Study design and setting
This was a cross-sectional retrospective study conducted in the city of Harare. Harare is the capital city of Zimbabwe with a total population of 2 123 132. The average household size is 4 and 74% of households are male headed. A third of the population (29%) are home owners whilst 48% are lodgers (12). The city is divided into four districts (Eastern, Southern, Northern and Western districts). At the time the study was conducted the city was divided into 9 districts (northern, north eastern, eastern, south eastern, southern, south western, western, north west and central business district). The study was health facility based. Using growth monitoring attendance figures, primary care clinics with the largest attendance of individuals from low socio economic districts were purposively selected . These were Hatcliffe and Borrowdale Poly clinic from Nothern District, Mbare Poly Clinic from Southern District, Dzivarasekwa Poly clinic from North Western District and lastly Hatfield Poly clinic from South Eastern District. In total this gave 5 poly clinics from 4 Districts.
Study participants
The following formula used to calculate sample size in cross sectional studies was used:
N = z2 p (1-p)/e2 (13,14)
Where:
N = sample size
Z = confidence interval (which is at 95%, 1.96)
P = proportion of children aged 6-23 months living with their mothers who are fed according to the three IYCF feeding practices (breastfeeding status, number of food groups, meal frequency for Harare was 15.4%) .
e = error level of precision (which is 0, 05)
The calculated sample size was 196. To adjust for attrition, a non-response rate of 10% was factored in to give a final sample size of 216 (approximately 43.2 mother infant pairs per clinic) . Eligible participants were enrolled as they came to attend the monthly growth monitoring programme. Researchers continued to visit the clinic for interviews until the sample size specifically calculated for the health facility was reached. The inclusion criteria was any mother with a child under the age of five years with no underlying health problems. We excluded mothers or caregivers who were recent visitors to the area, non consenting and none parents to the infant. The study was conducted based on the ethical principles of respect, justice and confidentiality summarised in the 2013 Declaration of Helsinki (15). The study was approved by the Harare City Health Services Department. Written informed consent was obtained from all participants prior to study procedures.
Data collection instruments
A structured interviewer-administered questionnaire was used to interview caregivers on child-feeding practices from the time the child was born. The questionnaire was based on WHO IYCF indicators (6) and comprised questions on demographic characteristics of the caregiver, previous and current child-feeding practices as well as social, economic and cultural factors that influenced the child-feeding practices. The questionnaire was pretested for ambiguity, validity and reliability at a clinic that was not participating in the study and revised accordingly. To assess for morbidity the caregiver was asked to recall if the child had been ill first in the 1 month preceeding the interview. This was confirmed using clinic card where possible. Height of the caregiver were measured to the last completed 1mm using a stadiometer and weight to the nearest 0.1kg using a Tanita scale (Tanita, IL USA). Body mass index (BMI) of the caregiver was determined by dividing the caregiver’s weight in Kg with the square of the height measurement in metres (16). Data was entered into Microsoft Excel 2010 and analyzed using SPSS software package version 21 (Chicago, Illinois USA). Frequencies and percentages were used to evaluate the feeding practices. Chi-square analysis was employed to test for association between categorical variables. Statistical significance was set at p<0.05.