Purpose: Dietary diversity is a key proxy indicator of nutrient adequacy; however, limited studies have been done on it among pregnant women in Ethiopia. The study aimed to examine the prevalence of sub-optimal dietary diversity and its associated factors among pregnant women in Gurage zone, South Central Ethiopia. Materials and Methods: A mixed-method approach, a longitudinal study complemented with an exploratory qualitative study, was conducted. In the longitudinal study, a consecutively included sample of 668 pregnant women was followed in three rounds of survey. Dietary diversity was assessed using the minimum dietary diversity score for women (MDD-W) tool. The average of three dietary diversity scores was used to define overall diversity. Consuming less than 5 of 10 standard food groups was considered as suboptimal dietary diversity. Multivariable logistic regression analysis was used to identify predictors of suboptimal dietary diversity. Qualitative data were analysed using the thematic analysis method. Results: During the 16 to 20, 28 to 29 and 36 to 37 weeks of gestation surveys, 75.0, 78.7 and 76.5% of the women had sub-optimal dietary diversity. In aggregate, 84.4% (95% CI: 81.6, 87.3) of the women had sub-optimal dietary diversity. Rural residents (AOR: 1.91, 95% CI: 1.01, 3.62), women with no formal education (AOR: 5.51, 95% CI: 1.96, 15.53) and from food insecure households (AOR: 2.44, 95% CI: 1.07, 5.59) had higher odds of suboptimal dietary diversity. Women with higher nutritional knowledge (AOR: 0.92, 95% CI: 0.87, 0.98) were less likely to have suboptimal dietary diversity. Food taboos, poor nutritional literacy and pregnancy complications were also reported as factors affecting dietary diversity. Conclusion: Majority of pregnant women in the area had sub-optimal dietary diversity. Improving the socio-economic status and promoting nutrition knowledge may improve women’s dietary diversity.
The Gurage Zone is one of the zones in the Southern Nations, Nationalities and Peoples Region of Ethiopia with its center at Wolkite town located 158 km west of Addis Ababa. In terms of population, it is a densely populated zone in Ethiopia with 441 people per square kilometer. It is composed of 14 districts and 5 town administrations.23 The zone has seven hospitals (two non-governmental and five governmental) and 72 health centers. The main source of food and economic activity is rain-fed agriculture. Common staple foods are enset (Ensete ventricosum), teff in the form of injera, green cabbage, maize and other cereals.23,24 Data were collected through an institution-based longitudinal study supplemented by an exploratory qualitative study. The study population was all pregnant women attending antenatal care (ANC) in selected districts of the Gurage zone. Women in their second trimester (16–20 weeks) with a singleton pregnancy, permanent residents (lived at least 1 year) and aged 15–49 years were included in the study. Pregnant women with known medical conditions like HIV/AIDS and diabetes mellitus were excluded. The participants in the focus group discussion (FGD) were pregnant women, health extension workers and women’s development army. The health extension workers are trained professionals providing essential health services at health post. The women’s development armies are community health workers with no formal professional education but chosen as a leader to serve five households within the same neighborhood based on their clear understanding and practice of health extension packages.25 The key informants selected for in-depth interviews were heads of health centers and heads of district health departments. The sample size was calculated using single population proportion formula by considering the following assumption: 55.2% prevalence of inadequate dietary diversity,13 95% of confidence level and 5% marginal errors. To accommodate for the multistage nature of the study, we applied a design effect of 1.5. Furthermore, 20% was added for compensating possible non-response. Ultimately, a sample size of 684 was reached. There were 6 FGDs, involving pregnant women, health extension workers and women’s development army. Eleven participants, including 3 heads of health centers, 3 heads of the district health office, 2 maternal and child nutrition coordinators and 3 community elders were contacted for in-depth interviews. The final sample of the qualitative study was 55. For the quantitative study, a multistage cluster sampling method was used to select pregnant women. Initially, the study area was divided into rural districts and urban town administrations. Six districts and two town administrations were taken to represent 14 districts and 5 town administrations within the Zone. Two representative healthy facilities were randomly selected from each district. Then, the sample size was allocated proportionally to 16 selected health facilities. The participants who fulfilled the eligibility criteria were consecutively included until the sample size was filled and those samples followed up to the end of pregnancy. For the qualitative study, a purposive sampling method was used to select participants whom we think could provide rich information on the existing maternal nutrition situation in the study area. An interviewer-administered structured questionnaire that was prepared partly by reviewing several published articles13–15,26,27 and partly by adopting standardized data collection tools2,28 was used for quantitative data collection. The questionnaire has different sections on socio-demographic variables, obstetric history, household food insecurity, maternal dietary diversity, maternal nutrition knowledge and practice. The questionnaire was prepared in the English language and translated to the local Amharic language. The consistency was checked by translating it back to English and was edited by a person with good knowledge of both languages. Data were collected by trained and supervised enumerators. The questionnaire was pretested among pregnant women not participating in the actual study but living in a similar setting. The actual data collection was done among women at their 16 to 20 weeks of gestation and maternal dietary diversity score was repeated in two subsequent phases; 28 to 29 weeks and 36 to 37 weeks of gestation. The completeness of data was checked each day at the end of data collection. Incomplete data was traced back and edited accordingly. The dependent variable of the study was maternal dietary diversity measured using the standard FAO’s minimum dietary diversity for women (MDD-W).2 This section of the questionnaire listed ten groups of food items: 1. Grains, white roots and tubers, 2. Pulses (beans, peas, and lentils), 3. Nuts and seeds, 4. Dairy products, 5. Meat, poultry, and fish, 6. Egg, 7. Dark green leafy vegetables, 8. Other vitamin A-rich fruits and vegetables, 9. Other vegetables, 10. Other fruits. Participants were asked whether they consumed items from each group in the preceding day from when they woke up in the morning, through the day and night for the subsequent 24 hours. Each food group consumed (scored 1) was summed up to a score ranging from 0 to 10. A score above 4 (women who consumed items from five or more groups) was categorized as optimal dietary diversity while those consuming food items from less than five groups were considered as having suboptimal dietary diversity.2,18 The average of sum of three dietary diversity scores was considered to determine overall dietary diversity across pregnancy. The independent variables included a basic socio-demographic profile of the study subjects including place of residence (urban or rural), marital status, religion, educational level, occupation, husband’s occupation, husband’s education, family size and monthly income, and reproductive history that were assessed using standard demographic and health survey (DHS) questionnaire.29 The other independent variable was household food insecurity that was assessed using the household food insecurity access scale (HFIAS). The scale had nine questions intended to assess the experience of household food insecurity that occurred within the previous month.30 The households were categorized into 4 levels of food insecurity: food secure and mildly, moderately, and severely food insecure as per guidelines.28 Maternal nutritional knowledge and practice-related variables were also considered as independent variables of the study. Maternal nutritional knowledge was assessed using a non-standard scale containing ten items. Then, ten nutrition-related questions asked were recorded as complete answers (scored 2), incomplete answers (scored 1) and wrong answers (scored 0). Ultimately, maternal nutritional knowledge was scored out of twenty. Maternal dietary practices including meal frequency, meal skip, avoidances of food, craving and household access to food aid were also assessed. Women asked whether they were getting at least three meals per day or not and other dietary practice variables were assessed by yes or no questions. A guide was used for the in-depth interviews and focus group discussions to explore factors contributing to dietary diversity among pregnant women, available in Appendix 1. The interview guide explored dietary practice in terms of diversity, knowledge on the advantages of a diversified diet and factors that influence dietary diversity during pregnancy. During data collection, each question was probed for further exploration and the responses were recorded in a notebook and digital electronic recorder for later transcription. The quantitative data were entered and cleaned using Epi-data statistical software and then exported to SPSS version 24. Frequency distribution, measure of central tendency and dispersion were used to describe the data. Numeric variables were checked for normality of the distribution by using probability plot and Shapiro Wilk test. The association between dependent and independent variables was assessed using bi-variable and multivariable logistic regression. To test independent predictors of dietary diversity, all independent variables with a p-value of less than 0.25 in the bi-variable logistic regression model were considered as candidate variables for the multivariable analysis. Then the relationship was presented using adjusted odds ratio (AOR) with its corresponding 95% confidence interval. For the qualitative study, research assistants who were familiar with the local language and culture were involved. The professional’s word-by-word transcribing of the data from audio records was done in the local language each day at the end of data collection. Then, the transcribed information was translated from the local Amharic language to the English language in the word document. The data were analyzed by thematic analysis using NVIVO-version 11 software. After re-organizing data, an exploration such as the summary of word frequency and word cloud was done. The relevant information was coded and categorized into different themes. Final analysis was done by exploring different factors and creating a hierarchy chart of themes.
N/A