Study area and period
This study was conducted in Amhara regional state teaching hospitals in Northwest Ethiopia, specifically in University of Gondar and Tibebe Ghion teaching hospitals, from March-April, 2020. The Amhara region is located in Northwestern and North central parts of Ethiopia. It has 10 adminstrative zones, one special zone, 181 woredas and 78 urban centers. According to the Amhara regional health office report in 2019, the region has two teaching hospitals, namely the University of Gondar and Tibebe-Ghion teaching hospitals. These teaching hospitals are the training centers for undergraduate and postgraduate medical students, dentists, nurses, midwives, pharmacists and others who shoulder the responsibilities to solve the health problems of the community and the country at large.
Study design and participants
An institution-based cross-sectional study design was employed. Medical interns studying in the University of Gondar and Tibebe Ghion teaching hospitals were participants of the study. But medical interns who were on a sick leave and week off during data collection period were excluded.
Sample size determination and sampling technique
The sample size was determined using single population proportion formula n= (Zα/2)2 (p (1-p)/d2 ; where n= the required sample size, level of confidence taken to be 95%, Zα/2 (1.96), a 5% margin of error (d=0.05), 50% proportion of EBM practice and n= the required sample size. The calculated sample size was 384 with 10% non-response rate, final sample size was 423.
A stratified random sampling technique was used to select participants from the University of Gondar and Tibebe Ghion teaching hospitals. By taking the list of medical interns from the human resource of each hospitals, we determine the proportionate sample to be taken, to proportionate number of study subjects for each hospital, the formula = (n) x (nf) / N was used where n = number of medical interns in each hospital, nf = total sample size and N = the total number of medical interns in the two hospitals. Then, considering their proportion of the population, simple random sampling method was applied.
Measurements and data collection techniques
Data was collected using self-administered questionnaire which had five sections; socio-demographic, EBM knowledge, attitude toward EBM, EBM practice and perceived barriers for practicing EBM. The questionnaire was adapted from previous published studies because these prior works have already been validated and can be used for an international comparison [6, 19]. Data was collected by a total of six health informatics students (BSc). The attitude of respondents was measured by 11 questions with five point Likert’s scale. All individual answers to attitudinal questions was computed to obtain total scores; then, mean score was calculated to categorize as having favorable attitude (if participants scored >=mean score) or unfavorable attitude (if Participants scored < mean score). In addition, respondents level of knowledge was measured by calculating the mean score of the 14 items and categorized as knowledgeable (if participants scored >=mean score of the correctly answered questions) or not knowledgeable (if participants scored <mean score of the correctly answered questions). The respondent’s level of practice was measured by 10 questions with five point Likert’s scale. All individual answers to practice questions was computed to obtain total mean scores and categorized as good practice (if participants scored >= mean score) or poor practice (if participants scored < mean score).
Data quality assurance and management
The data collectors and supervisors were trained prior to the actual data collection about the purpose of the study, sampling procedure, methods of data collection and ethical issues. Continuous follow-up and supervision were done by the supervisors and principal investigator throughout the data collection period. The data were checked daily for completeness and consistency.
The questionnaire was pretested on 5% of the similar population at Tikur Anbessa specialized hospital for consistency of response and validity of the questionnaire. Some adjustments were made while adapting to the Ethiopian context. Using the data obtained from the pre-test, the questionnaire was checked for reliability (internal consistency) using the Cronbach alpha test. The reliability for knowledge questions had a Cronbach’s alpha value of 0.839; attitude had a Cronbach’s alpha value of 0.764 and practice (Cronbach’s alpha 0.750). These values indicate that the questionnaire has very good reliability.
Data processing and analysis
The collected data were checked manually for comprehensiveness and uniformity of responses. Data were coded and entered in to Epi info 8.1 then exported to and analyzed by using SPSS 20.1. The descriptive statistical analysis was applied to compute the mean, standard deviation, frequency and percentages of the study variables. Bivariate analyses; Spearman’s rank correlation coefficient test was used to examine the bivariate correlations between the dependent and the independent variables. Selection of variables to be included in the final model was done by examining multi-collinearity among independent variables, with the purpose of excluding variables with strong correlation (rho-values >0.85). Finally, multivariable logistic regression analysis was done to control potential confounders and to identify the factors associated with practicing EBM. The variables which had significant association were identified on the basis of adjusted Odds Ratio (AOR), with 95% confidence interval (CI) and p-value ≤ 0.05.