|
PERSPECTIVEElectronic Medical Records And Health Care Transformation
Effective electronic medical care record (EMR) systems will make a critical contribution to health care transformation. However, we need to know more about the total costs of EMRs and the ways in which they will interact with existing health care systems to make compelling predictions about their clinical benefits or the savings they can enable.
Electronic medical records will make a critical contribution to health care transformation. At Geisinger we see it daily. Our 650 physicians create 100 percent of outpatient orders and 70 percent of outpatient office notes and operative notes electronically. Within six months after going live on the EMR, even our initially most skeptical physicians flatly refuse to practice without it. We distribute 120,000 same-day consultation reports (automatically distributed electronically or by U.S. mail, according to the recipient’s preference) annually. We provide 27,000 patients with access to their electronic record and Health Insurance Portability and Accountability Act (HIPAA)–compliant electronic messaging with their doctor’s practice (and let them schedule some appointments themselves). We are conducting six funded studies of the quality and efficiency effects of EMR-enhanced care processes in real-world practices, with five more under development. We can identify many benefits that we have realized—clinical and financial.
EMRs are complex systems, used by time-pressured, frequently interrupted users, many of whose workflows are (appropriately) too idiosyncratic for effective automation.1 Therefore, organizations implementing EMRs must redesign complex work processes and automate them simultaneously (in contradistinction to the telecommunications industry success story discussed below). Such complexity guarantees unexpected effects, some of which will cause errors. Richard Hillestad and colleagues’ assertion that "a carefully redesigned medication process supported by a modern CPOE [computerized physician order entry]" will not increase the overall error rate is likely correct.2 However, what reason do we have to think that the care processes in thousands of hospitals and practices will be skillfully redesigned and supported by an EMR that has been carefully configured and tested to support those processes?
An EMR implementation that is capable of supporting less error-prone care processes will require substantial resources for workflow analysis, software configuration, testing, and user training. These activities will be critical both before and for years after implementation. An implementation that provides adequate resources for these activities will cost far more than one that is just adequate to meet evolving EMR accreditation requirements. Since most organizations will have difficulty funding even a minimum implementation, few EMRs are likely to be robust enough to support meaningful work-process transformation. Even for organizations that have made sizable resource commitments at one point in time, it remains a challenge to continue funding workflow analysis, configuration, testing, and training. Recent reports of the errors caused by EMRs probably reflect this reality.3 This is not just a matter of money. Few organizations will be able to find, attract, train, and retain an information technology (IT) team with the critical skills (technical, clinical, process redesign, project management, and informatics) needed to implement an effective EMR. The lead time needed to train such people, and the number of people needed, can be inferred from the American Medical Informatics Association’s (AMIA’s) recently announced initiative to train 10,000 such workers by 2010.4 Although this is a large number, it amounts to only two workers for each U.S. hospital, with none left over for physician practices. Finally, and most importantly, the implementation of an effective EMR system requires an organization to be passionately committed to transforming the ways it cares for patients—and capable of effecting that transformation. Scarcities of both the passion and the capability are probably reflected in the low rate of EMR adoption and the high rate of EMR project failures.
Thomas Landauer has summarized a large body of econometric data that show no correlation between IT investment and productivity growth in the sixteen years from 1973 to 1989 in industries as diverse as insurance, banking, air transport, wholesale trade, retail trade, and manufacturing. Of the one exception—the telephone industry—Landauer notes, "The telephone companies had huge numbers of employees doing a number of well-subdivided, highly routinized tasks."5 As Hillestad and colleagues note, it was not until the late 1990s that widespread productivity gains could be attributed to IT expenditures. The complexity of health care makes it unlikely that we will achieve large productivity gains more quickly than other industries have. Given these caveats, how have Hillestad and colleagues advanced our understanding? They have made deliberate use of a well-designed methodology to estimate potential benefits and costs. However, the base case from which they extrapolate is problematic. The few organizations from whose research the extrapolation might be made are academic medical centers with custom-built EMRs, which have little in common with the vast majority of hospitals and physician practices (including other academic centers). Nor have they produced a robust research literature demonstrating the benefits of EMRs. In a study that is among the most widely cited as demonstrating the safety benefits of computerized physician order entry (CPOE), preventable adverse drug events were not greatly reduced.6 Other research suggests that in addition to an effective EMR, many other system components are needed to support improved physician performance.7 These include academic detailing of physicians, patient activation, financial incentives, and audit feedback of physician performance. We (and others) are working to implement and test performance improvement initiatives that add these components to EMRs, but no one has reported as much as 60 percent combined physician and patient adherence to the hundreds of validated recommendations that will be needed to achieve the benefits that Hillestad and his colleagues project. We have enough estimates. They suggest, as persuasively as such estimates can, that well-implemented EMRs have the potential to improve health care at an acceptable cost. To enable providers and payers to make serious commitments to implementing EMRs, we need real-world demonstrations of how commercially available EMRs can support improved care processes cost-effectively. We need standards for interoperability and a Consumer Reports for EMRs. We need to know more about cost-effective design, implementation, and technical support of EMRs. We need powerful financial incentives—such as pay-for-performance—that will reward organizations for using EMRs to improve the quality and efficiency of U.S. health care. We need to start yesterday.
James Walker (jmwalker{at}geisinger.edu) is chief medical information officer at Geisinger Health System in Danville, Pennsylvania.
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Complexity of EMR systems.
Resource requirements.
