OUP user menu

Open source electronic health records and chronic disease management

Jason C Goldwater , Nancy J Kwon , Ashley Nathanson , Alison E Muckle , Alexa Brown , Kerri Cornejo
DOI: http://dx.doi.org/10.1136/amiajnl-2013-001672 e50-e54 First published online: 1 February 2014

Abstract

Objective To study and report on the use of open source electronic health records (EHR) to assist with chronic care management within safety net medical settings, such as community health centers (CHC).

Methods and Materials The study was conducted by NORC at the University of Chicago from April to September 2010. The NORC team undertook a comprehensive environmental scan, including a literature review, a dozen key informant interviews using a semistructured protocol, and a series of site visits to CHC that currently use an open source EHR.

Results Two of the sites chosen by NORC were actively using an open source EHR to assist in the redesign of their care delivery system to support more effective chronic disease management. This included incorporating the chronic care model into an CHC and using the EHR to help facilitate its elements, such as care teams for patients, in addition to maintaining health records on indigent populations, such as tuberculosis status on homeless patients.

Discussion The ability to modify the open-source EHR to adapt to the CHC environment and leverage the ecosystem of providers and users to assist in this process provided significant advantages in chronic care management. Improvements in diabetes management, controlled hypertension and increases in tuberculosis vaccinations were assisted through the use of these open source systems.

Conclusions The flexibility and adaptability of open source EHR demonstrated its utility and viability in the provision of necessary and needed chronic disease care among populations served by CHC.

Keywords
  • open source
  • chronic disease
  • chronic care model
  • community health center
  • electronic health record

Background and significance

Chronic disease is the leading cause of death and disability in the USA, with over 133 million people having at least one chronic illness, and 25 million being disabled as a result of their condition.1 The increasing prevalence of chronic disease has increased the cost of healthcare for patients as the estimated total cost is currently US$659 billion, divided between direct expenditures and indirect costs, which accounts for lost productivity and non-reimbursed personal costs.2 Despite these costs, care for chronic illness is poor, as the aging of the population, combined with the increase in risk factors such as obesity, smoking and lack of an active lifestyle, has accelerated ahead of the advances that could assist in the management and treatment of these diseases.3 Furthermore, there is a growing chasm in the type of care provided and the quality of outcomes achieved with chronic disease among various socioeconomic and ethnic groups. Studies have indicated that 51% of Hispanic and 46% of African-American individuals over 55 years of age are limited by chronic disease, compared to 23% of white individuals.4

In addition, those with income levels 200% below the poverty level often lack insurance, and receive less preventive care, have lower vaccination rates and have a higher incidence of chronic illness in addition to poor quality and outcomes.5 The prevalence of certain risk factors, such as tobacco use, sedentary lifestyle, body weight, elevated blood pressure and excessive consumption of alcohol are positively related to adults with a low socioeconomic status.6 Furthermore, the rates of chronic disease and morbidity and mortality are also strongly associated with socioeconomic status.6

These individuals are usually served by safety net providers, which include community health centers (CHC), who have invested health information technology (IT) to improve the management in chronic disease resulting from the adoption and implementation of electronic health records (EHR).7 The benefits of EHR technology have been particularly pronounced in CHC where many patients are transitory and experience the effects of poor chronic disease management. Many centers extensively use registry functionality and disease management/preventive care capabilities to create effective quality improvement strategies to assist patients in the treatment of their chronic conditions.8 These functions include providing reminders at the point of care for providers to identify high-priority clinical areas for patients.

However, the functionality itself is usually not sufficient to manage chronic disease appropriately. Many successful interventions in chronic disease management adhere to the chronic care model (CCM) developed by Ed Wagner at the MacColl Institute for Healthcare Innovation, and require the delegation of responsibility by the primary care doctor to team members for ensuring that patients receive the appropriate clinical and self-management support services.9 The features of an EHR to coordinate care effectively between team members, provide evidence-based clinical management, support comprehensive self-management and extend the amount of time clinical team members spend with patients underscore the importance of this technology to manage chronic care. While a number of commercial EHR systems contain this functionality, the use of open source EHR systems, such as the Resource Patient Management System (RPMS) and OpenMRS have been particularly effective in assisting CHC to redefine the way chronic care is delivered within their setting by aligning to these tenets, has made a noticeable improvement in patient outcomes.

The term ‘open source’ means that the source code of an application is available for anyone to review, critique, modify, and redistribute to others.10 The open source approach to software development is more flexible and transparent than other processes and products may more closely reflect the collective interests of a community.11 Open source is sometimes inappropriately referred to as public domain, which is software that is released without any licensing agreement that accompanies it. Source code within an open source product is distributed through a licensing agreement. The original authors or developers of the product do not release their intellectual property rights over the original code. Private products (that are not publically distributed modification and enhancements to open source code) can thus remain private.

NORC at the University of Chicago was contracted by the Office of the National Coordinator for Health Information Technology in April of 2010 to undertake a study and report on open source health IT systems to assess and study open source EHR within safety net settings. The report was designed to gain a larger understanding of the experience of CHC in the acquisition, implementation and utilization of open source EHR to achieve meaningful use and to serve their patient populations. However, in the course of the study, it was also discovered that these systems could also be effectively modified to support the CCM and improve chronic disease management.

Methods and materials

The NORC team developed a methodology for a comprehensive environmental scan that included a literature review of relevant academic and trade literature, white papers, non-published literature, copies of testimony by experts in the areas of open source, speeches and presentations. The search terms included ‘open source EHR’, ‘Veterans Health Information Systems and Technology Architecture’, ‘Federally Qualified Health Centers (FQHCs),’ ‘open source health IT communities’, and ‘RPMS’, among others. NORC also conducted searches using terms relating to specific classification of open source health IT systems, including ‘medical information’ and ‘practice management systems’. Overall, we identified over 200 abstracts that we pared down to 127 articles by removing articles that did not have an identified author or did not reference at least one topic related to open source and health IT.

The NORC team also conducted a series of key informant interviews with recognized experts in open source technology and healthcare, as well as conducting site visits to FQHC that were currently using, or were planning on using, an open source EHR. FQHC is a federal designation from the Health Resources and Services Administration, and refers to a health center that is located in or serves a federally designated medically underserved area/population.12 The site visits assisted the team in better understanding the factors that affected the utilization of open source, and provided insight into consistent practices, lesson learned or emerging trends. In particular, the visits identified strategies for overcoming barriers to selecting, implementing and using open source health IT systems. NORC developed a set of criteria to identify appropriate themes, participants, instruments and metrics to be used in the course of a visit. These criteria included the following:

  • The site must have a current open source EHR implementation with an active community of users.

  • The site must serve both Medicare and Medicaid beneficiaries, as well as uninsured patients.

  • The EHR must have the ability to incorporate e-prescribing and clinical decision support functionalities.

  • The sites represent a variety of technical and delivery approaches.

  • The site locations are geographically diverse.

As shown in Otable 1, there were a number of sites that met the criteria established by NORC, and during the visit two of these sites (Primary Care Systems, Inc. and JWCH Institute) were found to be using their EHR for chronic disease management.

View this table:
Table 1

Site visit locations

Name of facilityLocationType of open source EHR
Primary Care Systems, Inc (part of Community Health Network of West VirginiaClay, West VirginiaRPMS (MedLynks)
JWCH Institute, IncLos Angeles, CaliforniaOpenMRS
Wesley Community Health CenterPhoenix, ArizonaWorldVistA
Clinca AdelanteSurprise, ArizonaWorldVistA
Operation SamahanNational City, CaliforniaClearHealth
  • EHR, electronic health record; RPMS, resource patient management system.

Primary Care Systems, Inc. is a 501(c)(3) CHC that serves the residents of Clay County, West Virginia, and other surrounding areas. Primary Care Systems serves approximately 7200 patients with about 30 000 patient encounters annually. Of the patients served, over 70% are covered by Medicare or Medicaid, or are uninsured. The staff of Primary Care Systems currently included four full-time physicians and four full-time mid-level staff providing a range of primary care services, including laboratory, radiology, behavioral health, maternity and well-child services.13 JWCH Institute, Inc. is also a 501(c)(3) CHC that was established at the Attending Staff Association of the John Wesley County Hospital in 1960 by Los Angeles county physicians. The agency works to improve the health and wellness of under-served groups of the population of Los Angeles county through the direct provision or coordination of healthcare, health education services and research.14 The organization works through a variety of programs and activities, such as: medical outreach and referrals for medical care; HIV services and drug treatment; health education; psychosocial assessment and intervention; primary medical care; family planning services; and research.

NORC created a detailed semistructured questionnaire for both the informant interviews and the site visits. These protocols included open-ended questions to encourage informants to share their experiences and concerns. Each protocol identified specific themes to be covered to ensure that the appropriate areas were the focus of this study. Guidance and direction for this study were provided by a technical expert panel that was formed during the inception of the project, and included a list of recognized technical and content experts in the area of open source and health IT.

Results

The CCM, developed by Ed Wagner at the MacColl Institute for Healthcare Innovation, conceptualizes care as prepared practice teams in productive interactions with informed, activated patients.15 Six elements are important to help accomplish this objective: the (1) overall healthcare organization, which must support a (2) redesigned delivery system and (3) modern clinical information systems; (4) systematic decision support; (5) self-management support for patients and (6) links to available community resources. A number of studies have demonstrated that incorporated elements of the CCM into practice have shown modest improvements in patients with chronic disease in organized healthcare settings. In particular, the CCM has a relative effect on process and outcomes with chronic care in modest doses, such as having care teams adhere to established practice guidelines for chronic care published by organizations such as the American Diabetes Association, or reminders about blood glucose tests, eye examinations, or blood pressure checks that are sent to patients on a regular basis.16 Furthermore, in some studies, the presence of an EHR was not independently associated with improved process or outcomes. This supports a hypothesis, presented in previous reports, that an EHR does not improve chronic care unless it is supported by the redesign of care delivery within the healthcare setting; the incorporation of quality and process improvement activities across the care continuum; and the use of specific functionality within the EHR, such as clinical reminders and alerts. Through the site visits conducted with this study, the NORC team observed the intersection of an open source EHR with the incorporation of a number of the CCM elements as well as those noted as necessary to improve chronic care management appreciably, particularly in the area of diabetes and tuberculosis.

In 2010, there were approximately 32 million people in the USA with diabetes.17 West Virginia has the highest prevalence rate of diabetes in the USA, with a 12.5% rate as compared to a 7.5% rate nationally. Over 170 000 patients have currently been diagnosed with the disease, with over 80 000 estimated to have diabetes who are unaware of their condition. Furthermore, over 66% of those who were diagnosed with diabetes also had hypertension, while 62.8% had hyperlipidemia.18 These three conditions are known as the ‘West Virginia triad’ and are present among a significant majority of the population. The total cost of diabetes within the state approximates almost US$1 billion per year, with a third of these costs stemming from indirect costs such as lost work productivity, and two-thirds a direct result of medical bills.

In 2006, Primary Care Systems, Inc. became the first CHC within the USA to implement an adapted version of the RPMS, which they rebranded as MedLynks. This open source system was a health center-configured version of the RPMS platform that had been used by the Indian Health Service (IHS) to improve health outcomes for tribal populations within ambulatory care settings. MedLynks had templates and tools that could be adapted for CHC, such as clinical reminders and alerts. Primary Care Systems also redesigned their care delivery system to accommodate the functionality within MedLynks, and the emphasis on chronic disease management by formulating coordinated patient care teams, in conjunction with the CCM.

In 2005, they prepared for the implementation by training staff on the CCM as well as realigning clinicians within care teams. Care managers and coordinators were trained on disease management processes and patient self-management techniques. These processes were used to guide the configuration of MedLynks with specific emphasis on the clinical reminder functionality, health factor reports and patient education materials. Primary Care Systems leveraged the RPMS community for the modifications to the core system along with enlisting the assistance of a third-party vendor. By using the modified system, each member of the care team was aware of the necessary tests or consultations that would be needed by each patient so that effective care management was coordinated between each of the members and that necessary and appropriate support for self-management of the condition was provided. The use of these care teams led to greater concordance with treatment protocols and assisted patients in becoming better self-managers of their chronic illnesses.

In 2005, the number of Primary Care System patients whose diabetes could be measured as ‘controlled’ was at 68.5% compared to the national average of 70.0%. In 2007, the center achieved an increase to almost 80% in the number of patients with diabetes whose condition was controlled. The number of diabetes patients being evaluated increased by 33% during that same time period. In addition, in 2005, Primary Care Systems captured the body mass index (BMI) number in 18 patients who were children or adolescents who were diagnosed as obese, with only eight being referred to weight management counseling. In 2007, over 150 children or adolescents had their BMI recorded and were referred to counseling. The use of MedLynks assisted the clinic in identifying individuals who were missing a hemoglobin BA1c test or would automatically calculate a BMI score. The clinical reminders or alerts would then help guide the care team and patient to the appropriate treatment and creation of a specifically tailored care plan.

Homelessness has become a major factor in the resurgence of tuberculosis within the USA over the past two decades. This social status is associated with an increased risk of exposure to Mycobacterium tuberculosis, undetected and untreated infection, and subsequent progression to tuberculosis disease. Homeless persons with tuberculosis are more likely to be hospitalized longer than non-homeless persons, with the public sector paying most of the costs.19 In 2010, 11 182 (per 100 000 population) cases of tuberculosis were reported within the USA, with 10 541 of those cases being reported by state health departments as coming from individuals classified as homeless. The highest concentration of homeless individuals diagnosed with tuberculosis was identified as California.20

The Skid Row Homeless Healthcare Initiative was a partnership of 26 agencies in Los Angeles that strove to integrate services, reduce suffering and improve the health of homeless people in and around the community known as Skid Row, a 52-block area of downtown Los Angeles where it is estimated that 8 000–10 000 homeless people reside. Located less than a mile east of the center of downtown, the homeless that are assembled in this area have access emergency shelters, transitional and permanent housing, social and medical services, and food provided by a cluster of agencies that serve the largest homeless community in the USA.21

The JWCH Institute is a partner within the Skid Row Homeless Healthcare Initiative and took on the development of a community project to create a tuberculosis registry that is an IT system that allows medical, social service, and housing providers access to shared tuberculosis records to provide a medical home, coordinate services and reduce unnecessary duplication. JWCH chose OpenMRS, a community developed, open source, enterprise electronic medical record system framework.22 JWCH funded WebReach, Inc., to develop a customized version of OpenMRS, which they named OpenMRSLA and included a tuberculosis module to help track the status of the homeless population they serve.

JWCH also created a comprehensive data sharing agreement with the Los Angeles Department of Social Services and the Department of Housing and gave authorized users access to OpenMRSLA so that they could identify where an individual was currently living, what medical program (such as Medi-Cal or HealthyWay LA) they were on or eligible for, and whether their tuberculosis status was current. Although at the time of the site visit there was limited information on the effectiveness of the program, the information gathered through the medical directors at JWCH indicated that they felt that over 75% of their patient population was in OpenMRSLA and that each of them were current on their tuberculosis vaccinations.

Discussion

The results of this study demonstrated both the impact and the effectiveness of an open source EHR on managing chronic disease. It also provided insight into how these systems could be adapted and modified to incorporate elements of the CCM and accommodate changes in care delivery to manage chronic disease more effectively. In table 2, the breakdown of each of the elements of the CCM and their alignment to both MedLynx and OpenMRSLA are shown.

View this table:
Table 2

Alignment of the CCM to open source EHRs

CCM elementMedLynxOpenMRSLA
Create a culture, organization and mechanisms that promote safe, high quality careUpon implementation of MedLynx, care team were formed to provide chronic care with their responsibilities mapped to the EHRThe OpenMRSLA system was specifically redesigned for JWCH Institute Inc., to track and monitor the tuberculosis status of the homeless population they serve
Assure the delivery of effective, efficient clinical care and self-management supportMedLynx incorporated clinical reminders and alerts for the management and treatment of diabetes, hypertension and hyperlipidemia as well as supporting collaborative care between the provider and patientJWCH coordinated housing and social services along with OpenMRSLA to track and identify those individuals who needed a current tuberculosis vaccination or who had previously received one
Promote clinical care that is consistent with scientific evidence and patient preferencesThe clinical reminders and alerts used with MedLynx are aligned with current evidence-based practice in diabetes careTuberculosis is a treatable and manageable condition and inoculations were given to those who were found within OpenMRSLA to not be in compliance
Organize patient and population data to facilitate efficient and effective careMedLynx would evaluate each patient's vital signs and medical history to determine the procedures and/or tests need for diabetes careOpenMRSLA would coordinate housing, social service and medical data to determine the tuberculosis status of an individual, the medical program they were on, and their current housing situation
Empower and prepare patients to manage their health and healthcareEducational modules with MedLynx provided information to patients about appropriate diabetes careBy tracking the tuberculosis and housing status of an individual, JWCH could provide more effective outreach
Mobilize community resources to meet needs of patientsPrimary Care Systems used MedLynx to coordinate care within the community health network of West Virginia and provide information on a patient's chronic care statusJWCH worked within the larger SSRHI initiative to improve health services for the homeless by providing a patient's current condition and tuberculosis status through OpenMRSLA
  • CCM, chronic care model; EHR, electronic health record; SSRHI, Skid Row Homeless Healthcare Initiative.

A significant number of EHR have the necessary functionality to support effective chronic disease management, so the question remains as to why an open source EHR was utilized in these settings and is viewed as more effective in aligning with the CCM. For Primary Care Systems, Inc., the prevalence of diabetes, hypertension and hypolipidemia was so significant that it was necessary to redesign the way of delivering care in order to provide the appropriate preventive care and care management services to patients. The incorporation of the MedLynx system enabled the providers and clinicians to form care teams, utilized clinical reminders and alerts to tailor individual care plans, and extended the amount of time a provider spent with a patient. For the JWCH Institute, it was imperative to utilize system that would coordinate vital demographic information with tuberculosis status, as well as interfacing with social services to identify the housing status of each homeless individual. The specific use of an open source EHR helped accelerate these changes as it enabled both organizations to redesign their care delivery systems before the acquisition and implementation of their respective systems, with the ability to modify these systems to accommodate these changes. The value of open source is that those changes were made to support each center and the modification aligned with those changes before the installation to support chronic disease management. This is akin to designing and building a house, in which the acquisition of a pre-built model will allow a family to use the rooms within the house as they see fit, but they will not be able to make any significant modifications. However, if they were able to build the house themselves, they could structure the design to reflect the needs of their family. An open source system allowed both of these facilities to structure their chronic disease programs in a manner that was effective and aligned with the CCM model. Over time, each organization noticed a significant difference in the compliance their patient population had with prescribed diabetes treatment protocols and tuberculosis vaccinations. The prevalence rate for each decreased and the care teams in both organizations were able to provide more individualized and effective care.

Conclusion

Even with these advantages, both Primary Care Systems and JWCH had staff with sufficient technical expertise to help oversee the necessary modifications and implement the system successfully. In addition, there were strong clinical champions within each facility who focused on the need to support successful chronic disease management through the EHR and a redesign of their care delivery system. Without this expertise and dedication, the use of the open source system may not have been as successful nor provided the benefits that both organizations realized. Further research is needed to examine open source EHR in other settings and compare them directly to commercial systems to understand both the benefits and disadvantages as they pertain to the management of chronic disease.

Contributors

NJK assisted in the development of the methodology that guided this study, developed the content on results from the site visits, and reviewed and gave final approval before the paper was submitted. AN assisted in the development of the methodology for the site visits and key informant interviews, assisted in the development of the background and significance section, and reviewed and gave final approval before the paper was submitted. AEM assisted in the development of the site visit protocols, assisted in the development of the results section, and reviewed and gave final approval before the paper was submmitted. AB developed the background section on open source, contributed content on open source governance, and reviewed and gave final approval before the paper was submitted. KC conducted a significant part of the literature review for the project and the paper, assisted in the development of the key findings and conclusions, and reviewed and gave final approval before the paper was submitted.

Funding

This work was supported by the Department of Health and Human Services, Office of the National Coordinator for Health Information Technology, contract number OS28589.

Competing interests

None.

Provenance and peer review

Not commissioned; externally peer reviewed.

Acknowledgements

The authors would like to acknowledge Wil Yu for his guidance and support as well as Margaret Moore and Maya Das for their contributions.

References

View Abstract