Study area and period
The study was conducted from July 1, 2022 to July 30, 2022 in public health centers of Addis Ababa, Ethiopia. Addis Ababa has almost 4 per cent of the Ethiopian populations with 11 per cent of under five-year children. The total population of Addis Ababa is 3,602,000; 1,134,150 are under five children according to 2019 projection based on the 2007 census(14). In Addis Ababa there are around 116 health centers in eleven sub cites (15).
Study Design: An institutional based cross sectional study design was conducted
Study Population and eligible criteria
All mothers who have infants less than six months in Addis Ababa were considered in the study. Those mothers who have infants less than six months at selected health center were included in the study. Mothers who have repeated visits during the study period will be excluded from the study.
Sample Size Determination
The sample size required for the study is calculated using epi info application or the formula of simple proportion estimation: n = n = [(Zα/2)2 (p) (1-p)]/d2
Where; n- The minimum sample size required, P- proportion of bottle feeding associated factor in study done at woldia hospital, maternal age 35–50 years (23.4%)(16). Z𝛼2⁄- Standard normal value at (1- α) 95% confidence level, = 95%, margin of error (d) = 5% and design effect (de) = 1.5 is used,
The sample size “n” for associated factor of bottle feeding = (1.96)2(0.234) (1-0.234)/ (0.05)2 = 275, since study uses multi stage sampling, 1.5 design effect was used then the sample size would be 414.The final sample size after addition of 10% non-response rate was 456.
Sampling technique
Multi stage sampling technique was employed to include study participants in to the research. First from a total of 11 sub cities three sub cities were selected using simple random sampling technique. Since 20–30% of the study population represents the total population, average 25% of sub cities (SC) was used to represent the population.
In the first stage of sampling, 25% of eleven sub cities (three sub cities) Areda, Kolife Karanio and Nifas silk lafito sub cities were selected by using simple random sampling technique. These three sub cities have a total of 27 health centers (Areda = 7, Kolife Karanio = 9 and Nifas silk lafito = 11); 25% of 27 (seven) were used to allocate study participants in each selected sub cities. The seven health centers (HC) (Areda = 2, Kolife Karanio = 3, Nifas silk lafito = 3) are allocated according to the number of health centers in each sub cities.
In stage second sampling, Beta and Simegn Kebede health center from Areda sub city, woreda 1, 3and 4 health center from Kolife Karanio sub city and woreda 2, 5 and 10 HC from Nifas silk lafito sub city were selected by lottery method.
The registration of mothers who have infant less than 6 months on expanded program of immunization register was used to get list of children from each Health centers. Sample from each health center was determined by using proportion allocation to the sample size.
Finally, systematic random sampling technique was used to select a representative population in the immunization unit. The sampling interval (K) was established by dividing the average number of women who visited the immunization unit in the previous two months before data collection (average n = 2660) by the sample size, which produced K = 2660/456 = 6. Women were recruited daily by simple random sampling technique using the lottery method, and the next respondent was selected by adding a sampling interval to the number of the selected women. The same procedure was done on subsequent days until the required sample size was reached.
Proportional allocation
In three selected sub cities there were 2660 children less than 6 months of age. Out of this 335 in Beta health center, 277 in Simegn Kebede health center from Areda sub city, 360 in woreda 1 health center, 431 in woreda 3 health center and 245 from woreda 4 health center of Kolife Karanio sub city and 365 in woreda 2 health center, 347 woreda 10 health center and 300 from woreda 5 health center of Nifas silk lafito sub city.
Based on proportional allocation to size, 456 study subjects were distributed to Beta health center (57), Simegn Kebede health center(47) from Areda sub city; woreda 1 health center (42), woreda 3 health center(74) and woreda 4 health center(62) of Kolife Karanio sub city and woreda 2 health center (63), woreda 10 health center (60) and 300 from woreda 5 health center(51) of Nifas silk lafito sub city.
Data collection procedure
Data was collected by face-to-face interview using semi-structured pretested questionnaire. The questionnaire was administered through face-to-face interview by five (5) trained data collectors (diploma nurse) and the data collection was supervised by two BSC health worker.
The questionnaire was constructed by adopting and modifying from previous studies(2, 10, 12, 16). First, the questionnaire was prepared in English then translated into Amharic (local language) and then back to English to check consistency.
The quality of the data collection process was monitored, and clear uniform instructions were given to all data collectors. All data collected as a part of this study was checked by the principal investigator.
Operational definitions
Exclusive breastfeeding: A mother gives to infant only breast milk (5). Maternal illness: Mothers difficulty to breastfeed an infant due to illness(breast illness related problem) (5). Mixed breastfeeding: if an infant took breast milk with addition of liquid foods like cow milk and formula milk and soft foods(5). Adequate knowledge: If a mother answers half and above correctly from questions which are asked to measure bottle feeding knowledge (17). Inadequate knowledge: If a mother answers below half correctly from questions which are asked to measure bottle feeding knowledge (17). Late initiation of breastfeeding: initiation of breast milk after one hour after delivery(5). Bottle feeding: Mother uses bottle for Brest milk and or breast milk substitutes (formula or others)(18).
Data Quality Control
In order to assure the quality of the data; a pre-test was conducted on 5% of the sample size at Abinet health center from mothers’ infants visiting for immunization before the actual data collection to ensure the completeness and consistency. In addition, one-day training was given to data collectors and the supervisor about the aim of the study, about the data collection tools, and how to obtain information. The principal investigator undergoes quality assessment on ongoing supervision each day during data collection to ensure quality of data by checking filled format for their completeness’s and consistency. Data reliability was checked by considering the Cronbach alpha which is greater than or equal to 0.5 and Multicollinearity between variables was also checked by considering variance inflation factor (VIF) which is less than 2 which indicates there is no correlation between independent variables.
Data management and analysis
Data obtained from institution-based survey was checked for completeness & inconsistencies, then coded, entered and cleaned in the Epi info version 7.1 and Microsoft excel and analyzed using Statistical package for Social Sciences (SPSS) version 26. After categorizing and defining variables, descriptive analysis was carried out for each of the independent variables using frequencies, cross tabulation and percentages: then presented by text, tables and chart.
Binary logistic regression analysis model was used to assess the association between bottle feeding and associated factors. Variables with p-value < 0.2 in bivariable binary logistic regression analysis were entered to multivariable binary logistic regression analysis. Finally, variables with p-value < 0.05 in multivariable logistic regression were taken as significant factor. Adjusted odds ratio with 95% CI was used to show the strength of association between the dependent and independent variables. Model fitness was tested with Hosmer and Lemeshow Goodness of Fit test and fit with a p-value = 0.440. In addition, there was no problem of Multicollinearity among explanatory variables, with variance inflation factor < 3 for all variables.
Ethical considerations
Ethical clearance was obtained from the institutional review board of Saint Paul hospital millennium medical collage ethical review committee (ref.no. PML3/757, on 9/6/2022) and Addis Ababa health bureau public health research and emergency management department (ref.no. A/A/16172/227, on 23/6/2022). The study participants were appropriately informed about the purpose of study in their local language and provide their willingness by informed consent. The participants have the right to refuse or terminate at any time. All information obtained from the study participants were kept strictly confidential.