The study was conducted from September 2017 to June 2018 in selected health facilities found in East, West and Kellem Wollega districts. In these three districts, there were about 4,322,357 total populations with the mean of 1,440,786 in each districts with 1:1 gender ratio. From these total populations, there were 893,536 women of reproductive age group (15-45 years old) and about 710,163 under five children in these three Districts .
From the three districts, Seven health centers and six hospitals having ART and PMTCT clinics were randomly selected for the study. Data were collected by 13 ART nurses from the respective health institutions who were working at ART and PMTCT department. Three supervisors recruited to look after the overall of data collection. In these three selected districts, there were about 2,365 HIV positive women on antiretroviral therapy (ART), where a total number of mothers with HIV/ exposed infants (HEI) were 415.
Study Design, population and sampling
Cross-sectional study was conducted among randomly selected samples of 219 HIV positive mothers with children under six months and attending PMTCT and ART clinics in those selected public hospitals and Health centers found at East, West & Kellem Wollega Districts. In this study, we used a mixed methods for data collection involving both qualitative and quantitative data. Qualitative data were mainly used to triangulates the quantitative findings.
Inclusion criteria: Mothers/guardians living with HIV/AIDS having 6 or less months’ children, voluntarily consented to participate in the study were included.
Exclusion criteria: Mothers/guardians living with HIV/AIDS having 6 or less months’ children, who were not consented to participate and who were seriously ill and unable to provide information were excluded from the study.
Sample Size Determination
The study was conducted in randomly selected thirteen government owned health institutions (six hospitals and seven health centers) proving ART and PMTCT services for mothers living with HIV/AIDS. These institutions were selected considering availability of the services and adequacy of client flow.
The sample size (n) required for the study was calculated using a single population proportion with assumptions of 95% confidence interval, 5% desired precision, proportion of HIV positive mothers exclusively breast feed their infants 48.2%  and considering 10 % for compensation for non-response. The total samples calculated for the study was 219. These total samples were distributed to each selected health care institutions using probability proportional to size (PPS) and the numbers of HIV positive mothers required for the study in each PMTCT/ART clinics were determined. For each of the selected facility, sampling frame was prepared. Every respondent was selected from his/her respective health facilities by systematic random sampling technique from the sampling frame. Sampling intervals were determined by dividing N/n (415/219), i.e. every second intervals. The first respondent was selected by blindly picking one out of two pieces of paper numbered 1& 2. Mothers visited health facility for collecting their ARV medication and other purposes interviewed at every second interval.
Data Collection Procedures
An interview questionnaire was prepared from related literatures. Questionnaire was developed in English and translated into the local language then back to English to check for its consistency. Administration was done with “Afaan Oromo” versions. The questionnaire was pretested in similar setting. For qualitative data collection, discussion guide was used to facilitate focus group discussion.
Data were collected by 13 ART nurses from the respective health institutions who were working at ART and PMTCT department. Three supervisors and principal investigator performs overall controlling activities of data collections process.
Two days intensive training was given for data collectors and supervisor regarding the data collection process and tools. Data collection tools were pre-tested on mothers who were not included in the actual study in order to check its language clarity and consistency of questionnaire in similar set up.
Defining Study Variables
The dependent Variable was Exclusive Breast Feeding (EBF): Exclusive breast-feeding (EBF) is defined as the consumption of only breast milk with no supplementation of any type of food since birth except drops and syrups like; vitamins, minerals or medicines. For this study we have dichotomized EBF into ( ‘Yes’ and ‘No’). Those mothers exclusively breast fed their children for the first six months were labeled as ‘Yes.’ Those mothers who started complementary feeding withing the first six months were labeled as ‘No.’
Independent Variables assessed were: Socio economic status of mothers and households (educational status of mothers, income of mothers, antenatal care attendance, Occupation of mothers, disclosure of HIV status to spouse, age of mothers, Parity and mode of delivery), Mother’s Decision on the choice of infant feeding, Infant illnesses, baby hospitalization, birth weight, Hospital and Health service, Nutritional education, Antenatal Care, Post-Natal Care, Cultural norms on breastfeeding, child feeding practice and feeding system, mixed feeding (Breast -feeding with the addition of fluids, solid feeds and non-human milks in the first 6 months of age).
HIV Exposed Infants (an infant or child born to a mother living with HIV until the infant or child is reliably excluded from being HIV infected). HIV positive mothers (it refers to women belonging to the age group of 15 to 49 years who are on HAART, attending the antenatal and post-natal clinics in the selected hospitals and health centers.
Early termination of breast-feeding (the act of interrupting giving breast milk and making the child accustomed to other food before the child reach the age of two years). Seriously ill (Women who are sick, in bed and fail to give information).
Data processing and management
Data entered was done using EPI info Version 3.5.1, transferred to SPSS Version 20 for windows statistical package for analysis. Results were reported using standard data presentation techniques like frequency tables, measures of central tendencies and variations was done to see the nature of the data. Both bi-variable and multivariable logistic regression analysis were used to determine the association of each independent variable with the dependent variable. Candidate variables for multi-variable model were identified by considering P<0.05 at bi-variable analysis. Multi-variable logistic regression model was used to control for the effects of cofounders on the outcome variable. Associations were declared by assuming P<0.05 with their 95% confidence intervals and Adjusted Odds Rations (AOR) were reported to show the strength of associations were computed to identify the presence and strength of associations. Qualitative data were analyzed from voice records and the field notes. After checking, organizing, coding, conceptualizing and categorizing responses, then similar ideas were grouped and summarized based on thematic area and the key variables of the study. Concepts extracted from themes were presented in narrative and triangulated with quantitative results. Some quotes were presented in support to the quantitative findings with similar ideas.
Two days intensive training was given for data collectors and supervisors. Pre-testing of the questionnaire was done on mothers from non-selected health facility to test its clarity for both interviewer and respondent and to get experience to be applied with data collectors on actual data collection.
Ethical clearance was taken from the ethical clearance committee of Wollega University. Letter of clearance was obtained from each districts health office management to get the assurance of the study. Information sheet were attached to each questionnaire explaining the purpose, objectives, of the study were explained and written informed consent was secured from each participant.
Confidentiality was maintained at all levels of the study. Only consented participants were recruited on a voluntary basis. Participants wished to withdraw from the study at any point were informed to do so without any restriction.