The study was conducted at Wacha primary hospital which was established in 2005. It is found in Chena town, Kaffa Zone, Southwest Ethiopia which is located 541 km far from Addis Ababa, the capital city of Ethiopia and it is the only hospital in Chena district. The hospital has been giving health services for the total population of 30,891 people in its catchment area with different four major departments including medical, pediatrics, surgical, and obstetrics, and gynecology. It also provides outpatient service, ophthalmology, emergency, antiretroviral Therapy (ART) clinic, and ANC clinic. There were a total of 723 pregnant women who had attended ANC in 2020/21 and on average around 30–40 followers have been attended ANC clinic every day at Wacha primary hospital.
Study design and study period
A facility-based cross-sectional study was conducted from January to February 2021.
All pregnant women attended the ANC clinic at Wacha primary hospital, Kaffa zone.
All randomly selected pregnant women attending ANC clinic at Wacha primary hospital from 16 weeks gestational age were included in the sample.
Sample size determination
The required sample size was determined using single population proportion formula based on the assumption of high minimum dietary diversity of 55.2% in Bale zone, Ethiopia , 5% margin of error, 95% of Confidence Interval (CI), 10% of non-response rate, and by using the correction formula since the target population of the study area was below 10,000. Then the final sample sizes 274 of pregnant women were included.
The study participants were selected by using a systematic random sampling technique. The sampling interval was determined by calculating monthly average attendance for ANC follow-up divided by the required sample size, and then, the first study participant was selected randomly and then every third pregnant woman was included.
Pregnant women with greater than 16 week’s gestational age and who lived in the study area for at least one year were included in the study.
Pregnant women who were unable to speak&/hear and who have seriously ill during data collection.
The dependent variable of this study was dietary diversity status, and independent variables like age, marital status, occupation, residence, educational status, family size, household head, nutrition awareness of women, maternal health status, husband support, income, having mobile phone, radio and bank account, garden, water source, availability of latrine, marketing, and household food security and nutritional status of pregnant women.
Data collection technique
The data were collected through face-to-face interviews using pretested structured and semi-structured questionnaires adapted from different kinds of literature [2, 33, 36]. The data were collected by well-trained Nurse professionals. The questionnaire had three parts. The first part includes socio-demographic factors (age, marital status, residence, family size, education, occupation, and others). The second part was dietary-related information questionnaires which were adapted and modified from the Food and Agriculture Organization of the United Nations (FAO) 2016 . The dietary diversity questioner has ten different food groups based on their nutrients:1) grains, white root, tubers, and plantains, 2) pulses (beans, peas, and lentils), 3) nuts and seeds, 4) dairy, 5) meat and fish (poultry and fish), 6) eggs, 7) dark green leafy vegetables, 8) vitamin A-rich fruits and vegetables, 9) others vegetables, and 10) others fruits. It was assessed by using 24-h open dietary recall methods; one point was given to each food group consumed over the past 24 h before the survey period. The participants were asked about all food and beverage consumed during the day and night including any snack in the past 24 h and the interviewer were probing for any food types forgotten by participants. Each food or beverage that the respondent mentions was circled underlined on a predefined list. The foods not included on the predefined list were classify by the principal investigator on an existing predefined food group or recorded in a separate place on the questionnaire and coded and organized later into one of the predefined food groups .
The third part was household food security status which was assessed by using Household Food Insecurity Access Scale (HFIAS) . The household food security was categorized as food insecure for those who score 2 and above out of 27 household food insecurity indicators and while food secures categorized for pregnant women who scored less than 2 out of 27 household food insecurity indicators.
Nutritional status of pregnant women was assessed by using mid-upper arm circumference (MUAC) which was measured halfway between the olecranon process and acromion process by using -non-stretchable tape to the nearest 0.1 cm. Nutritional status was defined as MUAC less than 22 cm were undernutrition and while 22 cm or more was considered to be normal nutritional status [38, 39].
Data quality control
Initially, a survey questionnaire was prepared in English translated to the local language and translated back to English to check for consistency. A three days training was given for data collectors and supervisors and a pretest was done on 5% of the study sample size. The Cronbach’s alpha was calculated (P = 0.78). The data were checked by the principal investigator on the daily basis for completeness and consistency.
Data processing and analysis
After the data were cleaned and coded, the data were entered in Epidata 3.1 version software and exported to statistical package for social science (SPSS) version 21 for analysis. A minimum dietary diversity score (MDDS) was dichotomized as meet minimum dietary diversity for pregnant women who consumed 5 and above out of ten food groups coded as 1 and while unmeet minimum dietary diversity for those who consume less than 5 out of ten food groups in the past 24 h coded as 0. Descriptive analysis was done to calculate the mean, frequencies, and percentage distributions for the variables. A stepwise backward elimination logistic regression was done to identify variables associated with minimum dietary diversity. The model fitness was checked by using Hosmer and Lemeshow statistic (P = 0.47), which showed the model was fitted. Bivariate logistic regression was done to identify the covariate associated with minimum dietary diversity and independents variables and the variables with a p-value less than 0.05 were considered for multivariable logistic regression to control all possible confounders and to determine the strength of association between minimum dietary diversity and each explanatory variable. The strength of association was measured with an adjusted odds ratio (AOR) with a 95% CI. Finally, the variables with a P-value less than 0.05 were considered statistically significant.