Study site and population

Central Women’s Hospital, Yangon was purposefully selected because it was the largest public hospital providing free women health services and provided a setting where we could recruit women from various walks of life. Secondly, as our intervention uses a mobile phone platform and this setting currently has the best mobile network coverage in Myanmar. The average antenatal care clinics attendance was approximately 2,000 pregnant women per month [12]. During a preliminary site assessment, we observed that the majority of antenatal care clinic attendees were using mobile phones and from the hospital records appeared to live in areas with high mobile network coverage [12].

The study population comprised of selected pregnant women seeking antenatal care at the Central Women’s Hospital, accompanying family members and service providers.

Sampling

We applied purposive sampling and selected the pregnant women from the antenatal care clinic attendance register the Central Women’s Hospital. In general, a pregnant woman receives her antenatal care card after registration where basic information are recorded and updated at each visit. We selected women who were having their first pregnancy, had no known complications, had a single fetus, were 24 to 28 weeks of gestation, and lived in Yangon or nearby cities that had mobile phone network coverage. Other selection criteria were an absence of serious known illness that might potentially hinder breastfeeding, completed at least primary school education, and had access to a mobile phone. To explore the extent and nature of her husband’s support, gender roles and decision making power, we conducted in-depth interviews with accompanying husband. If husband was not available, mothers or mother-in-law were interviewed. For institutional staff, we included multiple stakeholders working in infant and young child feeding sector and purposively selected the respondents based on their expertise and areas of work.

We conducted ‘In –depth-interview’ (IDIs) with fourteen pregnant women, four husbands, one mother and one mother-in-law. One ‘Focus Group Discussion’ (FGD) was conducted with 8 nurses on duty (staff nurse, trained nurse and midwives), and KII with an obstetrician and a medical officer from the antenatal clinic at the hospital; program staff from National Nutrition Centre, Department of Health; Young Child Survival and Development Section, United Nations International Children’s Emergency Fund (UNICEF); Nutrition Section, Save the Children; Myanmar Maternal and Child Welfare Association (largest national NGO providing maternal and child health service,) and Ooredoo mobile company, one of the two recently launched mobile phone operators in Myanmar.

Data collection

We used semi-structured in-depth interviews (IDI), key informant interviews (KII) and focus group discussions (FGD) to collect the data. Before starting the interviews, separate guidelines were developed for all IDIs and FGDs and later translated into Burmese language. In February, 2014, the interview guidelines were pretested (by MPH) with three pregnant women and three accompanying husbands at the study hospital. Brainstorming sessions with UNICEF and Save the Children staff were conducted and guidelines were revised accordingly based on their feedbacks. The data collection took place during June and July, 2014. With the help of the hospital staff and nurses, MPH filtered the potential respondents based on the hospital attendance registry book to assess the eligibility, approached the selected woman individually, explained the nature of study, provided the ‘information statement and consent form’ written in Burmese and obtained a signed consent form. If the respondent woman came with accompanying family member, MPH approached him or her and conducted an interview after receiving informed consent. We also obtained consent to record interviews with a digital recorder. The data was coded as numbers, reported as anonymous, participant’s confidentiality is assured and individual cannot be identified in reporting.

The major topics explored with the pregnant women and their family members included their socioeconomic background, perceptions and sources of information about infant feeding, women’s intention to breastfeed, and the expected barriers and enablers to practice exclusive breastfeeding. We also elicited women’s understanding on infant feeding practices among their friends and families. As the intended intervention uses mobile phone communications, we explored the women and their family’s access to mobile phones, their skills on using a mobile phone, their perceptions about the intervention, the preferred method of receiving the educational messages (Short Message Service (SMS) or direct phone calls) the timing and frequency of receiving the educational messages. The guidelines for KIs and FGDs with institutional health staff included views and suggestions on current infant and young child feeding practices, barriers for appropriate breastfeeding practices, and perceptions and recommendations on the planned intervention. Interviews with the mobile operator explored mobile network coverage plan, possible challenges and feasibility of implementing a mobile communications intervention and the potential to expand to rural populations in the future.

Data analysis

All digitally recorded interviews were transcribed verbatim in Burmese by a local typist in Microsoft Word. A sample of transcript were as reviewed line by line by one of the authors (MPH) to check for accuracy and translated into English. These samples were then checked by AA for further quality check for the translation and interviewing process. Two of the authors (MPH and AA) discussed two initial transcripts to initiate the development of the codes. Subsequently all transcripts were carefully read by MPH to develop a list of thematic codes, which were independently reviewed and verified by AA. The data were manually coded by MPH for emerging themes, which was again verified by AA. Subsequently the text pertaining to each thematic code was discussed and summarised in a document that presents the findings for each theme using quotes and tables. Themes were triangulated using data collected through interviews and FGDs with different types of respondents.

Demographic characteristics of the women

The pregnant women participated in this study were mostly in their mid-twenties. All of them were between 24 to 28 weeks of gestation, the majority were from Bamar followed by Chin, Kayin, and Indian and Chinese ethnicity. All women came from nuclear families and majority were Buddhist and lived in urban slum area. They had varying levels of education. Among 14 pregnant women, majority of them (42%, n = 6) completed middle school, one fifth (n = 3) completed primary school, 15% (n = 2) completed high school and 15% (n = 2) were graduate while a woman was a university student (distance learning). Among interviewee women, half of them were employed and working as vendors, casual workers or government staff.

Perceptions of infant feeding and sources of information

Infant feeding practices in the community and influencing factors

All women noticed all mothers in their community fed breast milk to their babies while only one woman knew of a mother practiced exclusive breastfeeding for six months successfully. They felt that underlying socio-economic factors played an important role and work was the main reason for introducing infant formula. Two respondents from Ayeyarwady region (one of the most populous areas in Myanmar and occupy the delta region of the Ayeyarwady River) mentioned that in their villages, honey was given soon after delivery by traditional birth attendants due to cultural belief-for example- “the child needs honey to freshen up himself as he /she stays inside mothers’ womb for a long time”. Mixed feeding was the most commonly seen feeding pattern and water was found to be given as early as less than four weeks after delivery. Formula supplementation was commonly used before child reach 4 months and poor usually used formula made in China and the rich preferred branded infant formula such as Nestle or Dumex. A vast majority of respondents reported of knowing mothers introduced various kinds of soft and semi-solid foods at different age of children mostly at 2–4 months such as pasted rice powder or soft, mushy rice or chewed rice added by oil and salt. They noticed that mothers gave fruity ORS in hot weather, added blended carrots, beans or eggs while meat or dark green vegetables was not commonly given.

Intended feeding patterns and influencing factors for exclusive breastfeeding practice

Respondent use of mobile phones

Opinions about a mobile phone based intervention

In examining service providers opinions on the planned mobile phone based intervention, there were generally positive views on the intervention idea. A few participants mentioned that the study would use the integrated approach by leveraging the power of mobile phones at perfect timing which could enhance the community mobilization to participate in the study. Findings from FGD with nurses reported that sending health promotion SMS could attract mothers than traditional distribution of pamphlets, however younger nurses had higher positive attitudes toward using mobile phone messages than older staff. All respondents agreed that using SMS as a service delivery would be the best modality for current moment and suggested to send SMS at evenings or weekends to avoid mothers’ busy time. Perceived barriers found to be the unreliable mobile networks, affordability for handset, time needed to spend to send SMS, recipients’ willingness to read SMS if there were too many messages, and possibility of changing phone numbers in six months’ time. A staff reported of a potential downside as participants did not receive SMS as claimed to be due to natural disaster or unforeseen circumstances.

Recommendations made were to explore the best timing to send SMS to the selected women; to use simple, easily understandable and culturally appropriate terms, to use relevant ethnic languages such as Shan, Karen or Kachin in some areas; to add interactive or pictorial messages for better attraction. Other suggestions came out as conducting a proper evaluation after intervention, arranging for more media coverage, sharing and advocating for political commitment in order to scaling up of similar design to the national level.

In an in-depth interview with a staff from mobile operator revealed that, “We are aiming to cover 95% of the country in coming 3 years.” He supported the intervention idea in principle and reminded to work carefully in developing text messages in limiting numbers of characters (word count) to be put in a message and to consider and assess participants’ literacy level as reading text message need to have some level of education.

Incorporating findings into intervention design

Mobile phone, or use of SMS text messaging in particular, may provide an opportunity to improve health behaviors. According to the 2011 World Health Organization report, mobile phones provided a new communication channel for health promotion and community mobilization [13]. Several studies have shown feasible and promising results of using mobile phones and text messaging in smoking cessation [14–16], diabetes education [17], sexual health [18], diet and physical activity [19] as well as maternal and child health nutrition [20, 21]. Respondents preferences on the SMS over voice messages is consistent with study from Malawi [22] and we chose to use SMS message. As penetration of mobile phone has increased dramatically recently [23], unlike other countries, study participants were eager to receive health education messages to their mobile phones. In line with participants’ feedbacks, we created messages that were simple, locally acceptable, and culturally appropriate. For example, we included message as “Not to give honey after birth as breastmilk alone is the best for baby” to correctan incorrect belief commonly found in Ayeyarwady region. To be consistent with mobile operator‘s inputs, each message has approximately 160 characters written in Myanmar fonts (Zawgyi font¹). Due to financial and technology constraints, we did not add pictorial images in the planned intervention. In response to respondents’ suggestions, we designed breastfeeding promotional messages to be delivered three times a week (Tuesday, Thursday and Saturday) to the intervention group and non breastfeeding related messages, mainly pregnancy and child care messages, to the control group once a week in the evening. We added two ways SMS communications model to monitor the delivery status of women. We hope that this formative study could contribute in promoting exclusive breastfeeding practices in Myanmar to some extent as promotion of nutrition drew a policy maker’s attention [24]. Perceived barriers provided by the participants and service providers such as mobile network coverage, electricity, mobile handset cost and community SMS familiarity were similar to other study findings [13, 22] and we considered these factors to design the intervention.

The limited understanding of the exact definition of exclusive breastfeeding might reflect the possible main reason for gap in high awareness on breastfeeding benefits and low exclusive breastfeeding practice. Respondents had inappropriate knowledge of nutrient contents in breast milk, misperception that water could be added in practicing exclusive breastfeeding, preferred formula milk or influence of formula milk advertisement. A study with 409 first-time Australian mothers from Western Sydney revealed similar finding. A limited understanding of the term exclusive breastfeeding was also reported in this study. A study in Brazil described that respondents had doubts on the quality of breast milk resulting in the feeding of water to a quarter of children in first month [25–28]. Unpublished Myanmar studies reported that estimated 70% of mothers gave water to their babies before one month. Reasons given for giving water in our study appeared as ‘water was natural, prevent thirst, dry mouth and thrush’ and similar to the findings from Nepal and Vietnam [29, 30]. Women expected their child to be tall, gain weight and be smart and adding formula milk or mixed feeding perceived as the best source for nutrients and convenient for weaning similar to the studies conducted in China, Bangladesh and other countries [31–33], leading to higher odd of failed exclusive breastfeeding for 6 months. We therefore incorporated correct information into the messages development. Example messages were “Breast milk will give all the nutrients your baby need for healthy physical and brain development” “Don’t believe formula milk’s advertisement. Breast milk is the best”. Respondents claimed various sources as information and as per finding from China study, they also claimed of “having limited communications with hospital staff” [31].

Expected barriers to practice exclusive breastfeeding revealed as need to return to work earlier than the baby is six months old, worry of having insufficient breast milk and breast problems, limited skills to solve constraints and grandmothers influences. In general, these findings were in line with other several studies [25, 26, 31, 33, 34]. A secondary data analysis on five East and South East Asia countries supported mothers’ employment as barrier to exclusive breastfeeding [35] and we planned to send frequent messages on how to express and store breastmilk and dangers of formula milk. To empower respondents in dealing with expected breast problems, we incorporated skilled trainings approach on how to do correct position and attachment, manage cracked and sore nipples. Though husband supported breastfeeding, mothers-in-law tended to be the main influential persons for adding water, honey and formula milk. Our findings confirmed previous research showing that grandmothers play a vital role in feeding practices [36, 37]. In contrast, interviews with grandmothers from Nepal reported that they would not give water or pre-lacteal foods [38]. Cultural influences somewhat existed [38] as Chinese and Indian background women reported more of mothers-in-law’s influence compared to other ethnic groups. According to the respondents, potential influencing grandmothers would be those living with them and they could be respondent women or husbands’ mothers. In consistent with a recommendation from a UK study, [36], we added special messages addressing grandmothers.

In developing SMS for intervention group, alongside the findings of this study, we reviewed and utilized the infant feeding literature including UNICEF and World Health Organization (WHO) breastfeeding guidelines [3, 39], Australia breast and infant feeding guidelines [40, 41], Ministry of Health, Uganda breastfeeding counselling messages [42]. Information Education and Communications (IEC) materials focusing on the infant feeding educational messages, produced by the Department of Health, Myanmar [43] and Save the Children, Myanmar [44] and a similar study conducted in China [21, 45, 46]. Developed messages were then translated into Burmese, pre-tested with 15 pregnant women visiting study setting prior to the actual intervention.

This study and the related intervention will be, to our knowledge, a first ever research in Myanmar on using mobile phone as a tool to promote exclusive breastfeeding practice. This paper attempts to provide important insight into the study women’s perceptions, practices and barriers to practice exclusive breastfeeding for six months and theirs opinions and attitudes towards mobile phone based intervention. By incorporating findings, this study leads to the development of an effective randomized controlled trial intervention using SMS as a communications tool and we aim to apply similar study design in other health education promotion interventions in future.

Strengths and limitations

Our study has strengths as the researchers involved in whole process of developing guidelines, conducting interviews, transcribed verbatim and translating bilingual (English and Burmese), resulting in minimizing errors in translation and meaning. Secondly, the study includes views of women, their family members and stakeholders from the public, development and private sectors. The findings have evidence-based relevance for designing an intervention to improve exclusive breastfeeding in Myanmar. However, there are some issues that should be considered when interpreting the findings of this study. In this study, we did not aim to generalize the findings, rather we wanted to focus in prioritizing findings and feedbacks to inform the proposed intervention. The recommendations and findings were related to the planned intervention and has implications for that specific program. Future formative and or feasibility studies would be needed to inform the designs of intervention that targets any specific socioeconomic and ethnic groups in Myanmar or beyond.