A multicenter cross-sectional study was conducted to evaluate attitudes and associated factors related to EMR implementation among health professionals in EMR-equipped hospitals in Ethiopia. The study was conducted in 3 selected government hospitals actively using fully integrated EMR systems: Gizawu Hospital, located in Debre Berhan town, North Shoa, Amhara Region; Yekatit 12 Hospital Medical College; and Tirunesh Beijing General Hospital. Data were collected from April 1 to April 30, 2023.

The study population included health professionals from the 3 hospitals. The total sample size was proportionally allocated to each hospital based on the number of health professionals. A systematic random sampling approach was used to select participants who met the inclusion criterion of work experience of more than 6 months. Individuals on maternity leave, annual leave, sick leave, and those providing voluntary services were excluded. The sample size was determined using the single population proportion formula, accounting for a 5% nonresponse rate [13]. The initial sample size was 397 health professionals.

Study participants were selected using systematic random sampling within each profession. A detailed explanation of the sampling interval calculation, randomization process, and statistical justification is provided in Multimedia Appendix 1.

A pretested, structured, and self-administered questionnaire was used. The questionnaire was developed based on standard templates derived from a review of published articles [14-18]. Seven data collectors (5 GPs, 2 MPH graduates) and 2 trained GP supervisors monitored the data collection process. A comprehensive pilot study was conducted to ensure instrument validity and reliability. This included content validation by health informatics specialists (CVI=0.81), test-retest reliability (ICC=0.82), and cognitive interviews to refine question phrasing. Full details of the instrument validation, pilot testing, and comparative validation data are available in Table S1 in Multimedia Appendix 1.

Ethical approval was obtained from the Ethical Clearance Review Board of the Asrat Woldeyes Health Science Campus, Debre Berhan University (protocol number IRB 01/125/2015), in accordance with the Helsinki Declaration guidelines and regulations. Permission was obtained from the participating hospitals. Written informed consent was obtained from all participants. The authors have maintained the privacy and confidentiality of research subjects’ data and/or identity. The participants of this study were not compensated.

The collected data were entered into EpiData (version 4.6) and exported to SPSS (version 25.0) for analysis. A chi-square test was initially performed, followed by bivariable analysis. Variables with P<.25 in the bivariable analysis were included in the multivariable analysis. Adjusted Odds Ratios (AORs) with 95% CI were used to assess associations. Model fitness was evaluated using the Hosmer-Lemeshow test (P=.80).

The dependent variable was attitude toward EMR.

The independent variables were age, computer literacy, knowledge of EMR systems, refresher training on EMR, EMR functionality interruptions, adequate computer access, management support for EMR use, and presence of an ICT center.

The following terms and definitions were used in this study:

The study population consisted of health professionals from 3 hospitals, with the total sample size proportionally allocated to each hospital based on the number of health professionals (Table 1). A total of 382 health professionals participated in the study, with a response rate of 96.2%. Nonrespondent profiling was assessed by comparing the demographics of 39 nonrespondents with the sample frame using HER data. No significant differences were found in gender (P=.32), profession (P=.47), or EMR frequency (P=.19). To address concerns about causal inference in this cross-sectional study of EMR implementation attitudes, we implemented a multilayered methodological approach that combines design-based safeguards, advanced statistical techniques, and theoretical triangulation. These included EMR implementation timelines aligned with exposure measurement and hospital-specific rollout periods (6‐18 mo before the survey) using health ministry deployment records. We also captured technology adoption histories through work experience items (eg, “years using digital health tools”) to verify exposure duration.

The majority of respondents were male (n=236, 61.8%). The mean age was 30.04 (SD 4.43) years. Nearly two-thirds of the participants (n=246, 64.4%) had less than 6 years of work experience, and 69.6% (n=266) held a Bachelor of Science degree. A high proportion of study participants (n=339, 88.7%) demonstrated computer literacy, as measured by self-reported proficiency and task completion rates, and 26.6% (n=113) reported previous experience working in private hospitals (Table 2).

Among study participants, 251/382 (65.7%) demonstrated good knowledge of EMR (Table 3, Multimedia Appendix 2). A segmented bar chart showing health professionals’ knowledge of EMR system capabilities with 95% CIs (N=382) is shown in Figure 1. Notably, most respondents were aware of the diverse functions of EMR, including handling patient data, reviewing patient problems, obtaining test results, enabling private ward access, facilitating local network operations, and updating patient information. Furthermore, approximately three-quarters of participants were knowledgeable about EMR functionalities, including laboratory order requests, prescription capabilities, and disease codes, as well as data backup systems (Table 3).

Among the 382 study participants, only 184 (48.2%) demonstrated a favorable attitude towards the implementation of EMR (Multimedia Appendix 3). Please note that Multimedia Appendix 3 intentionally simplifies the data for visualization purposes only. Attitude responses on a 5-point Likert scale were grouped into 3 categories: positive (scores 4‐5), neutral (score 3), and negative (scores 1‐2). Of the 382 respondents, 210 (55%) expressed a positive attitude toward EMRs, 62 (16.2%) had a neutral attitude, and 110 (28.8%) reported a negative attitude. These distributions are detailed in Table 4 and represented in Figure 2. Percentages are accompanied by their absolute values (N=382). Study participants expressed strong agreement on several benefits of EMR, including cost-saving potential (n=117, 30.6%), ease of data entry (n=99, 25.9%), facilitation of timely decisions (n=93, 24.3%), acknowledgment of the need for training (n=166, 43.5%), reduction of medical errors (n=100, 26.2%), enhancement of productivity (n=106, 27.7%), and suitability for audit trials (n=110, 28.8%). Additionally, 156 (40.8%) strongly agreed, and 151 (39.5%) agreed that other hospitals should implement EMR (Table 4). Distribution of health professionals’ overall attitudes toward EMR use across three Likert scale categories: positive (4-5), neutral (3), and negative (1-2) as represented in Figure 3. Error bars represent 95% CIs for proportions (N=382). This 3-category structure reflects respondents’ sentiments and attitudes shown in Table 4.

A subgroup analysis regarding patterns of EMR attitudes by age group, profession, and facility type is visualized for all 17 benefit dimensions in Multimedia Appendix 4. We assessed multicollinearity using variance inflation factors (VIF), with all values ranging from 1.1 to 4.7, below the threshold of 5. Moderate correlations (eg, between years of EMR experience and technology self-efficacy) were addressed by combining variables into composites or residualizing overlapping predictors. Robustness checks, including principal component regression and elastic net regularization, confirmed the stability of the results.

Most respondents reported the fulfillment of certain organizational factors: 73% mentioned the assignment of an EMR focal person and duty person; 77.7% believed that management supports EMR implementation; and 86.4% witnessed the presence of an ICT (Information and Communication Technology) center. However, less than half of respondents received refreshment training on EMR (48.2%); only 38% perceived adequate computer access for EMR, and 39.8% responded that they had regular meetings (Table 5). A clustered bar chart (Figure 4) displays the percentage of health professionals in Ethiopian hospitals who reported technology and organizational readiness for EMR implementation across ten key factors in 2023. For each factor, the proportions of respondents answering “Yes” (indicating presence or support) and “No” (indicating absence or lack of support) are shown side by side. The chart visually highlights areas of strength, such as ICT center availability and management support, as well as gaps, including regular meetings on EMR and refresher training.

The study revealed a strong preference for EMR over paper-based systems among participants. A significant majority (320/382, 83.8%) reported finding EMR systems advantageous over disadvantageous for several reasons (Figure 5). Note that participants could select more than one option.

These advantages included:

Although most participants favored EMR, some reported challenges. Figure 6 highlights the reasons why a minority of participants expressed reservations about EMR use. The main concerns were the time required for data entry (mentioned as a reason by some who did not prefer EMR). Again, participants can select more than one option.

All independent variables underwent initial computation using the chi-square test, and variables with a P value ≤.05 were selected for bivariate analysis. Among these, variables with a P value of ≤.25 were considered for multivariate analysis. Following adjustment for other variables, the multivariable logistic regression analysis revealed that respondents aged ≤29 years were three times more likely to exhibit positive attitudes towards EMR implementation compared to those aged ≥35 years (AOR –3.05; 95% CI −1.58 to 5.9; P=.001). Furthermore, individuals who received EMR training were 2.8 times more inclined to have a favorable attitude towards EMR implementation than those without training (AOR −2.87, 95% CI 1.80‐4.56; P<.001). Negative signs before AOR values indicate inverse associations (eg, prior poor EMR experience reduces favorability).

The study administered EMR training with role-specific standardization (eg, physicians vs. nurses) but allowed site adaptations for workflow variations. While core curricula were maintained across sites using learning management systems, content delivery mixed future iterations, e-learning, and blended approaches based on staff roles and complexity levels. Health professionals with computer literacy displayed a 2.7 times higher likelihood of holding positive attitudes towards EMR implementation compared to those without computer skills (AOR 2.66; 95% CI 1.16‐6.09), while those with good knowledge of EMR were 1.8 times more likely to have a favorable attitude towards EMR than those with poor knowledge (AOR 1.8; 95% CI 1.10‐2.96) (Table 6).

The forest plot from the regression analysis (Table 6), showing AORs and CIs for attitude predictors, is represented in Figure 7. The forest plot presents the adjusted odds ratios (AORs) and 95% confidence intervals for predictors of positive attitudes toward electronic medical record (EMR) use among health professionals in EMR-implementing Ethiopian hospitals (N=382). Significant predictors identified in the analysis include younger age (≤29 years), computer literacy, participation in refresher training, and good EMR knowledge (P<.05) (Figure 7).

EMRs represent a significant advancement beyond traditional paper-based records, offering enhanced coordination and communication among health care departments and professionals, ultimately improving the quality of patient care [1,19]. This study found that just over half (51.8%) of health professionals in the surveyed Ethiopian hospitals reported a favorable attitude toward electronic medical records (EMRs). This rate, while indicating partial acceptance, is significantly lower than that observed in high-income nations like the United States (97%) and other developing countries such as Saudi Arabia (70%) and South Africa (67.2%) [10,17,20]. This disparity likely reflects differences in resource availability, baseline technology access, and national digital health infrastructure.

Our finding is also lower than some previous studies within Ethiopia, with reported favorability rates between 50.6% and 72.8% [16,19,21,22]. This variation may stem from our study’s focus on facilities with full EMR implementation, where staff experience the complete benefits and burdens, unlike settings with partial implementation or where staff are only anticipating a future system.

The primary predictors of a positive attitude were computer literacy (AOR 2.66), EMR training (AOR 2.87), and younger age (AOR 3.05). This finding is crucial, as it underscores that EMR acceptance is not passive; it must be actively cultivated through foundational training and skills development.

The strong influence of training and literacy suggests that in the Ethiopian context, fundamental barriers such as the skills to use the system remain the primary challenge. This contrasts with research in contexts like Canada, where factors like collaborative care and financial incentives are more prominent once baseline literacy is assumed [23]. The finding that younger clinicians (“digital natives”) are more receptive aligns with prior adoption models [16,22,24]. This highlights an urgent need for established, experienced professionals who may be less familiar with digital systems but whose buy-in is critical for successful implementation.

Interestingly, factors like management support, computer access, and the presence of ICT centers were not significantly associated with attitudes in our model. This finding contrasts with other studies in the Amhara region and Dire Dawa [5,21,22]. This discrepancy may suggest that while access to hardware is a prerequisite, it is not sufficient to foster positive attitudes. Rather, the skills and training to use that hardware effectively are what drive professional acceptance [16,23].

Our findings generate specific, actionable recommendations for Ethiopia’s digital health strategy.

This study has several limitations. The sample was drawn from urban hospitals, which may limit generalizability to rural settings with different infrastructure. Furthermore, the cross-sectional design is a snapshot in time and cannot establish cause-and-effect relationships; professionals with a positive attitude may be more likely to seek out EMR training, rather than the other way around.

To address these limitations, future research should employ qualitative methodologies. Longitudinal ethnography could track EMR adoption over time to identify evolving attitudes. We also recommend Participatory Action Research (PAR) to co-design training protocols with frontline staff, which can foster stakeholder buy-in. Finally, narrative inquiry—collecting stories from both "resistors" and "early adopters"—could reveal deeper, more nuanced experiential perceptions of EMR utility that our survey may have missed.

In this study, nearly half of the respondents had low attitudes towards the implementation of EMR. Variables related to socio-demographic (health professionals’ age), personal (health professionals’ knowledge of EMR), technology-related (health professionals’ computer literacy), and organization-related (health professionals’ EMR training) factors were significantly associated with health professionals’ attitudes towards the EMR system. 

Based on the findings of this study, it is important to focus on the gaps identified to implement EMR systems efficiently and effectively in all Ethiopian hospitals. Enhancing health professionals’ attitudes and contextualizing EMR training in the health care system are highly recommended for scaling up EMR use. For future researchers interested in EMR-related research in Ethiopia, employing mixed methods (quantitative and qualitative) could be highly beneficial.