Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 25
Computational Technology for Effective Health Care: Immediate Steps and Strategic Directions 3 Crossing the Health Care IT Chasm The committee observed a number of success stories in implementation of health care information technology (IT). For example, one organization had implemented a pharmacy/medication administration system in what appeared to be an exemplary fashion. Making extensive use of robotics and bar coding of medication, patients, and providers, this organization had implemented procedures and practices that apparently reduced error rates in dispensing and administration significantly. Another organization had almost completely transitioned to electronic clinical ordering and documentation in both its inpatient and outpatient facilities. Another had made progress in using evidence-based medicine through clinician-customizable order sets to decrease the variability of care. Another had implemented effective data support for management of clinical process improvement and was able to support systematic decisions about where to focus organizational energy and attention. Although seeing these successes was encouraging, in the committee’s judgment they fall far short, even in the aggregate, of what is needed to support the Institute of Medicine’s (IOM’s) vision of quality health care. Apart from a few exceptional examples, the IT-related activities of health professionals observed by the committee in these organizations were not well integrated into clinical practice [C1O1, C1O2, C1O3, C4O17, C5O22, C6O24]. Health care IT was rarely used to provide clinicians with evidence-based decision support and feedback [C1O4]; to support data-driven process improvement [C2O6]; or to link clinical care and research [C2O10]. The committee saw virtually no effective computer-based sup-
OCR for page 26
Computational Technology for Effective Health Care: Immediate Steps and Strategic Directions port of an integrative view of patient data [C1O1]. Care providers had to flip among many screens and often among many systems to access data; in some cases, care providers found it easier to manage patient information printed or written on paper. A reviewer of this report in draft form noted the non-intuitive behavior of most health care IT systems and the training requirements that result from that behavior. Hospitals often require 3- or 4-hour training sessions for physicians before they can get the user names and passwords for access to new clinical systems. Yet much of the computing software that these people use in other settings (e.g., office software) adopts a consistent interface metaphor across applications and adheres to prevailing design/interface norms. As a result, there is much less need for training, and the user manual need only be consulted when special questions arise. In contrast, health care IT lacks these characteristics of conventional software packages—a fact that reflects the failure of these systems to address some basic human interface considerations. The committee also saw little cognitive support for data interpretation, planning, or collaboration. For example, even in situations where different members of the care team were physically gathered at the entrance to a patient’s room and looking at different aspects of a patient’s case on their individual computers, collaborative interactions took place via verbal discussion, not directly supported in any way by the computer systems, and the discussions were not captured back into the system or record (i.e., the valuable high-level abstractions and integration were neither supported nor retained for future use). Instead, committee members repeatedly observed health care IT focused on individual transactions (e.g., medication X is given to the patient at 9:42 p.m., laboratory result Y is returned to the physician, and so on) and virtually no attention being paid to helping the clinician understand how the voluminous data collected could relate to the overall health care status of any individual patient. Care providers spent a great deal of time in electronically documenting what they did for patients [C1O3], but these providers often said that they were entering the information to comply with regulations or to defend against lawsuits, rather than because they expected someone to use it to improve clinical care. These shortfalls are not necessarily for lack of investment; although health care organizations as a whole spend a relatively smaller percentage of their revenues on IT than do other fields such as banking,1 one organization—a major integrated health care enterprise with yearly rev- 1 David W. Bates, “The Quality Case for Information Technology in Healthcare,” BMC Medical Informatics and Decision Making 2:7, 2002, available at http://www.biomedcentral.com/1472-6947/2/7.
OCR for page 27
Computational Technology for Effective Health Care: Immediate Steps and Strategic Directions enue in the billions—that the committee visited had invested over a half-billion dollars in IT in the past decade. The health care organizations visited demonstrated both deep and sustained organizational and financial commitment to using information technology to improve health care. Yet their health care IT implementation time lines are measured in decades, and it is common to see the implementation of a new generation of health care IT begin while rollout of the prior generation is still underway [C4O16]. Centralization of management and reduction in the number of systems are the predominant method for standardization [C4O15], whereas innovation requires systems that can adapt to local needs [C6O25]. System response times are often slow and long downtimes are common [C4O18]. Consistent with many other reports,2 the committee recognizes commitment to 21st century use of IT in health care as an essential part of achieving the IOM’s vision of 21st century health care. But health care IT is merely a means to the desired end, namely better and/or less expensive health care. The committee believes that clinicians and other providers will, appropriately, be drawn to IT only if, where, and when it can be shown to enable them to do their jobs more effectively. Blanket promotion of IT adoption where benefits are not clear or are oversold will only waste resources and sour clinicians on the true potential of health care IT. In short, the nation faces a health care IT chasm that is analogous to the quality chasm highlighted by the IOM over the past decade. In the quality domain, various improvement efforts have failed to improve health care outcomes, and have sometimes even done more harm than good.3 Similarly, based on an examination of the multiple sources of evidence described above and viewing them from the committee’s perspective, the committee believes that the nation faces the same risk with health care IT—that current efforts aimed at the nationwide deployment of health care IT will not be sufficient to achieve the vision of 21st century health care, and may even set back the cause if these efforts continue wholly without change from their present course. Success in this regard 2 See, for example, Institute of Medicine, Crossing the Quality Chasm: A New Health System for the 21st Century, National Academy Press, Washington, D.C., 2001; President’s Information Technology Advisory Committee, Revolutionizing Health Care Through Information Technology, National Coordination Office for Networking and Information Technology, Washington, D.C., 2004, available at http://www.nitrd.gov/pitac/reports/20040721_hit_report.pdf; Office of the National Coordinator for Health Information Technology, The ONC-Coordinated Federal Health Information Technology Strategic Plan: 2008-2012, U.S. Department of Health and Human Services, Washington, D.C., 2008, available at http://www.hhs.gov/healthit/resources/HITStrategicPlan.pdf. 3 See, for example, the studies of the Dartmouth Atlas Project at http://www.dartmouthatlas.org/.
OCR for page 28
Computational Technology for Effective Health Care: Immediate Steps and Strategic Directions Box 3.1 Four Domains of Information Technology in Health Care Motivated by a presentation from Intermountain Healthcare’s Marc Probst, the committee found it useful to categorize health care information technology (IT) into four domains: Automation. Automation is the use of IT to perform tasks that can be repeated with little modification—examples include bar code medication administration, generation of laboratory results, and issuing invoices for payment. Connectivity. Connectivity begins with physical infrastructure—ensuring base-level electronic connections between various physical facilities so that data can be transmitted electronically. Examples might include high-speed fiber lines and routing capabilities throughout a physical plant, wide-area networks, and the deployment of wireless infrastructure. Connectivity includes interfaces that map data from one system into another. At the highest level, connectivity involves connecting people to systems and to each other. Decision support. Decision support (DS) involves the use of IT-based applications to provide information at a high conceptual level to clinicians to facilitate or improve decisions made about care. For example, DS can include simple rule-based alerts such as reminders to physicians about possible drug interactions when medication orders are entered. DS can also involve the presentation of information to care providers in ways that make it easier for them to know how to direct their attention—a “dashboard” indicating patient status across an entire ward or for a physician’s 50 sickest patients would be an example of DS for presentation. Finally, DS can also refer to statistical and heuristic decision support reflecting an intelligent synthesis of information about the patient, information from the care setting, and biomedical knowledge—for example, a DS system might recommend a particular antibiotic based on the patient’s condition and a database of the recent sensitivity of microorganisms to different antibiotics in their hospital. Data-mining capabilities. Data-mining capabilities use knowledge discovery techniques to analyze various similar or dissimilar datasets to recognize known or unknown relationships. Data mining converts raw data signals into clinical variables and models to provide a rich source for new approaches to evidence-based medicine and personalized care. Examples range from identification of a marker for breast cancer therapeutic response from microarray data, through mining the text literature for little-known drug-drug interactions, to mining multimedia electronic health records to identify a patient’s condition from a text note or a change in heart size from a sequence of images, and extracting ideas or relationships from a recent publication in a leading journal and pushing the information to the physicians who are treating patients who may benefit from those findings. Data mining provides many of the inputs needed for decision support.
OCR for page 29
Computational Technology for Effective Health Care: Immediate Steps and Strategic Directions will require greater emphasis on the goal of improving health care by providing cognitive support for health care providers and even for patients and family caregivers on the part of computer science and health/biomedical informatics researchers. Vendors, health care organizations, and government, too, will also have to pay greater attention to cognitive support. This point is the central conclusion articulated in this report. So that the nation can cross the health care IT chasm, the committee advocates re-balancing the portfolio of investments in health care IT; adhering to proven principles for success; and accelerating research in computer science, social sciences, and health/biomedical informatics (and concomitant education about each field for practitioners in the others). Motivated by a presentation from Intermountain Healthcare’s Marc Probst, the committee found it useful to categorize health care IT into four domains: automation, connectivity, decision support, and data-mining capabilities. See Box 3.1. The majority of today’s health care IT is designed to support automation, with some investment in supporting connectivity, and little in support of data mining or decision support. Yet the IOM’s vision for 21st century health care expects health care IT that is capable of supporting cognitive activities and a learning health care system. These activities are much more about connectivity, decision support, and data mining than they are about automation. The health care IT investment portfolio must be re-balanced to address this mismatch.