1.1. Electronic Medical Record (EMR) Adoption in High-Income Countries

The potential benefits of electronic medical record systems (EMRs) in high-income countries have been well described. It has been suggested in the past that successful implementation may have a wide range of measurable benefits, including clinical decision support such as drug allergy warnings and drug incompatibilities [1], support for program monitoring including reporting outcomes, budgets, and supplies [2], support for clinical research, and management of chronic diseases such as diabetes, hypertension, and heart failure [3].

Additional studies in high-income countries have shown that EMRs have demonstrated significantly increased net monetary benefit per provider, cumulative net hospital savings, an increase in provider productivity, and improved provider efficiency [4, 5, 6]. A few studies have also shown evidence of improved clinical outcomes upon EMR adoption in high-income countries, including shorter hospital length of stay and lower 30-day mortality, a decrease in relative odds of death from myocardial infarction and CABG procedures, and improved outcomes for renal disease patients [7, 8, 9].

Despite the potential for many benefits, EMR adoption in high-income countries has also been associated with numerous drawbacks, including high cost of implementation and maintenance, changes to organizational workflow, temporary loss of productivity, and privacy and security concerns [10]. In addition, a 2014 survey of 6,375 American providers revealed that physicians who used EMRs had lower satisfaction with the amount of time spent on clerical tasks and higher rates of burnout [11]. Given these challenges and many others, it is important to the success of a given EMR to perform a detailed cost-benefit analysis.

1.2. EMR Expansion to Developing Countries

Given the potential benefits of EMR adoption demonstrated in high-income countries, efforts have been attempted to expand these systems to countries with limited resources. There are several examples of successful EMR and electronic health registry implementation in resource-limited hospitals in the current literature, including the Mosoriot Medical Record System in Kenya [12], PIH-EMR in Peru [13], HIV-EMR in Haiti, Careware in Uganda [14], the Baobab-ART (BART) system in Malawi [15], and an injury surveillance trauma registry in South Africa [16]. However, there is a general lack of studies that report on EMR systems that have been successfully implemented in hospitals in low- and middle-income countries (LMICs).

Despite successful implementation of the EMR systems described above, global adoption in resource-limited settings remains low. The reasons for this include: (1) resource and infrastructure limitations such as lack of reliable electricity, low-quality Internet access, and insufficient supply of devices; (2) lack of centralized organization such as a national health information technology (IT) agenda; (3) lack of explicit and broad legal regulation; (4) lack of common interoperability standards; and (5) lack of a trained workforce, including specialized IT workers and medical informaticians [17]. Additionally, EMR implementation represents a significant shift in workflow and organizational culture and thus may face significant barriers to adoption [18].

1.3. EMR Adoption in Guatemala

Guatemala is a country with poor health and economic indicators relative to the rest of Latin America and the rest of the world. Although Guatemala was reclassified by the World Bank in 2019 as an upper-middle-income country (UMIC) with a GNI per capita of US$4610, inequalities persist across geographic areas and among ethnic groups [19]. According to the most recent report in 2014 by the National Statistics Institute of Guatemala, 59.3% of the population of 16.58 million of Guatemala lived in poverty, and 8.7% lived below the lower-middle-income poverty line (defined as 8.3 in Guatemalan quetzal [2014] or US$5.5 per day in 2011 PPP) [20]. The GINI coefficient, a measure of inequality defined as the “extent to which the distribution among individuals or households within an economy deviates from a perfectly equal distribution,” was 0.43 as reported by the World Bank in 2014 and was consistent with moderate to severe economic inequality [19]. The life expectancy at birth was 72 years, compared to a life expectancy of 76 years for the associated World Health Organization (WHO) region [21].

Guatemala does not have a national universal health coverage policy or strategy, and it also lacks a national eHealth policy or strategy. EMR adoption in Latin America, including Guatemala, has been sparse, with only one example in the literature of an emergency surgery registry implemented in an urban hospital in Guatemala City [22]. However, after a one-year pilot in 2017 this registry was ultimately not adopted secondary to lack of optimization for hospital workflow, insufficient number of devices, and lack of end-user training.

Since 2015, the University of Virginia-Guatemala Initiative (UVA-GI) has worked on developing an EMR in a rural Guatemalan Hospital. Previous research groups identified Hospital Nacional Jose? Felipe Flores (Totonicapán Hospital) as a potential location for EMR implementation, and further researched what the required specifications for EMR software would be at this location [23, 24]. Totonicapán Hospital was selected for the EMR system pilot because of its small size, a previously established relationship with UVA-GI, and willingness to collaborate on software optimization based on provider feedback. These features reduced the complexity and costs of implementation and increased the likelihood of maintaining a sustainable project.

Totonicapán is a city of 500,000 residents in the hot and humid western highlands of Guatemala, approximately 100 miles from the Guatemalan capital, Guatemala City. An overwhelming 97% of the population identify as indigenous peoples speak the K’iche’ language, although Spanish is also widely spoken. Of households in Totonicapán in 2014, 40% were classified as living in extreme poverty, as compared to a national figure of 23.4% [20]. Totonicapán has many prominent population health concerns including high infectious disease prevalence, high infant mortality, and a lack of medical personnel and resources. The mortality rate by the age of 15 is 13.5% and only 44.5% of the population reach the age of 65 [21]. Totonicapán Hospital contains approximately 94 beds, cares for roughly 2,000–5,000 patients annually, and employs 28 doctors [23].

1.4. Objectives

To address the stated needs of the Totonicapán Hospital Emergency Department (ED), UVA-GI developed and implemented a point of care EMR named SABER [Simple, Accesible, Básico, Electronic Record] beginning in 2015. Given the need for more research on the implementation, evaluation, and continuing support for EMRs in developing countries and particularly in Guatemala, this study had three main objectives within the context of a resource-limited environment:

1. Evaluate the perceived benefits of EMR implementation.
2. Identify specific and unique challenges to the implementation of an EMR in such a setting.
3. Identify potential strategies that Totonicapán Hospital could use to overcome these challenges.

2.1. Software and Hardware Implementation

Implementation of the SABER program was conducted in multiple phases. First, we conducted a mixed-methods needs assessment with physicians, medical students, and hospital administration to determine the utility and feasibility of an EMR system in Totonicapán Hospital. Based upon the insight from previous research groups, an EMR, SABER 1.0, was developed in-house by two local Guatemalan UVA-GI programmers in 2015. The SABER interface is a Spanish language health record programmed primarily in PHP and Java that may be accessed via computers connected to a local area network (LAN) with a MySQL database. We chose to develop a novel EMR system as opposed to using an existing open-source platform (eg, OpenMRS, OpenEMR, GNUmed) because existing software was not easily customizable to unique needs of our hospital workflow, technological support for third-party programs was difficult to access internationally, and local computer engineers could offer on-site, in-person hardware and software support.

UVA-GI purchased a local Guatemalan internet connection for its program activities in the Totonicapán ED including cheap wireless routers to extend connectivity throughout the unit, but it was at that time not equipped to expand the internet service to the rest of the hospital or nearby clinics, laboratories, or other hospitals. Given these limitations, UVA-GI decided to pilot the SABER system exclusively in the Totonicapán Emergency Room. UVA-GI also purchased all of the necessary equipment to set up a fully functioning wireless LAN system, including a wireless LAN access point router, ethernet LAN switch, and ethernet cables and connectors. We also purchased a monthly subscription from a local internet service provider. Additionally, UVA-GI purchased an APC SMART-UPS 1000VA backup battery with the capability to provide up to 60 minutes of backup power for the system, a Canon iP2800 printer, a double-sided Canon laser printer, ink cartridges (color included), and multiple reams of paper. The total cost of these hardware components including internet service for one year was $1,584.

SABER collects data, including basic patient information, triage, initial evaluation, review of systems, physical exam, and evaluation and plan. It generates a PDF file based upon data entered into the browser form. The information collected is consistent with the patient information that is required to be reported to the Guatemalan Ministry of Health (GMH). Subsequent versions of SABER after the v1.0 pilot eliminated the need for filling out a paper chart in addition to inputting information in the EMR. These versions also addressed various bug fixes and expanded the capacity of the EMR to include order input. In the latest version of SABER (v. 6.0), all patient information stored electronically, including medication orders, can be printed directly for reporting to GMH. The GMH does not currently have capacity to accept or process data electronically, so printing capability is a critical feature of SABER.

2.2. Survey Development

In order to assess the continuing viability of SABER with regards to efficient collection of patient data and provider perceptions of the EMR system, we created a survey for medical students and emergency department healthcare providers. The survey was developed by the investigators, with guidance from the preliminary results of investigations conducted in prior years in the Totonicapán Hospital ED. We also consulted with other members of the UVA-GI team including healthcare professionals who had worked in Guatemala, in-country administrative professionals who work with the Totonicapán Hospital administration, and individuals involved in the networking and implementation of the SABER system. The survey was also informed by a systematic review of previous surveys that focused on electronic medical record adoption [25, 26].

The survey included questions designed to elicit basic demographic information, prior experience using an EMR, and Likert surveys evaluating providers’ perception of and experiences with the SABER EMR. A 5-point Likert survey was used to evaluate provider perceptions of the SABER EMR, whereas a 7-point survey was used to evaluate their self-reported proficiency with the program in order to better distinguish differences among computer skills.

The survey was administered in Spanish to 31 medical students on their clinical rotations between fourth and sixth year, and to six physicians in the Totonicapán ED. The surveys were given to the medical director of the ED, who subsequently distributed the surveys to the participants in-person. Participation in the survey was completely voluntary, and participants were instructed that they could withdraw their responses at any time. No personal identifying information was collected through the survey; participants were assigned a unique, randomized personal identification number. There was a total of 40 medical students and eight doctors working in the emergency department in 2018, thus yielding a survey participation of 80% and 75%, respectively. No participants declined to participate in the survey; participation was limited by providers who were off-service at the time of the investigation. Characteristics of survey participants are outlined in Table 1.

2.3. Qualitative Evaluation Process

In addition to surveys of healthcare providers in the Totonicapán Hospital Emergency Room, we also conducted a series of four focus group interviews in Spanish consisting of 4–6 individuals each of either physicians or medical students. Focus group interviews were conducted using a series of 10 scripted open-ended questions and lasted approximately 30–60 minutes. These interviews were facilitated by a medical student fluent in Spanish without prior relationship to the participants (the first author). Before taking part in focus group discussions, participants were informed of the purpose of the study and that their participation was completely voluntary and that their responses would be anonymous. Confidentiality and privacy of the participants were ensured. All focus groups were proctored by a professional transcriptionist who documented the general content of the discussion.

Initial analysis of focus group data was performed by the same individual who had conducted the interviews. The goal of the analysis was to identify themes related to the original objectives of the study. Attention was given to perceived benefits of and challenges to the SABER EMR, and common topics and/or challenges addressed by participants were given priority. The researcher who conducted the interviews engaged in ongoing discussion with other members of the research team, the director of the Totonicapán ED, and developers of the SABER program. Themes selected were: (1) organizational workflow; (2) critical technical system features; and (3) social and cultural system features that either helped or hindered EMR adoption. The results of both the survey and focus group interviews were deemed as critical to understanding the appropriate organizational issues, correct system features, and workplace culture needed to create an EMR that may be successfully adopted in a resource-limited setting.

3.1. Survey Results

Results of the survey as described in Tables 2 and 3 are summarized below.

3.2. Focus Group Results

Focus group results were organized into the three themes of organizational workflow, critical technical system features, and social and cultural system features as described previously.

4.1. Lessons Learned

The majority of participants had a favorable impression of SABER and seemed to indicate an understanding of why an EMR adoption could have a positive impact on both patient care and ED organization and workflow. However, we found that achieving these positive changes would require a shift in workplace culture and workflow not previously identified as a challenge. Unfortunately, nurses were not included as participants in the investigation because they were not end users at the time of study design. We learned during our investigation that future investigations must include nurses as key stakeholders to the future success of SABER. Their lack of inclusion in this study may have unfortunately amplified already existing political and cultural tensions among providers in the Totonicapán ED.

We also learned that efficient and effective hospital workflow is important in maintaining accurate medical records and continued use of an EMR. Without clearly defined roles for data input, oftentimes patient data was not collected in a timely fashion resulting in potential loss of information. Additionally, without a clearly defined triage system patients could end up waiting up to two hours to receive treatment and were often not identified as patients in the waiting room. This is a potential future application of SABER.

4.2. Recommendations

Two studies describing methods used to overcome social barriers to EMR adoption in developing countries appeared useful as analogues to the challenges faced by SABER. The first was a 2010 case study of an EMR system at a large hospital in India that used financial and social incentives to encourage chart completion using the EMR as opposed to the old method using paper charts. Use of the EMR was kept optional in order to reduce pressure on staff, but providers who chose to use the EMR received a small bonus as well as recognition in the EMR system via a “top 10” list. These incentives increased user participation without the need for administrative coercion [27].

A second cross-sectional study conducted in 2014 in six public and private hospitals in Saudi Arabia identified resistance to using new technology as a key social barrier obstructing EMR implementation. Importantly, the study results showed that 70% of excluded survey responses due to returned questionnaires were from nurses, indicating a relative lack of participation in the EMR adoption process. However, the authors of this study argued that nurses are “critical stakeholders who can affect EMR implementation either positively or negatively, and [that] significant attention should be directed to them in [EMR] adoption [28].” The authors found two key solutions to addressing user resistance. First was an adequate, strong, committed, and positive leadership team including engineers and training supervisors [29]. The second solution was conducting education training sessions to instruct the users [30]. These solutions are also applicable to the needs of the Totonicapán ED.

Based on the barriers to EMR implementation identified from survey responses, focus group discussions, and observation of delivery of care in the Totonicapán Hospital Emergency Room, we categorized our recommendations for future implementation of EMRs in low-resource settings into seven categories which had previously been identified as success criteria in such a setting in Table 4 [31]. Based on our observation that there was significant organizational and political resistance from providers in adopting the SABER electronic medical record, we suggest that these barriers to adoption be addressed first, before then addressing additional important technical, financial, ethical, functionality, and training barriers. Organizational and user resistance to EMR adoption in the Totonicapán ED appear to be mediating factors on other barriers.

4.3. Study Limitations

Given the size of the Totonicapán Emergency Department, we were faced with a limited sample size for our research given the small population size. For example, we only collected quantitative survey data from six doctors, which limited further statistical analysis of their responses. However, by collecting qualitative data from the participants we elicited valuable information regarding the continuing efficacy of SABER even with the limitations of our data analysis.

As previously mentioned, our investigation did not include nurses as participants as they were not end-users at the time the investigation was planned. However, adding nursing order functionality and encouraging nurses to use SABER would require input from nurses to best design a program that meets their unique needs. A successful EMR requires nursing feedback, investment, and participation [28]. Consequently, expansions to other units in the hospital will require similar input from future stakeholders.

We also recognize that SABER is an EMR in a very specific setting with a unique user interface, support structure, and patient demographic. Therefore, we caution against the generalizability of the results of this study when compared to EMR implementation in other care settings and/or countries. The hope is that this study provides information for others hoping to implement EMRs in low-resource settings, but not direct replication of our system in other settings without consideration of local contexts.

4.4. Future directions

We have several initiatives and research opportunities for future consideration that we argue would improve both quality of care and quality of data in the Totonicapán ED, including (1) expansion of SABER to include pharmacy data, (2) increasing provider education through development of an online and in-person laminated troubleshooting handbook, (3) integration with waiting room triage procedure, (4) correlation of SABER metadata to patient outcomes, and (5) potential for mobile health expansion. With these future initiatives, we hope to successfully expand SABER within the Totonicapán Emergency Department and eventually to the rest of the hospital, with the goal of providing a model for successful implementation of an EMR in a resource-limited setting.