Study setting and population
Health facility based cross-sectional design was employed to assess dietary diversity among lactating mothers in public health facilities of Aksum town. Aksum town is located in the Northern part of Ethiopia with an elevation of 2131 m (6991 ft) above sea level, with annual average temperature of 18.30c and annual rain fall of 652 mm. It is 1042 kms far from Addis Ababa, the capital of Ethiopia and 270 kms away from Mekelle (the capital city of Tigray regional state). Aksum is the ancient historical town which has an area of 3247 km2. Based on the town administrative office 2013 report, the total population is 59,300 (28,637 are males and 30,663 are females). Administratively, the town is divided in to 4 kebeles (the smallest administrative units). In Aksum town, there are 2 governmental health centers and 1 governmental general hospital. Lactating mothers who visit the public health facilities during the study period were considered as the study population. The inclusion criteria were being 15–49 years lactating mother and being resident of Aksum town for at least 6 months. Lactating mothers who were seriously ill and/or had difficulty to communicate were excluded. The data collection was undertaken in June, 2015.
Sample size and sampling procedures
The required sample size was determined using single-population proportion formula with the following assumptions: 50 % of proportion of lactating mothers who have below mean diet diversity score, 95 % of confidence interval and 5 % marginal error. The final sample size was 384. Respondents who fulfilled the inclusion criteria were selected using systematic random sampling technique. The interview was started by selecting a random start using lottery method. Then every 2nd lactating woman visiting the public health facilities were interviewed. Eventually, a total of 346 lactating mothers were participated in this study.
Data collection
Data were collected by exit interview using structured and pre-tested questionnaire adapted from different literatures mainly Food and Agriculture Organization (FAO) Guidelines for measuring household and individual dietary diversity, 2011 [15] and Household food insecurity access scale (HFIAS) to measure food security [23]. The questionnaire had four main contents: socio-demographic characteristics; source of food, feeding practice and other individual related factors; dietary diversity, and food security related questions. The questionnaire was first prepared in English and then translated to Tigrigna and translated back to English to observe its consistency. Finally, the questionnaire was also pretested on 5 % of the total sample size in Suhul general hospital found in Shire Endaselassie town of North Western zone of Tigray Regional state. Necessary modifications were made on the questionnaire based on the findings of the pretest.
Dietary diversity score (DDS) was collected and calculated as the sum of the number of different food groups consumed by the mother in the 24 h prior to the assessment. A total of nine food groups were considered in this study (i.e. starchy staples, dark green leafy vegetables, other vitamin A rich fruits and vegetables, other fruits and vegetables, fats and oils, meat and fish, eggs, legumes, nuts and seeds and milk and milk products [28]. Finally, respondents with < 3.4 mean food groups consumed were considered as having lower dietary diversity whereas those with ≥3.4 mean food groups consumed were considered as having moderate or high dietary diversity.
Food insecurity access was measured using items from the HFIAS [23, 29–31]. The HFIAS consists of 9 items specific to an experience of food insecurity occurring within the previous 4 weeks. Each respondent indicated whether he/she had encountered the following due to lack of food or money to buy food in the previous 1 month: (1) worried about running out of food, (2) lack of preferred food, (3) the respondent or another adult had limited access to a variety of foods due to a lack of resources (4) forced to eat un preferred food due to lack of resources, (5) eating smaller portions, (6) skipping meals, (7) the household ran out of food, (8) going to sleep hungry, and (9) going 24 h without food. Finally, individuals were classified as food secure if the individuals responded ‘no’ to all of the items and insecure if the individuals responded ‘yes’ to at least one of the 9 items included on the HFIAS tool.
Two public health officers, who were fluent in the local language (Tigrigna), participated in the data collection process. The data collection process was supervised by two senior Public Health professionals and the principal investigators. Both data collectors and supervisors were trained for 1 day about the contents of the questionnaire and on how to collect the data properly in order to minimize errors. The principal investigators and supervisors had made daily supervision during the whole data collection process. The questionnaire was reviewed and checked for completeness, accuracy and consistency by the supervisors and investigators daily and at the end of the whole data collection process.
Operational definition
Diet diversity score
Is the sum of food groups eaten in the previous 24 h, serves as a proxy indicator of nutrient adequacy of an individual’s diet. Low dietary diversity (<3.4 mean food groups consumed) and moderate or high dietary diversity (≥3.4 mean food groups consumed). The diet was classified according to the 9 food groups recommended by the FAO [28].
Food security
Food insecurity access was measured using items from the HFIAS [23, 29–31]. The HFIAS consists of 9 items specific to an experience of food insecurity occurring within the previous 4 weeks. Each respondent indicated whether he/she had encountered the 9 items included in the HFIAS due to lack of food or money to buy food in the previous 1 month. Finally, individuals were classified as food secure if the individuals responded ‘no’ to all of the items and insecure if the individuals responded ‘yes’ to at least one of the 9 items included on the HFIAS tool.
Data analysis
After all the relevant data were collected, the data were coded on pre- arranged coding sheet by the investigators. Dietary intake was converted to 9 food groups to assess dietary diversity of the subjects. Data were entered, cleaned and then analyzed using SPSS version 20. Descriptive summaries using frequencies and proportions were used to present the study results. In this study, the dependent variable was dietary diversity coded 1 as lower dietary diversity and 0 as moderate or high dietary diversity. In the bivariate analysis, the independent variables with a P-value less than 0.2 with the dependent variable were fitted in to a multivariate logistic regression model to identify their independent effect on dietary diversity. Independent variables included in the multivariate analysis include educational status, occupation, residential address, monthly income, source of drinking water, availability of latrine, home gardening, history of illness, duration of menstruation (in days) and daily eating pattern in the previous 7 days. The association between the dependent and the independent variables was measured using odds ratio (OR) with 95 % Confidence Interval (C.I). Those variables with p-value of less than 0.05 in the multivariate analysis were considered as significant.