The American Medical Informatics Association (AMIA) held its annual symposium in San Francisco this year. AHRQ Health IT grantees, contractors, and staff were featured in symposium workshops, panels, and demonstrations, which addressed the “Current Priorities” listed below. One of our panels also highlighted some cross-cutting work that made use of EHR log data to answer research questions in the areas of provider burden, workflow effectiveness, and health IT safety.
I also had the opportunity to speak with AMIA leaders about some of our emerging areas of interest, including artificial intelligence (AI). In addition to co-sponsoring the JASON report "Artificial Intelligence for Health and Health Care" (PDF, 2.5 MB), AHRQ funds a number of investigator-initiated grants that address how AI can improve health care. Simply search our AHRQ-funded projects using search terms like ‘machine learning’ or ‘natural language processing’ to find descriptions of our AI-related grants.
AI is also a personal interest of mine and I’m happy to report that I did receive one of the books I’d asked for this holiday season: The Deep Learning Revolution. Since I’ve been hosting holiday company, I’ve been able to read only the first few chapters so far, but it’s very clear to me that, as with other industries already affected by AI, health care will face tremendous disruption and we need to start planning for and conducting research around how best to integrate new, AI-related technologies. If you have any thoughts about priority areas where health IT research is needed, feel free to send them to me: email@example.com.
Chris Dymek, Ed.D.
The Health IT Program at AHRQ provides foundational research to ensure that health IT is designed and implemented in ways that improve quality and safety, while not resulting in excessive burden on physicians and other members of the care team. Current research areas include health IT usability, consumer health IT, clinical decision support, health IT safety, and patient-reported outcomes.
AHRQ’s usability research focuses on how to design and implement electronic health records (EHRs) so that they are more intuitive to use and more readily support clinical workflow. Two areas of research, supported by the AHRQ projects below, are how to effectively reduce documentation burden for physicians and how to make data within EHRs more usable for clinical decisionmaking.
- Many physicians are overly burdened by writing notes in EHRs. To address this problem, Dr. Thomas Payne, Professor of Medicine at the University of Washington and past Board Chair of the American Medical Informatics Association, implemented and tested a new voice-generated enhanced electronic note system (VGEENS), which integrates voice recognition and transcription with natural language processing and links to the EHR to improve note creation efficiency and note accuracy.
- Reviewing clinical notes, a necessary part of making diagnostic and therapeutic decisions, is hindered by many factors including the sheer volume of electronic clinical data in the EHR, Discovery and Visualization of New Information from Clinical Reports in the Electronic Health Record Project Profilesuboptimal text user interface design, and limited time to interact with patients. Dr. Genevieve Melton-Meaux at the University of Minnesota is refining computational methods to identify new information in clinical notes and then assessing a new tool to help clinicians better visualize that new information, thereby improving clinicians’ efficiency, decision-making, and satisfaction with documentation mechanisms in the EHR.
- A National Web Conference on Effective Design and Use of Patient Portals and their Impact on Patient-Centered Care
- Addressing Hospital Patient Information Needs Using a Personal Health Record Portal Project Profile
- Engaging Diverse Patients in Using an Online Patient Portal Project Profile
- Improving Healthcare Quality with User-Centric Patient Portals Project Profile
Increasingly, innovative computer and information systems are being developed to help people manage health concerns, monitor important indicators of their health, and communicate with their caregivers. AHRQ supports research to determine how these patient-facing technologies can best improve the quality and effectiveness of care. For example, a recent AHRQ National Web Conference highlighted the results of AHRQ’s patient portal research and described portal impacts on patient engagement and satisfaction; and identified features that make it both easier and harder for patients to use patient portals; and showed ways to make patient portals more responsive to patient needs.
In particular, Web Conference attendees heard about the following evidence, useful for informing their patient portal implementations:
- How a plain-language encyclopedia hyperlinked directly to unfamiliar terms, such as endometriosis, was used frequently by patients accessing their medical records via a portal and provided particular value and enhanced satisfaction; and
- How a tailored online video instruction for utilizing particular portal functions, for example how to log in to the portal and review the visit summary or medication list, can help overcome health literacy issues that prevent effective portal usage.
Clinical decision support (CDS) makes new and targeted evidence available to clinicians at the point of care, facilitating their ability to act more readily on this information. We know that CDS, when inappropriately implemented, can lead to alert fatigue, high override rates, and physician frustration. When effectively implemented, CDS provides the right information to the right audience in the right way and at the right time. AHRQ has a long history of investing in research about how to make CDS more effective and usable. In 2016, AHRQ embarked on an ambitious multi-component program to advance patient-centered CDS and to make CDS a more effective tool: to help advance research evidence into clinical practice and to help make the right thing to do, the easy thing to do.
AHRQ is also creating a CDS repository called “CDS Connect,” which will be the first national, public resource and platform for sharing CDS, including its interoperable, standards-based building blocks. Those developing, implementing, and testing CDS will be able to use CDS Connect to share their CDS and to learn from each other’s experiences.
Research is needed in two key areas: how health IT can improve patient safety and how health IT itself can be safely used and implemented. AHRQ-funded research provides critical evidence in both areas, and the Program recently renewed a special emphasis notice reaffirming the intent to support research regarding the safety of health IT systems. For example, Dr. Jason Adelman, Chief Patient Safety Officer and Associate Chief Quality Officer at Columbia University Medical Center, is investigating whether multiple, simultaneous, open records by individual users can affect the likelihood of wrong patient errors, and where a balance might exist between safety and efficiency. Hospitals need such practical research evidence to guide policy decisions when implementing EHRs, especially when those policy decisions can impact patient safety.
For many chronic illnesses, major treatment goals become not only lengthening a patient’s life but also maximizing the individual’s ability to function normally, reducing symptom burden, and improving patient health-related quality of life. The incorporation of brief, valid patient-reported outcome (PRO) measures to assess functional outcomes, such as symptom amelioration (e.g., reduced pain), elimination of emotional distress, and optimal physical and social functioning, is essential to the provision of meaningful and quality health care for individuals. While some EHR systems are currently able to capture some structured PRO data, such information is not commonly collected and integrated at the point of care. Thus, PRO data are not routinely available for clinical care, research, and quality improvement. AHRQ recently released funding opportunities for how to collect and use PROs using health IT, as well as scale and spread existing health IT models that currently incorporate PROs. Current AHRQ-funded research in this area includes the following projects:
- Developing Design Principles to Integrate Patient-Reported Outcomes (PROs) Into Clinical Practice Through Health Information Technology: Data, User Experience, and Workflow Requirements for PRO Dashboards Project Profile
- Rheumatology Informatics System for Effectiveness Patient-Reported Outcome (RISE PRO) Dissemination Project Profile
- Patient Outcomes Reporting for Timely Assessments of Life with Depression: PORTAL-Depression Project Profile
- Dr. David Flum, Associate Chief Medical Officer at the University of Washington, and his team are assessing healthcare practitioner perspectives and workflow needs to inform design principles for the meaningful integration of PROs into clinical practice;
- Dr. Jinoos Yazdany, Associate Professor in the Division of Rheumatology, Department of Medicine at University of California, San Francisco is using a novel health IT platform of the Rheumatology Informatics System for Effectiveness (RISE) patient registry to scale, spread and evaluate the collection of PROs across a nationwide network of rheumatology practices; and
- Dr. Neda Laiteerapong, Associate Director at the Center for Chronic Disease Research and Policy at the University of Chicago Medicine is integrating the Computerized Adaptive Test for Mental Health (CAT-MH) into an EHR to evaluate the effectiveness of collecting depression symptoms via a patient portal (MyChart) versus usual collection during clinic visits in two randomized controlled trials.
The United States has invested heavily in the widespread adoption and use of health IT. The next critical step is maximizing this investment by gathering evidence on how best to utilize health IT to generate, integrate, and synthesize disparate electronic data and evidence to support systems and processes that continuously improve patient outcomes. Health systems that effectively apply data and evidence to improve patient outcomes are called “learning health systems (LHSs).”
- Achieving Individualized Precision Prevention (IPP) through Scalable Infrastructure Employing the USPSTF Recommendations in Computable Form Project Profile
AHRQ’s health IT program explores how health IT can best support LHSs, particularly in ambulatory care settings. AHRQ issued a funding opportunity calling for exploratory and developmental research in this area. Current AHRQ-funded research in this area includes work conducted by Dr. Charles Friedman, Chair of the Department of Learning Health Sciences and the Josiah Macy Jr. Professor of Medical Education at the University of Michigan Medical School, who is applying LHS methods to adapt the US Preventive Services Task Force guidelines and to develop individualized precision prevention tools and recommendations.
AHRQ Division of Health IT Past Achievements
Since 1968, AHRQ has invested in research grants and contracts awarded to over 180 institutions in 46 States and the District of Columbia, producing foundational work on many uses of IT to improve health delivery including telemedicine, health information exchange, and e-prescribing.
Select examples of some of the many contributions of AHRQ’s health IT work follow.
Telehealth can extend patient care by enabling clinicians to engage with patients or with other teams of clinicians remotely when they cannot meet in person. Project ECHO developed and evaluated a tele-consultation platform to provide clinical support to rural primary care physicians to help them to care for complex patients. Through Project ECHO, rural physicians were able to care for patients with Hepatitis C, achieving outcomes equal or exceeding patients cared for at an academic medical center. The model has been expanded to provide specialty consultation to primary care teams caring for people with asthma, heart failure, chronic pain, mental health conditions, and complications of pregnancy. The VA is now implementing Project ECHO across the country to improve access to quality care.
Health Information Exchange
Health information exchange (HIE) allows doctors, nurses, pharmacists, other health care providers, and patients to access and share a patient’s medical information electronically across organizations and systems, for example between the emergency department and physician’s offices. This can enable better coordination of care and increased safety. AHRQ funded the first large-scale demonstrations of HIE in the United States. The projects developed best-practices regarding all aspects of HIE including establishing health information organization, technical considerations for implementation, and demonstrating value. These projects laid the foundation for national health information exchange efforts and informed the ONC State HIE grant program. There are currently active HIEs in many states, and the success of these HIEs is due to the lessons learned from AHRQ’s investments.
E-prescribing allows clinicians, pharmacies, and health plans to communicate prescription information electronically, which can make coordinated care safer and more efficient. AHRQ funded the evaluation of a statewide program that implemented an e-prescribing system that indicated whether medications were covered by insurance and found it to be cost saving. Another project developed an electronic system to create and securely transmit prescriptions for controlled substances. This research was used by DHHS and the Drug Enforcement Agency (DEA) to issue national e-prescribing standards for controlled substances, resulting in over one billion electronic prescriptions in 2014 and allowing e-prescribing of controlled substances in 49 States.
Clinical Decision Support
Clinical decision support (CDS) brings together patient-specific information with different treatment options in a way that allows clinicians and patients to choose the most appropriate care. AHRQ-funded demonstration projects created processes and tools for translating clinical knowledge and narrative guidelines into formats that can be used by multiple EHR systems for implementing CDS across a range of health care settings. One team implemented evidence-based guideline recommendations that address coronary artery disease, diabetes, and hypertension management; while another focused on prevention of pediatric obesity and chronic management of asthma for children. The tools, standards, and technology created by these demonstrations are being used by ONC and have become the basis for health IT standards established by ONC to make it easier to share CDS between systems.
Learning Health Systems
AHRQ’s previous work in this area helped individual patients and their doctors find answers to questions by pioneering innovative ways to gather and analyze multiple sources of health care data for evidence about what works. The Inflammatory Bowel Disease Registry (ImproveCareNow), as part of the AHRQ-supported Enhanced Registries grant program, improved health outcomes in children with inflammatory bowel disease. The remission rate of children in the registry improved from 55 percent in 2007 to 79 percent in 2015, enabling thousands of children to attend school more regularly, actively engage in sports, and overall improve their health and well-being. Use of these data also enabled 49 percent of these children to sustain remission for at least 1 year and 94 percent have achieved satisfactory growth. Another example, the AHRQ-supported Surgical Care and Outcomes Assessment Program (SCOAP)/CERTAIN project produced a safety and quality checklist for the operating room, which is being implemented in all hospitals in Washington State. The sharing of surgical checklists, and clinical data about procedures and outcomes, among Washington hospitals involved in SCOAP has resulted in reduced complication rates, shorter stays, and fewer 30-day rehospitalizations.
Distributed Research in Primary Care
Distributed research allows information to be shared across organizations without sharing data, which improves our understanding of barriers to care delivery and ability to test solutions that improve quality of care. AHRQ pioneered a distributed research network in primary care through the DARTNet project funded in 2006. DARTNet demonstrated the feasibility of using routine clinical data collected in EHRs in diverse primary care sites for comparative effectiveness research in diabetes and depression. Although the United States Preventive Services Task Force (USPSTF) has recommended screening for depression in primary care for many years, few practices have implemented this recommendation. DARTNet customized screening for depression to workflow of each practice site. The customization enabled a majority of DARTNet practices to continue with depression screening even after the end of AHRQ support for that project. AHRQ’s initial investment to create the DARTNet network met a critical need and led to its rapid growth independent of AHRQ. DARTNet has grown into a non-profit institute that now supports 12 research networks that offer access to approximately 12.5 million patient visits per year, five million patient lives, and five billion data points. These networks blend quality improvement, effectiveness, and translational research with a data driven-learning system. It is important to note that several networks focus on under-served patients including SAFTINet, Appalachian Research Network, and LA Net.
Creating Learning Communities
The Electronic Data Methods (EDM) Forum supported a learning network of researchers, clinicians, informatics experts and other stakeholders. It developed new open-access resources (e.g., eGEMs journal, issue briefs, Webinars, toolkits, Web sites) to share innovations and lessons learned, and to advance methods by supporting small-scale, rapid turnaround collaborative methods projects. The EDM Forum has also created several resources to improve governance of data and information for research and quality improvement, including a governance toolkit. Among the achievements: eGEMs, less than 3 years old, is now in PubMed Central and has over 58,000 downloads and over 100 papers; the patient-centered portable consent project led to an e-consent module, which is present in all five applications of Apple’s ResearchKit.
Current AHRQ work builds on this extensive body of evidence to support projects that develop, evaluate, and scale health IT interventions that facilitate practice improvement and better patient outcomes.