Study area
The study was conducted in Meket Woreda, Ethiopia. According to the 2007 Central Statistics Agency report, the woreda has a total population of 226, 644. From which, 112,246 are females. The woreda was identified as one of the most drought-affected and food insecure woreda in the Amhara Region. According to the woreda education bureau report, in the woreda, there are 50 full-cycle primary schools with 51,869 students. Of these, the SFP was implemented in 5 schools with 5,079 students.
Study design and population
A school-based comparative cross-sectional study was conducted among primary school students, from April 1-27/2015. All full-cycle primary school students in the woreda were the source population. Students in selected schools during the study period were the study population. In the study area, the school meals were given to the students starting from grade one. Students from grade four and above (students who took school meals three or more years) were included in this study. Similarly, students from grade four and above were also included in schools had no school feeding program. Students who transferred from other schools and not available during the data collection period were excluded from the study.
Sample size, sampling procedure, and data collection
The required sample size of the study was calculated using two population proportion formulas considering the following assumptions: a 95% confidence level and 80% power, 50% proportion of undernutrition among students who did not take meals at school and 40% proportion of undernutrition among students who took meals at school since there was no previous study. The population allocation ratio is 1:2; the design effect of 2 and 10% non-response rate was added. The final sample size was 1176. Of these, 392 students were from schools had no SFP and 784 students were from schools had SFP.
The study participants were selected using a multi-stage stratified sampling method. Using a list of kebeles (the smallest administrative unite in Ethiopia) as a sampling frame, 12 kebeles were selected from 45 kebeles in the woreda by simple random sampling (SRS) technique (lottery method). Next, schools in selected kebeles were stratified as the SFP schools have and non SFP schools.
In selected kebeles, there were 12 schools (8 schools without SFP and 4 schools with SFP). Two schools with SFP and four schools without SFP were selected by the SRS technique using the list of schools in selected kebeles as a sampling frame. The calculated sample size was allocated to the selected schools based on proportion to the size of the students from each school. Finally, using the school registration logbook as a sampling frame, students were selected by the SRS technique. The interview with students’ mothers was conducted in the school compound considering privacy.
Data were collected by the Amharic version (local language) structured interviewer-administered questionnaire. The questionnaire was taken from similar literature (20) and it included socio-demographic characteristics, environmental hygiene, and feeding practice related questions. Feeding practice was assessed using a 24hour recall method. Eight diploma and two BSc nurses were recruited as data collectors and supervisors, respectively.
The weight and height of the students were measured using SECA Germany weighing Scale and stadiometer, respectively. During weight measurement, weighing scales were calibrated each day prior to the actual data collection using a known weight material. Weight was measured to the nearest 0.1kg using a digital scale. The scale was adjusted before weighing every student by setting it to zero. The students were lightly dressed during having the weight taken.
Height was measured to the nearest 0.1 cm. During taking height, each student stood keeping normal anatomical position without shoes and heels, buttock, shoulder, and back of the head touched measuring board. Then, the headpiece of the measuring board touched the top of the head. For both weight and height, two readings were recorded and the computed averages were used in the analysis.
The z-score values for BMI-for-age and height for age were calculated using WHO Anthro-Plus software. Calculated z-scores of BMI-for-age and height for age were used to classify thinness and stunting using the new WHO 2007 reference value, respectively.
Operational definition
School feeding program (SFP) participants- are students who took meals at school.
School feeding program (SFP) non- participants- are students who did not take meals at the school.
Height-for-age at ≤ -3 SD, ≤ -2 SD and >-2 SD of the median value of the WHO international growth reference were defined as severely stunted, stunted and not stunted, respectively.
BMI-for-age at ≤ -3 SD, ≤ -2 SD, between +1 SD and > -2 SD, between >+1 SD and +2 SD and >+2 SD of the median value of the WHO international growth reference were defined as severely thin, thinness, normal, overweight and obesity, respectively(21).
Data Quality Control
Using trained data collectors and supervisors, pre-testing the questionnaire and checking the weighing scale for functionality were measures taken to assure the quality of data. Moreover, the collected data were reviewed and errors were returned to the data collectors for correction on a daily base. Supervisors and investigators closely supervised the data collection procedure.
Data processing and analysis
Data were entered and analyzed using SPSS version 20 software. The nutritional status of the students was compared. Bivariate and multivariable logistic regressions were done to identify factors associated with stunting and thinness. Hosmer-Lemeshow Goodness-of-fit-test
was done to check model fitness. Correlation between independent variables was checked using the Pearson Correlation Coefficient. The model was built with backward elimination. The crude Odds ratio was done and p-value ≤0.2 was taken as a cut-off point to select variables for the final model. Age was controlled during the multivariable logistic regression analysis. The adjusted Odds ratio was computed to determine the strength of association and control confounders. The p-value of less than 0.05 was considered statistically significant.
Ethical approval
The study was approved by the Ethical Review Board of Bahir Dar University. A letter of permission was taken from zonal and woreda health bureaus as well as the school administrators. Since it causes less than minimum risk, verbal consent was taken from parents and assent was taken from students. Privacy and confidentiality were maintained throughout the study period by excluding personal identifiers from the data collection tools.