Advances in biomedicine are often slow to be discovered and even slower to translate into practice.1 Addressing this is a top priority for the United States, as evidenced by substantial investments and initiatives focused on accelerating clinical and translational research as well as improving healthcare effectiveness, quality, and efficiency.2–5 In addition, investments and incentives are ongoing to accelerate the adoption and meaningful use of interoperable electronic health records (EHRs), with an eye toward creating what the Institute of Medicine has termed the “learning health system.”6,7
Increasingly, the success of such research and improvement initiatives relies, either explicitly or implicitly, on the ability to leverage point-of-care activities and systems to generate knowledge through routine practice. Indeed, the research and improvement activities that are a focus of ongoing investments often benefit greatly from, and sometimes necessitate, clinician involvement to fully leverage clinical information and system resources for research purposes. Examples of such clinician participation in research-related activities include systematically collecting relevant data during clinical practice, generating research questions informed by practice, and recruiting patients for clinical studies. Beyond facilitating traditional research activities, there are also emerging approaches such as “pragmatic clinical trials” and “practice-based evidence for clinical quality improvement” that have the potential to more rapidly and efficiently advance biomedical and healthcare knowledge that is highly relevant to everyday practice.8–10 Overall, the potential benefits of leveraging activities and information derived from healthcare delivery to drive the generation of knowledge holds great promise.
Despite the potential benefits and imperatives to systematically learn from routine clinical activities to advance the healthcare evidence base, the ability to leverage such activities, data, and resources to advance knowledge is often challenged by a host of well-characterized but persistent factors ranging from regulatory and policy conflicts to socio-organizational and informatics challenges.1,11–13 Examples of such regulatory and policy conflicts, socio-organizational barriers, and informatics challenges include: (1) poor engagement by physicians in research participant recruitment and systematic data capture at the point-of-care; (2) inadequate attention paid by health information technology (IT) vendors to the development of capabilities in EHR platforms that support research-related information needs; (3) resistant to data reuse for research purposes based on privacy and regulatory concerns, ethical or human subjects protection frameworks, and/or impacts on operational efficiencies and financial performance; and (4) the relative paucity of resources and expertise dedicated to advancing systematic learning and knowledge generation from routine practice environments.
Although many of these factors have certainly been described in the peer-reviewed literature, and although there are certainly examples in the literature where groups have overcome these challenges to demonstrate benefit,14–24 such endeavors too often require substantial efforts and resources, are slow, and are difficult to scale or disseminate.11,25–27 Indeed, these widespread challenges often frustrate the efforts of individuals charged with accelerating the generation of evidence at the point-of-care, including informaticians, health services researchers, and others methodologists. Such frustrations increasingly take on the appearance of “headwinds” to progress, increasing costs, and impeding progress, and they must be removed if the widespread successes envisioned by ongoing investments are to be realized.
SOURCE OF THE HEADWINDS: THE CURRENT RESEARCH-PRACTICE PARADIGM
Although there are certainly many possible reasons for the current state of slow progress to realize systematic learning through routine practice, we suggest that many of the issues faced by those working to accelerate point-of-care research activities stem from a fundamental conflict in how our society regards such activities. On one hand, we are clearly investing heavily in and counting on the benefits that will accrue from leveraging healthcare activities and data for systematic learning. However, on the other hand, we have developed policies, regulations, payment models, organizational structures, and even implemented health information systems in ways that are often run counter to the goals of point-of-care research and the “secondary use” or “reuse” of clinical data for research and improvement activities.28
Instead, our current healthcare system is designed and optimized to enable the care of one patient at a time rather than to support, or even allow in some cases, the conduct of knowledge-generating activities during practice. Indeed, fundamental to the design and operation of our current healthcare system is the prevailing healthcare paradigm that views research and practice as distinct endeavors that share, at best, a unidirectional relationship. This prevailing view of a unidirectional relationship between research and practice indicates that research findings should be applied to clinical practice, ideally using an evidence-based medicine (EBM) approach.29 Although this view is very well suited to individual patient care, such a unidirectional paradigm is increasingly at odds with and often impedes the very research and healthcare improvement initiatives in which our institutions and the public (through our governmental entities) are increasingly investing. Guided by this paradigm, relevant organizational, financial, and policy decisions are often poorly harmonized, impeding the integration of research and practice that is needed to achieve our collective goals. This paradigm also fosters a counter-productive perception about research among individuals, including many practitioners and patients, like the commonly held view that participating in or contributing to research during regular clinical practice is optional or even inappropriate.30
DRIVING BIDIRECTIONAL EVIDENCE FLOW: EVIDENCE GENERATING MEDICINE (EGM)
The current challenges facing our research enterprise, combined with a growing demand for more personalized, effective, efficient, and cost-effective healthcare delivery, contribute to our emergent understanding that generating evidence in the course of practice has become central to enabling biomedical advances and a “learning healthcare system.”6,31,32 However, to achieve such advances, we must first recognize that the prevailing research-practice paradigm is flawed. Just as pressures built-up 20 years ago to drive a paradigm shift from expertise-based clinical decision-making to EBM,29 the challenges facing our healthcare and biomedical research enterprises have also revealed a need for a research-practice paradigm shift.33 Rather than characterizing the research-practice relationship as unidirectional (Fig. 1A), the new paradigm must make explicit that a bidirectional relationship between practice and research is essential to achieving a virtuous cycle of evidence generation, application, and refinement. By enabling such an evidence cycle, our health system can evolve into one that enables excellent individual patient care while also ensuring that we learn from each patient encounter to improve the health of future patients and populations.
Therefore, we propose the concept of EGM as the missing and transformational link in the current EBM model that is needed to realize a true evidence cycle (Fig. 1B). With this new paradigm in mind, we define EGM as:
The systematic consideration and incorporation of research and improvement activities into the organization and practice of healthcare to accelerate biomedical discovery and improve the health of individuals and populations.
ENABLING FACTORS TO REALIZE EGM
Given the above-mentioned definition, EGM can be understood to encompass a range of activities across the healthcare research and delivery landscape. Although some examples of EGM-type activities exist today, achieving widespread EGM activity will remain challenging without major changes to how organizations currently structure and conduct healthcare delivery and research. Enabling factors to operationalize EGM fall into 3 key categories, each spanning stakeholders from individual researchers, practitioners, and patients to institutions, government agencies, and private-sector concerns (Fig. 1B).
One category of factors that must be addressed to enable EGM includes policy and organizational factors. Pressing issues include the need to harmonize policies and regulations that govern healthcare and research. For instance, striking a balance between respecting patient privacy and the need to systematically learn from and improve upon healthcare delivery is essential. Both advancing biomedical knowledge through research and respecting and protecting individual privacy are ethical “goods,” and both must be reconciled. Some have even argued that participation in research activities should be considered a responsibility or obligation.30 However, current regulatory frameworks are often developed, implemented, and interpreted with research as an afterthought, and existing regulations relevant to research often conflict with each other. Too often, this results in the unintended consequence of erecting barriers between research and practice, restricting access to and use of clinical information and materials for both scientific and improvement endeavors.11,34 Beyond laws and regulations, accepted standards and norms spanning healthcare and research environments also require realignment to engender a scientific and learning culture wherein data sharing and knowledge advancement is both expected and seen as essential by all stakeholders.35 At all levels, engaging researchers and practitioners along with consumers in data sharing, strategic planning, and development processes can help ensure that research and systematic “learning” from the activities of healthcare are prioritized and harmonized alongside clinical care–focused requirements.
Examples of EGM-related activities to address these policy and organizational factors might include: (1) the development of institutional policies and procedures that streamline approval of novel pragmatic, point-of-care research activities; (2) the implementation of organizational governance models that prioritize research activities alongside clinical care to enable systematic learning during routine practice; and (3) the modification and harmonization of current federal and state privacy and regulatory frameworks that currently impede scientific progress by creating barriers to data sharing for research.
A second category involves fiscal and administrative factors. Most prevailing business models and organizational structures emphasize administrative and fiscal separations between research and healthcare, creating significant and persistent barriers to their integration.10 Existing models must be reconciled and reengineered to support and promote EGM alongside EBM in healthcare settings. This will likely necessitate changes to reimbursement mechanisms, the elimination of administrative silos, and the realignment of research funding models. In addition, new incentive structures are needed to encourage research activities in increasingly time-constrained and resource-constrained practice environments.
Examples of EGM-related activities to address these fiscal and administrative factors might include: (1) implementing a system to track and incentivize research-related activities at the point-of-care, such as using the recently proposed Relative Research Unit analogue to the Relative Value Unit measure of clinical productivity36; (2) investments by institutions in the research informatics infrastructure at a level comparable to the clinically-focused health IT infrastructure; and (3) the inclusion and empowerment of research and informatics leadership to make strategic decisions that drive EGM-focused activities within healthcare organizations.
A third category involves Informatics and Health IT factors. Biomedical informatics advances and Health IT investments hold great promise to facilitate evidence generation from routine practice.7,37 As noted above, ongoing efforts to implement and meaningfully use EHRs, for instance, are essential for achieving many of the goals set forth and envisioned for a “learning health system.”6 Beyond the potential to leverage discretely recorded EHR data for so-called “secondary use” or “reuse” activities, ongoing development and refinement of approaches to extract data from narrative text using natural language processing capabilities hold promise for further advances.38–40 However, even if such methods are perfected, simply digitizing data through widespread EHR adoption will not automatically accelerate research. Information collected during practice is often too incomplete, inaccurate, or not sufficiently granular for research purposes, and current Health IT systems are often inadequate to the task of supporting such multi-use scenarios.13 To fully realize their potential to facilitate EGM activities ranging from research subject recruitment to systematic data collection, Health IT platforms must be designed, managed, and used with research and improvement activities in mind. In addition, workflow changes will likely be needed to enable efficient and consistent data collection to enable “learning” while still enabling practitioner efficiency and productivity. As such, while adoption, “meaningful use,” and interoperability of EHRs to support a “learning health system” are necessary, these likely are insufficient to facilitate the activities proposed. Solutions will likely require: collaborations between informatics professionals knowledgeable in research and their healthcare IT professional counterparts; consideration of research needs by EHR vendors and those advancing informatics standards; and efforts to enable EGM functionality and workflows by those implementing systems locally.
Examples of EGM-related activities to address these informatics and health IT factors might include: (1) enhancement of Health IT platforms to support rapid and efficient deployment of standardized, systematic data capture and extraction (eg, registry) tools and interfaces; (2) leveraging EHR capabilities to identify and facilitate the recruitment of research participants at the point-of-care14; and (3) the widespread adoption of pragmatic reference information models, terminologies, and semantics for data sharing across and between Health IT and research computing platforms.
In addition to the factors already mentioned, education and training related to the practice of EGM will also be important to operationalizing it. As was proposed, when EBM was first proposed, teaching about and modeling EGM during practice must also become part of the education and training of healthcare professionals and researchers if it is to become routine across healthcare and research cultures. Beyond teaching and modeling to clinical trainees, there is also a need to develop and expand the workforce of methodological experts who will be required to instrument and operationalize the healthcare system for EGM and EBM. These will certainly include, but not be limited to, expansion of the workforces in health services research, biostatistics, and informatics, with particular emphasis for the latter on principles and methods in the emerging discipline of clinical and translational research informatics.41
In conclusion, we believe that the challenges and limitations to progress facing the healthcare and research enterprise are related to a currently flawed research-practice paradigm and that a paradigm shift is needed to realize the benefits of the envisioned learning health system.6,31 We suggest EGM as a critical and missing element to realizing an evidence cycle necessary to reap the full benefits of ongoing research and healthcare transformation initiatives. Guided by an EGM paradigm, enabling and fostering research activities during practice becomes not just acceptable, but essential, and every patient encounter becomes an opportunity to learn. By facilitating broader participation in and integration of research throughout the healthcare enterprise, EGM should lead to accelerated and generalizable findings, making it more likely that evidence will exist to improve the care of future patients. Furthermore, by better enabling research from routine practice, this approach will help ensure the answer to the critical EBM question, “is this evidence applicable to my patient,” is “yes.” Finally, we suggest that healthcare organizations, particularly those with a research and scholarly missions, are well positioned to lead in developing and operationalizing key elements of EGM, thereby modeling the integration of research and practice to accelerate biomedical science and improve human health.
The authors thank the following colleagues for their helpful feedback and comments: Dr Erin Holve, Dr Joel Tsevat, Dr Christopher Lindsell, and Dr Jonathan Silverstein.
1. Sung NS, Crowley WF Jr., Genel M, et al..Central challenges facing the national clinical research enterprise.JAMA.2003;289:1278–1287.
2. Zerhouni EA, Alving B.Clinical and translational science awards: a framework for a national research agenda.Transl Res.2006;148:4–5.
3. Steinbrook R.The NIH stimulus—the recovery act and biomedical research.N Engl J Med.2009;360:1479–1481.
4. Kaiser J.Research funding. Health bill backs evidence-based medicine, new drug studies.Science.2010;327:1562.
5. Lauer MS, Collins FS.Using science to improve the nation’s health system: NIH’s commitment to comparative effectiveness research.JAMA.2010;303:2182–2183.
6. Friedman CP, Wong AK, Blumenthal D.Achieving a nationwide learning health system.Sci Transl Med.2010;2:57cm29.
7. Blumenthal D.Launching HITECH.N Engl J Med.2010;362:382–385.
8. Horn SD, Gassaway J, Pentz L, et al..Practice-based evidence for clinical practice improvement: an alternative study design for evidence-based medicine.Stud Health Technol Inform.2010;151:446–460.
9. Saag KG, Mohr PE, Esmail L, et al..Improving the efficiency and effectiveness of pragmatic clinical trials in older adults in the United States.Contemp Clin Trials.2012;33:1211–1216.
10. Westfall JM, Mold J, Fagnan L.Practice-based research—“Blue Highways” on the NIH roadmap.JAMA.2007;297:403–406.
11. Ness RB.Influence of the HIPAA Privacy Rule on health research.JAMA.2007;298:2164–2170.
12. D’Avolio LW.Electronic medical records at a crossroads: impetus for change or missed opportunity?JAMA.2009;302:1109–1111.
13. Weiner MG, Embi PJ.Toward reuse of clinical data for research and quality improvement: the end of the beginning?Ann Intern Med.2009;151:359–360.
14. Embi PJ, Jain A, Clark J, et al..Effect of a clinical trial alert system on physician participation in trial recruitment.Arch Intern Med.2005;165:2272–2277.
15. Atreja A, Achkar J-P, Jain AK, et al..Using technology to promote gastrointestinal outcomes research: a case for electronic health records.Am J Gastroenterol.2008;103:2171–2178.
16. Roden DM, Pulley JM, Basford MA, et al..Development of a large-scale de-identified DNA biobank to enable personalized medicine.Clin Pharmacol Ther.2008;84:362–369.
17. Denny JC, Ritchie MD, Basford MA, et al..PheWAS: demonstrating the feasibility of a phenome-wide scan to discover gene-disease associations.Bioinformatics.2010;26:1205–1210.
18. Murphy SN, Weber G, Mendis M, et al..Serving the enterprise and beyond with informatics for integrating biology and the bedside (i2b2).J Am Med Inform Assoc.2010;17:124–130.
19. Sagreiya H, Altman RB.The utility of general purpose versus specialty clinical databases for research: warfarin dose estimation from extracted clinical variables.J Biomed Inform.2010;43:747–751.
20. Kozak MS, Mugavero MJ, Ye J, et al..Patient reported outcomes in routine care: advancing data capture for HIV cohort research.CLin Infect Dis.2012;54:141–147.
21. Maro JC, Platt R, Holmes JH, et al..Design of a national distributed health data network.Ann Intern Med.2009;151:341–344.
22. Pace WD, Cifuentes M, Valuck RJ, et al..An electronic practice-based network for observational comparative effectiveness research.Ann Intern Med.2009;151:338–340.
23. Kho AN, Pacheco JA, Peissig PL, et al..Electronic medical records for genetic research: results of the eMERGE consortium.Sci Transl Med.2011;3:79re71.
24. Pathak J, Wang J, Kashyap S, et al..Mapping clinical phenotype data elements to standardized metadata repositories and controlled terminologies: the eMERGE Network experience.J Am Med Inform Assoc.2011;18:376–386.
25. Yawn BP, Graham DG, Bertram SL, et al..Practice-based research network studies and institutional review boards: two new issues.J Am Board Fam Med.2009;22:453–460.
26. Embi PJ, Leonard AC.Evaluating alert fatigue over time to EHR-based clinical trial alerts: findings from a randomized controlled study.J Am Med Inform Assoc.2012;19:e145–e148.
27. Murphy SN, Dubey A, Embi PJ, et al..Current state of information technologies for the clinical research enterprise across academic medical centers.Clin Transl Sc.2012;5:281–284.
28. Safran C, Bloomrosen M, Hammond WE, et al..Toward a national framework for the secondary use of health data: an American Medical Informatics Association White Paper.J Am Med Inform Assoc.2007;14:1–9.
29. .Evidence-based medicine. A new approach to teaching the practice of medicine.JAMA.1992;268:2420–2425.
30. Schaefer GO, Emanuel EJ, Wertheimer A.The obligation to participate in biomedical research.JAMA.2009;302:67–72.
31. Olsen L, Aisner D, Michael McGinnis J.The Learning Healthcare System: Workshop Summary.2007.Washington, DC:The National Academies Press.
32. Slutsky JR.Moving closer to a rapid-learning health care system.Health Aff (Millwood).2007;26:w122–124.
33. Kuhn TS.The Structure of Scientific Revolutions.1962.Chicago:University of Chicago Press.
34. Nass SJ, Levit LA, Gostin LO.Beyond the HIPAA Privacy Rule: Enhancing Privacy, Improving Health Through Research.2009.Washington, DC:National Academy Press.
35. Piwowar HA, Becich MJ, Bilofsky H, et al..Towards a data sharing culture: recommendations for leadership from academic health centers.PLoS Med.2008;5:e183.
36. Embi PJ, Tsevat J.Commentary: the relative research unit: providing incentives for clinician participation in research activities.Acad Med.2012;87:11–14.
37. Embi PJ, Kaufman SE, Payne PR.Biomedical informatics and outcomes research: enabling knowledge-driven health care.Circulation.2009;120:2393–2399.
38. Denny JC, Choma NN, Peterson JF, et al..Natural language processing improves identification of colorectal cancer testing in the electronic medical record.Med Decis Making.2012;32:188–197.
39. Elkin PL, Froehling DA, Wahner-Roedler DL, et al..Comparison of natural language processing biosurveillance methods for identifying influenza from encounter notes.Ann Intern Med.2012;156:11–18.
40. Savova GK, Olson JE, Murphy SP, et al..Automated discovery of drug treatment patterns for endocrine therapy of breast cancer within an electronic medical record.J Am Med Inform Assoc.2012;19:e83–e89.
41. Embi PJ, Payne PR.Clinical research informatics: challenges, opportunities and definition for an emerging domain.J Am Med Inform Assoc.2009;16:316–327.