), and responsiveness (Cronbach alpha: men =0.77, women=0.76)). Mean scores were calculated from all scores, from 1 (extremely not supportive) to 5 (extremely supportive) for each breastfeeding support component .
The data was analysed using IBM SPSS Statistics version 23 (IBM Crop. Released 2015. IBM SPSS Statistics for Windows, Version 23.0. Armonk, NY: IBM Crop). After data cleaning and coding, descriptive statistics were conducted. Sociodemographic characteristics were presented using frequency and percentage or mean with standard deviation. Normality tests were performed for each continuous variable using skewness test. Based on these tests further analyses were selected. Independent T-test or ANOVA, and Mann-Whitney U Test or KRUSKAL Wallis test were used for normal distribution and non-normally distributed data, respectively. Once the explanatory variables were fitted with the dependent variable (gender) variables with p-value <0.05 were considered as having significant difference among mothers and fathers. Reporting follows the STROBE guidelines for cross-sectional studies.