📄 medicine

Current practices, challenges and enabling factors for electronic medical record implementation in Uganda: a convergent mixed- methods study

This convergent mixed-methods study of four Ugandan health facilities reveals that inconsistent Electronic Medical Record (EMR) adoption is driven primarily by facility-level conditions like staff shortages, time-consuming workflows, and weak IT support rather than one-time training, necessitating strategies that prioritize reliable infrastructure, responsive technical escalation, and continuous, workflow-embedded training for sustainable scale-up.

Original authors: Moses Mugerwa, Priase Nimusiima, Dong (Roman) Xu, Dadong Wu

Published 2026-08-14

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Moses Mugerwa, Priase Nimusiima, Dong (Roman) Xu, Dadong Wu

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a world where every time you visit a doctor, your medical history is instantly available on a screen, helping the doctor make the best choice for your health. This is the promise of Electronic Medical Records (EMRs). Think of them as a super-powered digital diary for your health, replacing dusty paper files that can get lost, damaged, or left in the wrong room. In many places, these digital diaries are already the norm, but in countries with fewer resources, getting them to work smoothly is like trying to build a high-speed train on a bumpy dirt road. The challenge isn't just buying the computers; it's making sure the electricity stays on, the internet doesn't vanish, and the doctors and nurses actually want to use the system instead of sticking to their old paper notebooks. This is the story of how technology meets reality in the real world of healthcare.

Now, let's zoom in on a recent investigation in Uganda, a country working hard to get these digital health records running in its hospitals and clinics. The researchers wanted to know: Why does the system work great in some places but barely get used in others? They didn't just look at the software; they looked at the people, the power outlets, and the daily chaos of a busy clinic. They studied four different health facilities, asking 302 health workers about their experiences and watching how they actually worked.

Here is the big surprise: The researchers found that simply giving someone a one-time training class on how to use the computer doesn't guarantee they will use it. It's like teaching someone to drive once and then expecting them to be a perfect driver forever without any practice or a mechanic nearby. The study showed that the most important thing for using the system was experience. The more years a worker had spent using the EMR, the more likely they were to use it every day. In fact, workers with zero experience were almost never using the system, while those with three to five years of experience were using it constantly.

But even the experienced workers faced some huge hurdles. The biggest problems weren't that the software was "bad" or "confusing." Instead, the problems were about workload and broken tools.

The Double-Book Problem: Imagine you have to write a story in a notebook, and then immediately type the exact same story into a computer, just in case the computer crashes later. That's what the workers were doing. Because the computer reports didn't match the official paper forms the government needed, they had to do the work twice.
The Power Outage: When the electricity or internet went out, the whole system stopped. Workers said, "If the screen goes black, we just go back to paper." One hospital that had solar power (a backup battery system powered by the sun) had much higher usage than the others, proving that reliable power is just as important as the software itself.
The "Help Me" Void: When the computer acted up or got slow, there was often no one to call for help. The study found that workers felt they had no reliable technical support and no "peer champions"—friendly experts on the staff who could show them the ropes when things got tricky.

The researchers also looked at what helped the system work. They found that the environment was actually quite weak in support. Less than one in four workers felt they had a strong support system. The things they missed the most were reliable IT support (someone to fix the tech quickly) and peer support (a colleague to ask for help).

So, what is the takeaway? The paper suggests that to make these digital health records stick, we need to stop thinking that "training" is a one-time event. Instead, we need to treat electricity and internet like essential medical tools—just as important as a stethoscope or a bandage. We need to fix the workflow so doctors don't have to write everything twice, and we need to build a team of tech-savvy helpers right inside the clinics. The technology is ready, but for it to truly work, the whole ecosystem around it needs to be fixed first.

Technical Summary: Current Practices, Challenges, and Enabling Factors for Electronic Medical Record Implementation in Uganda

Problem Statement

Despite the recognized potential of Electronic Medical Records (EMRs) to strengthen health service delivery in low- and middle-income countries (LMICs), real-world adoption in Uganda remains uneven. While the Ministry of Health, supported by the METS programme at Makerere University, has deployed the open-source UgandaEMR+ (based on the OpenMRS 3 framework) to approximately 1,000 of the country's 8,400 health facilities, routine use is inconsistent. Deployment does not guarantee adoption; facilities often suffer from parallel paper-based documentation, intermittent system use due to infrastructure failures, and a workforce lacking ongoing technical support. Existing literature often relies on single-facility case reports or high-income country contexts, leaving a scarcity of primary, multi-site, mixed-methods evidence that simultaneously quantifies specific barriers and elucidates the qualitative mechanisms of implementation failure or success in routine clinical practice.

Methodology

The study employed a convergent mixed-methods design conducted between March and May 2025 across four purposively selected public health facilities in Uganda: Lira Regional Referral Hospital, Hoima Regional Referral Hospital, Kyangwali Health Centre IV, and Kawaala Health Centre IV. These sites were chosen to maximize variation in facility level, geography, and deployment duration.

Quantitative Strand: A cross-sectional survey of 302 health workers (stratified by cadre: clinical officers, nurses/midwives, lab technicians, pharmacy staff, and records officers) was administered via KoBoToolbox. The instrument measured:
- EMR use frequency (dichotomized into "high" vs. "low" use).
- A 10-item Perceived Challenges scale (CH1–CH10) covering infrastructure, workload, and support.
- A 10-item Perceived Facilitators scale (FA1–FA10) covering leadership, training, and technical support.
- Statistical analysis included descriptive statistics, Pearson's chi-square tests, and multivariable binary logistic regression. Internal consistency was verified (Cronbach's α = 0.84–0.86).
Qualitative Strand: 10 semi-structured key-informant interviews (KIIs) with facility leaders, clinicians, data officers, and national implementers, alongside structured workflow observations in outpatient, maternity, lab, pharmacy, and records departments. Data were analyzed using reflexive thematic analysis with 20% double-coding.
Integration: Quantitative and qualitative findings were merged using a joint-display matrix to identify convergent, complementary, or divergent patterns, framed within the NASSS (Non-adoption, Abandonment, Scale-up, Spread, and Sustainability) framework.

Key Results

1. EMR Use Patterns and Predictors

Facility Variation: EMR use differed significantly by facility ( $\chi^2, p=0.005$ ). High-use proportions ranged from 29.6% at Kawaala HC IV to 60.0% at Kyangwali HC IV.
Primary Predictors: Years of EMR experience was the strongest predictor of high use ( $p < 0.001$ ). All respondents with 3–5 years of experience were high users, while 100% of those with zero experience were low users.
Training Impact: One-time formal training showed no significant association with high EMR use ( $p=0.465$ ). Facility-level conditions and accumulated experience were the only significant independent predictors in the multivariable model.

2. Perceived Challenges

Overall Severity: The mean perceived challenge score was 3.56 (SD 0.61), with 75.2% of respondents reporting moderate to high challenge levels.
Top Barriers: The most endorsed challenges were workload-related rather than purely technological:
1. Staff shortages at peak periods (58.3%).
2. Time-consuming data entry (58.3%).
3. Duplicate documentation (57.6%).
Infrastructure: Electricity instability, network issues, and system downtime were also significant but secondary to workflow capacity issues.

3. Perceived Facilitators

Overall Weakness: The mean facilitator score was low (2.66, SD 0.85), with fewer than one in four respondents perceiving a strong support environment.
Critical Gaps: The most strongly absent facilitators were reliable IT support (57.9% disagreed it existed) and peer-champion support (59.6% disagreed).

4. Qualitative Insights
Four themes contextualized the quantitative data:

Hybrid Workflows: Parallel paper-digital documentation is the norm because EMR outputs do not align with national reporting forms (e.g., HMIS 105).
Infrastructure Fragility: Power outages and network failures trigger immediate reversion to paper, creating backlogs that discourage future digital use.
Weak Escalation: There is no structured pathway for technical escalation; facility data officers are often the only resource and lack capacity for system-level fixes.
Training Deficit: Training is treated as a one-time event at deployment, with no follow-up for new staff, shift rotations, or system updates.

Key Contributions

Empirical Evidence on Training: The study challenges the assumption that initial training drives adoption, demonstrating that accumulated user experience and facility-level conditions are more predictive of sustained use than one-off workshops.
Workload-Centric Barriers: It identifies that the primary implementation barrier is not the software itself, but the mismatch between EMR documentation requirements and existing workflow capacity, specifically the burden of duplicate documentation.
Support Infrastructure Gap: The research highlights that the absence of responsive IT escalation pathways and peer-champion networks is a critical failure point, even in facilities where the software is installed.
NASSS Application: By applying the NASSS framework, the study moves beyond a checklist of barriers to a sociotechnical understanding of how technology, organization, and context interact to determine sustainability.

Significance and Claims

The paper claims that EMR implementation in Uganda is shaped more by the local implementation ecology than by the technology alone. The authors argue that for UgandaEMR+ to move from variable adoption to sustainable, routine clinical use, national strategies must shift focus:

Infrastructure as Safety: Treat power and connectivity as critical clinical safety infrastructure, not just IT utilities.
Support Systems: Build responsive technical support with clear escalation pathways and peer-champion networks as core implementation infrastructure.
Workflow Alignment: Align EMR reporting outputs with national requirements (e.g., HMIS) to eliminate the need for duplicate documentation.
Training Redesign: Shift from one-off training events to continuous, workflow-embedded learning models that include mentorship and competency-based refresher modules.

The study concludes that addressing these implementation ecology factors is essential for scaling EMRs effectively in Uganda and similar LMIC contexts.