Assessing the Safety of Iron Supplementation in Micronutrient Powder (MNPs)

For morbidity assessment 65 eligible children per cluster (18 intervention and 18non-intervention villages) a total of 2340 will be randomly selected The sample size was calculated based on an expected 12% increase in the incidence of overall diarrhoea and 30% decrease in anemia in the intervention group. The investigators assume that the investigators will be able to detect this increase with 80% power and 95% confidence interval, with an assumed 10% drop-out and design effect of two.

• Morbidity from infectious diseases will be assessed by means of a standardized recall questionnaire. Data will be collected on the following symptoms: diarrhea, respiratory signs, fever and other illnesses. Children who require medical treatment will be examined by a Health Officer at the health center following standard protocols of the Federal Ministry of Health.
• Biochemical iron status will be assessed from non-fasting venous morning blood samples. Blood samples will be kept cool and serum separated and frozen at -20oC within 3 hours of collection. Hemoglobin, serum ferritin, C-reactive protein (CRP), soluble transferrin receptor and AGP will be assessed following standardized methodologies at the laboratories of the Ethiopian Public Health Institute (EPHI).
• Anthropometry: Trained field staff will collect duplicate measurements every four months of weight, recumbent length (infants) and height (children > 12 months) using standardized methods and high quality equipment. Z-scores for length-for-age (LAZ), weight-for-age (WAZ) and weight-for-length (WLZ) will be calculated, using the 2006 WHO Child Growth Standard; and stunting, under-weight and wasting will be defined as LAZ, WAZ and LAZ <-2 SD, respectively.
• Infant feeding status. Information on infant feeding status will be collected every four months by means of a standardized questionnaire.
• Compliance to the program intervention: will be assessed by counting the number of returned and used MNP sachets every 2 weeks.
• Social, economic and demographic characteristics: will be collected in baseline questionnaires using Demographic and Health Survey methods.

Data will be analysed on an intention-to-treat basis, using the SPSS 19.0 statistical program. Random effect models will be used to account for repeated measures or episodes within children, and multiple children in each cluster.

For morbidity, the number of actual surveillance days will be counted by subtracting the days on which no recall data were available from the total days of follow-up. For the analysis of disease incidence, differences between groups will be assessed with random effects regression including number of episodes as dependent, control- or intervention as independent, and total days of actual surveillance as off-set term in the model. Longitudinal prevalence of disease will be calculated for each individual as the percentage of actual surveillance days with illness. All continuous variables (longitudinal prevalence, body iron status, mean and total weight and height gain) will be analysed using linear mixed models using cluster, household and child as random effects to account for clustered observations. Mixed effects logistic regression analysis will be used to evaluate binary outcomes (anemia, iron deficiency and iron-deficiency anemia), using cluster, household and child as random effects. All models will include the following fixed effects as co-variates: child's sex, socio-economic status, and baseline values. Non-normal distributed values will be log-transformed. P-values of < 0.05 will be considered statistically significant.