Identification of Patients in Need of Advanced Care for Depression Using Data Extracted From a Statewide Health Information Exchange: A Machine Learning Approach.

BACKGROUND: As the most commonly occurring form of mental illness worldwide, depression poses significant health and economic burdens to both the individual and community. Different types of depression pose different levels of risk. Individuals who suffer from mild forms of depression may recover without any assistance or be effectively managed by primary care or family practitioners. However, other forms of depression are far more severe and require advanced care by certified mental health providers. However, identifying cases of depression that require advanced care may be challenging to primary care providers and health care team members whose skill sets run broad rather than deep. OBJECTIVE: This study aimed to leverage a comprehensive range of patient-level diagnostic, behavioral, and demographic data, as well as past visit history data from a statewide health information exchange to build decision models capable of predicting the need of advanced care for depression across patients presenting at Eskenazi Health, the public safety net health system for Marion County, Indianapolis, Indiana. METHODS: Patient-level diagnostic, behavioral, demographic, and past visit history data extracted from structured datasets were merged with outcome variables extracted from unstructured free-text datasets and were used to train random forest decision models that predicted the need of advanced care for depression across (1) the overall patient population and (2) various subsets of patients at higher risk for depression-related adverse events; patients with a past diagnosis of depression; patients with a Charlson comorbidity index of ≥1; patients with a Charlson comorbidity index of ≥2; and all unique patients identified across the 3 above-mentioned high-risk groups. RESULTS: The overall patient population consisted of 84,317 adult (aged ≥18 years) patients. A total of 6992 (8.29%) of these patients were in need of advanced care for depression. Decision models for high-risk patient groups yielded area under the curve (AUC) scores between 86.31% and 94.43%. The decision model for the overall patient population yielded a comparatively lower AUC score of 78.87%. The variance of optimal sensitivity and specificity for all decision models, as identified using Youden J Index, is as follows: sensitivity=68.79% to 83.91% and specificity=76.03% to 92.18%. CONCLUSIONS: This study demonstrates the ability to automate screening for patients in need of advanced care for depression across (1) an overall patient population or (2) various high-risk patient groups using structured datasets covering acute and chronic conditions, patient demographics, behaviors, and past visit history. Furthermore, these results show considerable potential to enable preventative care and can be easily integrated into existing clinical workflows to improve access to wraparound health care services.

Pediatric decision support using adapted Arden Syntax.

BACKGROUND: Pediatric guidelines based care is often overlooked because of the constraints of a typical office visit and the sheer number of guidelines that may exist for a patient's visit. In response to this problem, in 2004 we developed a pediatric computer based clinical decision support system using Arden Syntax medical logic modules (MLM). METHODS: The Child Health Improvement through Computer Automation system (CHICA) screens patient families in the waiting room and alerts the physician in the exam room. Here we describe adaptation of Arden Syntax to support production and consumption of patient specific tailored documents for every clinical encounter in CHICA and describe the experiments that demonstrate the effectiveness of this system. RESULTS: As of this writing CHICA has served over 44,000 patients at 7 pediatric clinics in our healthcare system in the last decade and its MLMs have been fired 6182,700 times in "produce" and 5334,021 times in "consume" mode. It has run continuously for over 10 years and has been used by 755 physicians, residents, fellows, nurse practitioners, nurses and clinical staff. There are 429 MLMs implemented in CHICA, using the Arden Syntax standard. Studies of CHICA's effectiveness include several published randomized controlled trials. CONCLUSIONS: Our results show that the Arden Syntax standard provided us with an effective way to represent pediatric guidelines for use in routine care. We only required minor modifications to the standard to support our clinical workflow. Additionally, Arden Syntax implementation in CHICA facilitated the study of many pediatric guidelines in real clinical environments.

Experience in implementing the OpenMRS medical record system to support HIV treatment in Rwanda.

The challenge of scaling up HIV treatment in Africa has led to a new emphasis on improving health systems in impoverished areas. One aspect of this is the development and deployment of electronic medical record systems to support HIV and TB treatment. In this paper we describe the design and implementation of a new medical record architecture to support an HIV treatment program in rural Rwanda. The architecture is called OpenMRS and it has been developed to address the problem of configuring EMR systems to suit new sites, languages and diseases. OpenMRS uses a data dictionary called the concept dictionary to represent all the possible data items that can be collected. This allows new items to be added to the system by non-programmers. In addition, there are form creation tools that use drag and drop web technologies to simplify form construction. The OpenMRS system was first implemented in Kenya in February 2006 and then in Rwanda in August 2006. The system is now functioning well and we are developing extensions to improve the support for the clinic. These include improved, easy to use reporting tools, support for additional clinical problems including nutrition and child health, better database synchronization tools, and modules to collect laboratory data and support the pharmacy. The system is also in use in South Africa and Lesotho and is being deployed in Tanzania and Uganda.

The AMPATH medical record system: creating, implementing, and sustaining an electronic medical record system to support HIV/AIDS care in western Kenya.

Providing high-quality HIV/AIDS care requires high-quality, accessible data on individual patients and visits. These data can also drive strategic decision-making by health systems, national programs, and funding agencies. One major obstacle to HIV/AIDS care in developing countries is lack of electronic medical record systems (EMRs) to collect, manage, and report clinical data. In 2001, we implemented a simple primary care EMR at a rural health centre in western Kenya. This EMR evolved into a comprehensive, scalable system serving 19 urban and rural health centres. To date, the AMPATH Medical Record System contains 10 million observations from 400,000 visit records on 45,000 patients. Critical components include paper encounter forms for adults and children, technicians entering/managing data, and modules for patient registration, scheduling, encounters, clinical observations, setting user privileges, and a concept dictionary. Key outputs include patient summaries, care reminders, and reports for program management, operating ancillary services (e.g., tracing patients who fail to return for appointments), strategic planning (e.g., hiring health care providers and staff), reports to national AIDS programs and funding agencies, and research.

An Incremental Adoption Pathway for Developing Precision Medicine Based Healthcare Infrastructure for Underserved Settings.

Recent focus on Precision medicine (PM) has led to a flurry of research activities across the developed world. But how can understaffed and underfunded health care systems in the US and elsewhere evolve to adapt PM to address pressing healthcare needs? We offer guidance on a wide range of sources of healthcare data / knowledge as well as other infrastructure / tools that could inform PM initiatives, and may serve as low hanging fruit easily adapted on the incremental pathway towards a PM based healthcare system. Using these resources and tools, we propose an incremental adoption pathway to inform implementers working in underserved communities around the world on how they should position themselves to gradually embrace the concepts of PM with minimal interruption to existing care delivery.

Child Health Improvement through Computer Automation: the CHICA system.

Clinical guidelines are prevalent but frequently not used. Computer reminder systems can improve adherence to guidelines but have not been widely adopted. We present a computer-based decision support system that combines these elements: 1) pediatric preventive care guidelines encoded in Arden Syntax; 2) a dynamic, scannable paper user interface; and 3) a HL7-compliant interface to existing electronic medical record systems. The result is a system that both delivers "just in time" patient-relevant guidelines to physicians during the clinical encounter and accurately captures structured data from all who interact with the system. The system performs these tasks while remaining sensitive to the workflow constraints of a busy outpatient pediatric practice.

An Evaluation of Two Methods for Generating Synthetic HL7 Segments Reflecting Real-World Health Information Exchange Transactions.

Motivated by the need for readily available data for testing an open-source health information exchange platform, we developed and evaluated two methods for generating synthetic messages. The methods used HL7 version 2 messages obtained from the Indiana Network for Patient Care. Data from both methods were analyzed to assess how effectively the output reflected original 'real-world' data. The Markov Chain method (MCM) used an algorithm based on transitional probability matrix while the Music Box model (MBM) randomly selected messages of particular trigger type from the original data to generate new messages. The MBM was faster, generated shorter messages and exhibited less variation in message length. The MCM required more computational power, generated longer messages with more message length variability. Both methods exhibited adequate coverage, producing a high proportion of messages consistent with original messages. Both methods yielded similar rates of valid messages.

Targeted screening for pediatric conditions with the CHICA system.

OBJECTIVE: The Child Health Improvement through Computer Automation (CHICA) system is a decision-support and electronic-medical-record system for pediatric health maintenance and disease management. The purpose of this study was to explore CHICA's ability to screen patients for disorders that have validated screening criteria--specifically tuberculosis (TB) and iron-deficiency anemia. DESIGN: Children between 0 and 11 years were randomized by the CHICA system. In the intervention group, parents were asked about TB and iron-deficiency risk, and physicians received a tailored prompt. In the control group, no screens were performed, and the physician received a generic prompt about these disorders. RESULTS: 1123 participants were randomized to the control group and 1116 participants to the intervention group. Significantly more people reported positive risk factors for iron-deficiency anemia in the intervention group (17.5% vs 3.1%, OR 6.6, 95% CI 4.5 to 9.5). In general, far fewer parents reported risk factors for TB than for iron-deficiency anemia. Again, there were significantly higher detection rates of positive risk factors in the intervention group (1.8% vs 0.8%, OR 2.3, 95% CI 1.0 to 5.0). LIMITATIONS: It is possible that there may be more positive screens without improving outcomes. However, the guidelines are based on studies that have evaluated the questions the authors used as sensitive and specific, and there is no reason to believe that parents misunderstood them. CONCLUSIONS: Many screening tests are risk-based, not universal, leaving physicians to determine who should have a further workup. This can be a time-consuming process. The authors demonstrated that the CHICA system performs well in assessing risk automatically for TB and iron-deficiency anemia.

Evaluation of computer-generated reminders to improve CD4 laboratory monitoring in sub-Saharan Africa: a prospective comparative study.

OBJECTIVE: Little evidence exists on effective interventions to integrate HIV-care guidelines into practices within developing countries. This study tested the hypothesis that clinical summaries with computer-generated reminders could improve clinicians' compliance with CD4 testing guidelines in the resource-limited setting of sub-Saharan Africa. DESIGN: A prospective comparative study of two randomly selected outpatient adult HIV clinics in western Kenya. Printed summaries with reminders for overdue CD4 tests were made available to clinicians in the intervention clinic but not in the control clinic. MEASUREMENTS: Changes in order rates for overdue CD4 tests were compared between and within the two clinics. RESULTS: The computerized reminder system identified 717 encounters (21%) with overdue CD4 tests. Analysis by study assignment (regardless of summaries being printed or not) revealed that with computer-generated reminders, CD4 order rates were significantly higher in the intervention clinic compared to the control clinic (53% vs 38%, OR = 1.80, CI 1.34 to 2.42, p < 0.0001). When comparison was restricted to encounters where summaries with reminders were printed, order rates in intervention clinic were even higher (63%). The intervention clinic increased CD4 ordering from 42% before reminders to 63% with reminders (50% increase, OR = 2.32, CI 1.67 to 3.22, p < 0.0001), compared to control clinic with only 8% increase from prestudy baseline (CI 0.83 to 1.46, p = 0.51). Limitations Evaluation was conducted at two clinics in a single institution. CONCLUSIONS: Clinical summaries with computer-generated reminders significantly improved clinician compliance with CD4 testing guidelines in the resource-limited setting of sub-Saharan Africa. This technology can have broad applicability to improve quality of HIV care in these settings.

Changing course to make clinical decision support work in an HIV clinic in Kenya.

PURPOSE: We implemented computer-based reminders for CD4 count tests at an HIV clinic in Western Kenya though an open-source Electronic Medical Record System. Within a month, providers had stopped complying with the reminders. METHODS: We used a multi-method qualitative approach to determine reasons for failure to adhere to the reminders, and took multiple corrective actions to remedy the situation. RESULTS: Major reasons for failure of the reminder system included: not considering delayed data entry and pending test results; relying on wrong data inadvertently entered into the system; inadequate training of providers who would sometimes disagree with the reminder suggestions; and resource issues making generation of reminders unreliable. With appropriate corrective actions, the reminder system has now been functional for over eight months. CONCLUSION: Implementing clinical decision support in resource-limited settings is challenging. Understanding and correcting root causes of problems related to reminders will facilitate successful implementation of the decision support systems in these settings.

The OpenMRS Implementers Network.

OBJECTIVE: OpenMRS (www.openmrs.org) is a configurable open source electronic medical record application developed and maintained by a large network of open source developers coordinated by the Regenstrief Institute and Partners in Health and mainly used for HIV patient and treatment information management in Africa. Our objective is to develop an open Implementers Network for OpenMRS to provide regional support for the growing number of OpenMRS implementations in Africa and to include African developers and implementers in the future growth of OpenMRS. METHODS: We have developed the OpenMRS Implementers Network using a dedicated Wiki site and e-mail server. We have also organized annual meetings in South Africa and regional training courses at African locations where OpenMRS is being implemented. An OpenMRS Internship program has been initiated and we have started collaborating with similar networks and projects working in Africa. To evaluate its potential, OpenMRS was implemented initially at one site in South Africa by a single implementer using a downloadable OpenMRS application and only the OpenMRS Implementers Network for support. RESULTS: The OpenMRS Implementers Network Wiki and list server have grown into effective means of providing implementation support and forums for exchange of implementation experiences. The annual OpenMRS Implementers meeting has been held in South Africa for the past three years and is attracting successively larger numbers of participants with almost 200 implementers and developers attending the 2008 meeting in Durban, South Africa. Six African developers are presently registered on the first intake of the OpenMRS Internship program. Successful collaborations have been started with several African developer groups and projects initiated to develop interoperability between OpenMRS and various applications. The South African OpenMRS Implementer group successfully configured, installed and maintained an integrated HIV/TB OpenMRS application without significant programming support. Since then, this model has been replicated in several other African sites. The OpenMRS Implementers Network has contributed substantially to the growth and sustainability of OpenMRS in Africa and has become a useful way of including Africans in the development and implementation of OpenMRS in developing countries. The Network provides valuable support and enables a basic OpenMRS application to be implemented in the absence of onsite programmers.

The CHICA smoking cessation system.

Environmental tobacco smoke (ETS) exposure remains an important cause of morbidity and mortality in children. Pediatricians are well positioned to help smoking parents quit. Parents who smoke may be particularly responsive to advice to quit, repeated smoking cessation messages can be effective, and parents visit the pediatrician 8-10 times for well care in the first two years of their child's life. Yet most pediatricians do not provide smoking cessation advice. We developed a parental smoking cessation module for an established pediatric primary care decision support system (CDSS) that runs as a front-end to the Regenstrief Medical Record System. The system collects data directly from parents and guides the physician through smoking cessation counseling, using stages of change. We present the CDSS and the smoking module as well as descriptive data from our smoking cessation system. We also describe a randomized controlled trial of the system that is now underway.

Concept dictionary creation and maintenance under resource constraints: lessons from the AMPATH Medical Record System.

The challenges of creating and maintaining concept dictionaries are compounded in resource-limited settings. Approaches to alleviate this burden need to be based on information derived in these settings. We created a concept dictionary and evaluated new concept proposals for an open source EMR in a resource-limited setting. Overall, 87% of the concepts in the initial dictionary were used. There were 5137 new concepts proposed, with 77% of these proposed only once. Further characterization of new concept proposals revealed that 41% were due to deficiency in the existing dictionary, and 19% were synonyms to existing concepts. 25% of the requests contained misspellings, 41% were complex terms, and 17% were ambiguous. Given the resource-intensive nature of dictionary creation and maintenance, there should be considerations for centralizing the concept dictionary service, using standards, prioritizing concept proposals, and redesigning the user-interface to reduce this burden in settings with limited resources.

Using Arden Syntax and adaptive turnaround documents to evaluate clinical guidelines.

Clinical guidelines translate complex research findings and expert opinion into actionable recommendations. However, the effectiveness of even evidence-based guidelines is rarely tested as a whole in a real clinical environment. We have developed a decision support system for implementing clinical guidelines in a busy pediatric practice. We have added to this system the ability to randomize patients to receive care with or without system support of the guideline or guideline components. The randomization is part of the Arden Syntax that implements the system logic. The result is a relatively effortless process for testing guidelines, as they are implemented, to assure that they are effective. We describe the system and the process by which this guideline evaluation functionality was built in, using two guidelines (asthma management and maternal depression screening) both of which have been applied to thousands of patients to date.

Collaboration between the medical informatics community and guideline authors: fostering HIT standard development that matters.

Clinical guideline authors, health information technology (HIT) standards development organizations, and information system implementers all work to improve the processes of healthcare, but have long functioned independently towards realizing these goals. This has led to clinical standards of care that often poorly align with the functional and technical HIT standards developed to realize them. We describe the shortcomings and inefficiencies inherent in this current process and introduce two national initiatives that attempt to unite these communities. The mission of these two initiatives is to create examples of unambiguous, decidable, and executable clinical guidelines which both utilize and inform HIT terminology and logical expression standards. All of the products of this work aim to facilitate enterprise-wide guideline implementation and create a rising tide which lifts all ships.

Cooking up an open source EMR for developing countries: OpenMRS - a recipe for successful collaboration.

Millions of people are continue to die each year from HIV/AIDS. The majority of infected persons (>95%) live in the developing world. A worthy response to this pandemic will require coordinated, scalable, and flexible information systems. We describe the OpenMRS system, an open source, collaborative effort that can serve as a foundation for EMR development in developing countries. We report our progress to date, lessons learned, and future directions.

The OpenMRS system: collaborating toward an open source EMR for developing countries.

OpenMRS is an open source infrastructure for the creation of medical record systems in developing countries. Produced and maintained collaboratively across multiple institutions, this framework consists of an open source data model, a set of core application functions, and a default implementation. The goal of this implementation is to provide the beginnings of an EMR that is suitable for all groups involved with healthcare in developing countries.

Shortcomings in infant iron deficiency screening methods.

BACKGROUND: Screening for iron deficiency anemia is a well-established practice in pediatrics, but numerous challenges surrounding current recommendations raise questions about the effectiveness of this strategy. OBJECTIVE: To evaluate iron deficiency anemia screening approaches, by assessing rates of follow-up testing and resolution among patients meeting screening criteria in a primary care setting. METHODS: A retrospective cohort study was performed. We extracted electronic medical record data on complete blood counts for infants who received primary care in our clinics in the past 10 years. We calculated rates of positive screening results with 9 different measurement criteria and determined rates of follow-up testing and of documented correction of iron deficiency among those who screened positive. RESULTS: Our cohort consisted of 4984 children who were screened at 9 to 15 months of age, between 1994 and 2004. There was a wide distribution of positive detection rates (range: 1.5-14.5%) among the 9 screening criteria. Follow-up testing rates were low. No more than 25% of infants who screened positive by any criterion underwent a repeat complete blood count within 6 months. Moreover, no more than 11.6% (range: 4.4-11.6%) had documented correction of their laboratory abnormalities. CONCLUSIONS: Significant shortcomings exist in current iron deficiency anemia screening practices. A widely agreed-on, specific, and inexpensive screening criterion, with increased emphasis on systems-based approaches to iron deficiency screening, is needed.

AMPATH Medical Record System (AMRS): collaborating toward an EMR for developing countries.

Millions of people are either living with or dying from HIV/AIDS; most of this living and dying is taking place in developing countries. There is an immediate need for electronic medical record systems to help scale up HIV/AIDS prevention and treatment programs, reduce critical human errors, and support the research necessary to guide future efforts. Several groups are working on this problem, but most of this work is occurring within silos. To be more effective, we must find ways to collaborate. We describe a system built on the 30+ years of experience at Regenstrief Institute to serve as the seed for building toward a common infrastructure. We discuss the design goals, data model, and implementation of a data entry component. Further details are available online at amrs.iukenya.org.

Human and system errors, using adaptive turnaround documents to capture data in a busy practice.

Capturing coded clinical data for clinical decision support can improve care, but cost and disruption of clinic workflow present barriers to implementation. Previous work has shown that tailored, scannable paper forms (adaptive turnaround documents, ATDs) can achieve the benefits of computer-based clinical decision support at low cost and minimal disruption of workflow. ATDs are highly accurate under controlled circumstances, but accuracy in the setting of busy clinics with untrained physician users is untested. We recently developed and implemented such a system and studied rates of errors attributable to physician users and errors in the system. Prompts were used in 63% of encounters. Errors resulting from incorrectly marking forms occurred in 1.8% of prompts. System errors occurred in 7.2% of prompts. Most system errors were failures to capture data and may represent human errors in the scanning process. ATDs are an effective way to collect coded data from physicians. Further automation of the scanning process may reduce system errors.