Menu
Menu
Menu
Menu
About half of Ethiopians belong to the Orthodox Tewahedo religion. Annually, more than 200 days are dedicated to religious fasting, which includes abstaining from all types of food, animal source foods, and water. However, the association of fasting with undernutrition remains unknown in Ethiopia. Therefore, dietary pattern and nutritional status of lactating women during lent fasting and non-fasting periods were studied, and predictor variables for maternal underweight were identified. To achieve this, lactating mothers in lent fasting (N = 572) and non-fasting (N = 522) periods participated from rural Tigray, Northern Ethiopia. Average minimum diet diversity (MDD-W) was computed from two 24-h recalls, and nutritional status was assessed using body mass index (BMI). Binary logistic regression was used to identify potential predictors of maternal underweight. Wilcoxon signed-rank (WSRT) and McNemar’s tests were used for comparison of the two periods. The prevalence of underweight in fasting mothers was 50.6%. In the multivariate logistic regression model, younger age, sickness in the last four weeks preceding the survey, fasting during pregnancy, lactation periods, grandfathers’ as household decision makers, previous aid experience, non-improved water source, and not owning chicken were positively associated with maternal underweight. In WSRT, there was no significant (p > 0.05) difference on maternal body weight and BMI between periods. The average number of meals, diet diversity, and animal source foods (ASFs), consumption scores were significantly increased in non-fasting compared to fasting periods in both fasting and non-fasting mothers (p < 0.001, p < 0.05, and p < 0.001, respectively). Consumption of dark green leafy vegetables was higher in the fasting period (11%) than non-fasting (3.6%), in the study population. As a conclusion, Ethiopian Orthodox fasting negatively affected maternal nutritional status and dietary pattern in rural Tigray, Northern Ethiopia. To reduce maternal malnutrition in Ethiopia, existing multi-sectoral nutrition intervention strategies, should include religious institutions in a sustainable manner.
The study was conducted in the Genta Afeshum woreda of rural Tigray, Northern Ethiopia. The woreda covers an area of 1636 km2 with a total population of 99,112, and almost all people (99%) are Orthodox Christians. The woreda reside at an altitude between 2045 and 3314 masl. The woreda is classified as a hotspot for food insecurity [31,32,33]. In the woreda, drought, hail storms, and livestock diseases are the major disaster risks; followed by human diseases, crop diseases, pests, and flooding. Additionally, deforestation, water pollution, and soil erosion are the major environmental problems; whereas, high dependency syndrome, poor economic conditions, land shortage, severe shortage of drinking water, and poor saving are among the major vulnerability factors at the household level [34]. The study had a community-based longitudinal survey design, to assess the nutritional status and dietary pattern of lactating mothers. The data was collected during the Ethiopian Orthodox lent fasting period (Fasting of Jesus Christ, 15 February 2017 to 15 April 2017) and non-fasting periods (1 May 2017 to 30 May 2017). The sample size was calculated based on the prevalence of underweight in lactating mothers in the Tigray region, using the formula for estimating single population proportion and considering a 95% of confidence interval for true prevalence, and a relative precision (d) of 5%. In lactating mothers, the prevalence of underweight (BMI < 18.5 kg/m2) was 25% elsewhere in Tigray [5]. The total number of lactating mothers was estimated to be 3369, which was less than 10,000; therefore, the finite source population size correction formula was used. Additionally, 10% was considered as non-responses and dropout rates. Moreover, a 1.5 design effect was used on the final calculated sample size, and the final total sample size was 575. Multi-stage systematic random sampling was applied to obtain representative samples for the study. At first, Genta Afeshum was randomly selected out of the three GIZ Ethiopia, Nutrition Sensitive Agriculture (NSA) project woreda’s in Tigray region. Out of twenty rural kebeles in the Genta Afeshum woreda, seven were randomly selected. Then, the list of households which had lactating mothers with children aged between 6 to 23 months old, who fulfilled the inclusion criteria, was prepared for the seven kebeles (lowest local administrative unit) by health extension workers at the nearby health posts. Subsequently, the samples were chosen using systematic random sampling techniques. Ten trained and well experienced data collectors who were fluent in Tigrigna, Amharic, and English languages were recruited. Additionally, before conducting the main survey, the questionnaire was translated to Tigrighna by a professional translator and verified by data collectors. Then the translated questionnaire was pre-tested for its appropriateness, by administering it to lactating mothers around Mekele, and corrections were made. Structured and semi-structured questionnaires were prepared to collect information on socio-demographic and economic characteristics, maternal and child characteristics, water, sanitation and hygiene (WASH), feeding practices, and household food security indicators [35]. Before conducting the study, the whole study protocol was ethically approved by the Institutional Review Board of the College of Health Sciences at Hawassa University and the Tigray Region Health Bureau in Ethiopia; and the ethical review committee of Landesärztekammer Baden-Württemberg, Germany. Permissions from Genta Afeshum Woreda Health Office were also obtained. After the purpose of the study was explained to the study participants, agreement to participate in the study was documented by signing the informed consent. Each participant was also told that the collected information was confidential, and whenever she wanted to discontinue, withdrawal from the study was possible. The minimum-diet diversity score (MDD-W) was obtained by (a) collecting two 24-h dietary recalls; (b) categorizing as consumed or not consumed of the food group, considering the minimum amount (15 g) of any food items or the sum of food items eaten under a given food group and giving a score of 1 if consumed, otherwise 0 if not; (c) calculating the diet diversity score for each two days using 10 food groups, as the summation of consumed food groups or scored 1 for each day separately; and (d) taking the average diet diversity score of the two days, as an individual MDD-W score. The 10 food groups used for calculating MDD-W score were grains, white roots and tubers, and plantains; pulses (beans, peas, and lentils); nuts and seeds; dairy; meat, poultry, and fish; eggs; dark green leafy vegetables; other vitamin A-rich fruits and vegetables; other vegetables; and other fruits [36]. Household food insecurity data was collected using the household food insecurity access scale (HFIAS). It is the measure of the degree of food insecurity (access) in the household in the past 4 weeks (30 days). The questionnaire encompassed nine questions, which assess the occurrence of food insecurity in increasing level. Under each of the nine questions, frequency of occurrence questions were a follow up to determine how often the condition (1 = rarely, 2 = sometimes, 3 = often) occurred. The Household Food Insecurity Access Prevalence (HFIAP) status indicator, was used to determine prevalence of food insecurity to report household food insecurity. Using the HFIAP indicators four categories, the households were categorized into four levels of household food insecurity (access). These were food secure, mild, moderately, and severely food insecure. After creating these four categories, the HFIAP was calculated as the number of households with a given food insecurity category divided by the total number of households with household food insecure access category, multiplied by 100 [35]. Principal component analysis (PCA) was carried out to compute the wealth index. To achieve this, 17 variables: (Transport animals (horse/donkey/mule), goat and/or sheep, household owns kerosene or lamp, owns bed, chair, table, radio, electric-mitad, bicycle, mobile phone, non-mobile phone, animal drawn cart, motor bicycle, TV, electricity, windows, and separate room for animal), which could indicate the living standard of the surveyed area were included in the analysis. The first factor that explained most of the variation (86.3%) was used to group study households. Finally, the wealth tertile was performed and categorized as higher, medium, and lower [37]. Weighing body mass of the mothers was conducted using a portable digital scale (Seca 770, Hanover, Germany), working with a powered battery and measured to the nearest 0.1 kg. For height measurement, a dissembling plastic height measuring board with a sliding head bar was used and measured to the nearest 0.1 cm. During weight and height measurements, the mothers were advised to remove their jackets until they had light clothes to minimize the weight due to clothes. The measurements (height and weight) were carried out using standardized equipment and procedures in duplicate and the average values were used. Additionally, the BMI of the mothers was calculated as the weight of the mothers in kilograms divided by the square of their height in meters. The BMI values of mothers were classified in three categories as underweight, normal, and overweight (<18.5, 18.5–24.99, and ≥25 kg/m2), respectively [38]. Before submitting the data, variable coding was conducted in SPSS version 20. Following this, the data was entered, cleaned, and analyzed. First, frequency and crosstab were conducted to determine completeness of data and to present the results in descriptive statistics (frequency and percent). Association between outcome and potential explanatory variables, was assessed using bivariate analysis with a confidence level of 95% to declare the statistical significance. Out of all the independent variables entered in bivariate logistic regression, seventy variables with p-values of less than 0.25 were entered for multivariable logistic regression to identify predictor variables for maternal underweight (BMI < 18.5 kg/m2). p-value < 0.05 was used to declare the variables as predictors for the outcome variable. Hosmer and Lemeshow test and C-statistics (AOC) were conducted to assess fitness of the final model. Meanwhile, multi-collinearity was checked using the variance inflation factor (VIF) and standard error with <10 and <2 as a cutoff point, respectively. Maternal nutritional status was defined as underweight (BMI < 18.5 kg/m2) and normal if BMI ≥ 18.5 kg/m2, since the interest of this study was being underweight for logistic regression. Normality of continuous data was checked using the Kolmogorov-Smirnov test. Non-normally distributed data was analyzed using the Wilcoxon signed-rank test; whereas the dichotomous data was analyzed using McNemar’s test to detect a difference between fasting and non-fasting periods, of fasting and non-fasting mothers for their dietary pattern, separately.
N/A
