Study area: - The study was conducted in Dabat Demographic and Health Surveillance system site (DHSS). The Dabat DHSS covers 13 kebeles (four urban and 9 rural) divided into 83 clusters. The altitude of the district is divided into highland, Midland, and lowland climatic conditions. Dabat town, the capital of the district, is located at about 821 Km from Addis Ababa and 75 Km from Gondar town to the north. The district has six health centers, three health stations, and thirty-one health posts providing health services to the community. The total population of the district in 2012/13 was estimated to be 158, 250 of whom about 70,611 people were the population of HDSS site with a sex ratio nearly 1:1. in the HDSS there are 7,918 under-five children from 6,314 households included in this study (39).
Study design and population: the community-based cross-sectional study design was carried out among rural and urban households of HDSS site from April to December 2016. Heads households in the HDSS site during the study period were study participants. In the absence of heads of the household during visit day; mother or other senior household members beyond 18 years were interviewed for the tool after consenting.
Data collection tool and data collection procedure: -a pre-tested interviewer-administered structured questionnaire prepared with local language was used to collect data on socio-demographic, maternal and child health service utilization and mothers/caretakers food preference habits for their under-five children. FANTA III food insecurity access measurement scale version 3 was used to collect data. A five-day intensive training was provided for 15 data collectors and 5 supervisors. A pre-test was conducted in the rural and urban kebeles of Dabat district which are not included in the HDSS and necessary modification of the tool was made according to the inputs obtained from the pre-test.
Data processing and analysis: collected data were entered into Epi data template prepared by Amharic language having similar content with the data collection question to avoid data entry error related to the translation of the responses. Data entry was performed by five experienced data entry clerks working for Dabat HDSS. The overall data entry process was supervised by the data manager working HDSS. Entered data was transported to STATA version 14.1 software. Before actual data analysis data management and clearance was performed by the data manager. Before fitting univariate and multivariate logistic regression models the relationship of the dependent variable with each independent variables was assessed by Pearson's chi-square test. Then variables with significant Pearson's chi-square were considered for the univariate ordinal logistic regression model. All variables significant at the univariate ordinal logistic regression model were fitted into multivariate ordinal logistic regression models to identify predictors for the worsening state of household food insecurity. Proportional or cumulative odds ratio assumption was checked by significant Pearson's chi-square.
Study variables
Dependent variable: - state of household food insecurity
Independent variables
Socio-demographic characters: - (wealth status of the household, educational status, marital status and residence of head of the household number of children in the household, age of children in the household, sex of children in the household , family size, religion ethnicity, occupation and availability of garden and source of food item for the household )
Maternal and child health utilization: - (TT vaccine, Iron supplementation , ANC visit, and Place delivery for the last pregnancy, PNC service for the last birth, VitA supplementation for children in household , Child deworming, BCG, polio, DPT, Penta Valiant, Rota ,PCV , measles vaccine, illness and Treatment for illness )
Food preference related factors: - (preference of balanced diet, specific food preference, and preference of family food)
Operational definition
Household food security:-if the interviewed heads of the household answered rarely for only Q1a
Mild food insecurity:- if the interviewed heads of the household answered sometimes or often for Q1a or rarely, sometimes or often for Q2a or rarely for Q3a or Q4a it was considered as mild household food insecurity.
Moderate food insecurity: - if the interviewed heads of the household answered sometimes or often for Q3a or Q4a or rarely, sometimes Q5a or Q6a it was considered as moderate household food insecurity.
Sever food insecurity:- if the interviewed heads of the household answered often for Q5a or Q6a or rarely, sometimes or often for questions Q7a-Q9a it was considered as severe household food insecurity
Food preference:- if parents choose to feed food with the same caloric content more than once per day it is considered as preferring to feed specific food preference, if they tend to feed any available food or the food prepared for adult family members it is considered as preference to feed family food and if there is habit of balancing child food from locally available food items it is preference to feed balanced diet.