Validation of the CIMI-ethiopia program and seasonal variation in maternal nutrient intake in enset (False banana) growing areas of Southern Ethiopia

Background: Tools for the rapid and accurate analysis of nutrient intakes from diets of individuals in Southern Ethiopia are lacking. The Calculator of Inadequate Micronutrient Intake program for Ethiopia (CIMI-Ethiopia) has been developed to overcome this problem. CIMI-Ethiopia also computes protein and energy intakes from the diet. The objectives of the current study were to validate CIMI-Ethiopia for the dietary pattern of Southern Ethiopia, and assess the nutrient intakes in postharvest dry and lean wet seasons. Methods: 24-h dietary recall (24HR) data was collected from 578 women of a reproductive age in postharvest dry and lean wet seasons in 2017. For analysis, 24HR data was entered into NutriSurvey (NS), which was the reference nutrition software, and then into CIMI-Ethiopia. For validation, the mean and standard deviation (SD) of the difference between CIMI-Ethiopia and NS were computed. The percentage of participants with an inadequate intake was calculated. The correlation between CIMI-Ethiopia and NS results was determined. The nutrient intakes in postharvest dry and lean seasons were compared. Results: Among the nutrients, pantothenic acid, vitamin B1, and protein showed a very high accuracy in CIMI-Ethiopia calculation (|difference (D)| < 5.0% of the NS result). Nutrients with a good accuracy (|D| = 5%–15%) were iron, zinc, magnesium, vitamin B12, vitamin B6, and energy. The accuracy for calcium, niacin, and vitamin A was moderate (|D| = 15%–30%). The intakes calculated by CIMI-Ethiopia and NS of iron, zinc, magnesium, calcium, B-complex vitamins, vitamin A, protein, and energy were highly correlated (r = 0.85–0.97, p < 0.001). NS analysis identified a significant reduction in the mean intake of iron; zinc; magnesium; pantothenic acid; vitamin B1, B12, and D; protein; and energy in the lean wet season; however, calcium and vitamin A intake increased. Conclusions: It has been found that CIMI-Ethiopia is a valid tool for estimating nutrient intakes at an individual level in Southern Ethiopia. The study demonstrated a decline in intakes of iron; zinc; magnesium; pantothenic acid; vitamin B1, B12, and D; protein; and energy in the lean wet season. This result provides some hint for fortification and supplementation programs that aim to combat maternal malnutrition in rural Southern Ethiopia. This study was conducted in Shebedino and Hula Districts from Sidama Zone in the Southern Nations Nationalities and Peoples Region (SNNPR). The regional capital city Hawassa is located nearly 278 km south of the national capital Addis Ababa. Sidama Zone has a total population of nearly 3.1 million [25]. Enset, maize, and coffee are among the major crops in Sidama Zone. Enset is a crucial drought-resistant crop with significant cultural value in the zone [26]. Shebedino and Hula are two of nineteen districts and two city administrations of the zone [27]. Nearly 99% of the households in the five kebeles (kebele is the smallest administrative unit of a district) sampled from the two districts produced enset in the l2 months before the survey [28]. This is a longitudinal study in which food and nutrient intake data were collected in postharvest dry (January to first week of February 2017) and lean wet seasons (June 2017). All women of a reproductive age with 24–59-month-old children were eligible for this study. In order to reduce those lost to follow-up, temporarily resident mothers were not included in the study. Besides, women were excluded from the survey when reported as being severely sick to avoid the effect of illness on the dietary intake due to a reduced appetite. For this particular analysis, a basic sample size of 319 was computed using a correlation coefficient (r) with the following formula [29], assuming a two-tailed test: α (type I error rate), 0.050; β (type II error rate), 0.050; and r (expected correlation coefficient), 0.20. Here, N = sample size; α = Zα = 1.960; β = Zβ = 1.645; and C = 0.5 * ln[(1 + r)/(1 − r)] = 0.203. On top of this, the design effect of 1.5 was considered, and 10% of the basic sample size was added to compensate for those lost to follow-up, resulting in a total of 510. However, a total of 625 mothers with 24–59-month-old children were enrolled at the beginning in order to obtain a sufficient sample size for additional analysis [28]. Nonetheless, 578 completed the study, while the rest left for different reasons: death (1), delivery (4), and sickness or moving to other places (42). A two-stage sampling technique was employed. Firstly, Shebedino and Hula Districts were selected from the Sidama Zone. Secondly, Fura, Howolso, Dila Gumbe, Worare, and Chirone Kebeles were randomly selected using probability proportional to size (PPS) from the two districts. Then, the sample size was distributed to each kebele using PPS. The samples were randomly selected based on the sampling frame prepared from a house-to-house listing of households with 24–59-month-old children [28]. Ethical clearance was received from the Institutional Review Board of Hawassa University, Ethiopia, and Ethik-Kommission, Landesäztekammer Baden-Württemberg, Germany (F-2016-127). Permission was obtained from health administrative offices of the study districts to access the study population and collect data. Informed written consent was received from mothers. Information was kept confidential with pseudonymous codes. Data collectors with previous experience were recruited. The principal investigator trained the data collection staff for one week. The same data collection team collected the lean season data after additional project-specific training. The pre-test of survey tools was carried out on five percent of the total sample size to check for appropriateness. Data collection was supervised by the principal investigator. A structured questionnaire was first prepared in English, and then translated to Amharic. The purpose of this questionnaire was to collect the sociodemographic data. A standard 24HR protocol was used to collect the dietary intake in the 24 h preceding the survey. To assess the usual dietary nutrient intake, the dietary consumption data was collected on days other than fasting and holidays. In order to facilitate mothers’ recall, food models and food charts were used. The amounts of foods and beverages consumed were estimated in terms of local units: coffee cup, ladle, soup spoon, water glass, and tea spoon. Printed pictures aided to identify portion sizes, for instance large or medium or small, for countable food items like potato. Data collectors probed the mothers to remember all food items consumed in the 24 h preceding the survey. A compact scale CS2000 (Ohaus Corporation, Parsippany, NJ, USA) was used to convert the weight equivalents of the local units into grams of foods consumed. Before data entry, the uniformity of the nutrient compositions of foods in CIMI-Ethiopia and NS were assured by adopting the compositions configured in the prior one. Following this, the dietary consumption data collected with 24HR was entered into NS 2007, which is a reference program used to compute nutrient intakes. Then, nutrient data was exported from NS into Microsoft Excel 2016, and subsequently into Statistical Package for Social Sciences (SPSS) ver. 20 (IBM Corporation, Armonk, NY, USA) for analysis. Secondly, the same dietary intake data from 24HR was entered into CIMI-Ethiopia. The nutrient intake data calculated by CIMI-Ethiopia was automatically transferred to a server when WIFI was available. Later, the data was downloaded from the server and imported into SPSS. Thirdly, the sociodemographic data was entered directly into SPSS. Then, the data sets were compiled together and checked for completeness before analysis. One participant was excluded from this analysis, as the amount of the food consumed was missing, though types of foods were recorded. Data normality was checked with the Kolmogorov–Smirnov test. For validation, the mean and standard deviation (SD), and median with 25th and 75th values of intakes, were calculated for iron, zinc, calcium, magnesium, vitamin A, vitamin B1, niacin, pantothenic acid, vitamin B6, vitamin B12, vitamin D, protein, and energy from CIMI-Ethiopia and NS. The Pearson’s test was used to determine the correlation between the nutrient results from CIMI-Ethiopia and NS. Furthermore, a Bland–Altman plot was used to determine how well the data fit. Besides, the percentage of participants below the threshold of an inadequate intake was computed. The nutrient intake less than two-thirds of the FAO/WHO RNI [22,23,24] was taken as a cutoff point for the threshold of an inadequate intake [14,30,31]. To determine the level of accuracy, the mean and SD of the difference between the intake computed by CIMI-Ethiopia minus the intake calculated by NS were computed. Accordingly, an absolute difference, i.e., |D (difference)| < 5.0% of the NS result was categorized as a very high accuracy, |D| = 5%–15% as a good accuracy, |D| = 15%–30% as a moderate accuracy, and |D| > 30% as a low accuracy [17]. In order to determine the seasonal differences of nutrient intakes, the Wilcoxon signed-rank test was used to compare non-normally distributed mean nutrient intakes between postharvest dry and lean wet seasons computed by NS. The cutoff point for statistical significance was a p-value ≤ 0.05.