A Progress Report on Computer-Based Patient Records in the United States
Paul C. Tang and W. Ed Hammond
Much has changed since the release of the first edition of The Computer-Based Patient Record: An Essential Technology for Health Care. The current environment in which health care is practiced and the information technology available to its practitioners are significantly different from that which existed when the study was completed in 1991. Changes in the health care environment produced fundamental shifts in the delivery of health care, favoring outpatient care over impatient care, primary care over specialty care, and guidelines-driven care over autonomous decision making. Technological advances have overcome some barriers to computer-based patient records (CPRs) (e.g., World Wide Web, applications that operate across distances on many different computers) and heightened the visibility of others (e.g., confidentiality policies and legislation). In this commentary, we describe some of the environmental and technological changes that have occurred since publication of the first edition and highlight the challenges that remain to be addressed. Probably the most significant nontechnological change that occurred since 1991 was the change in the health care practice environment. We begin by addressing the new environment.
The Changing Health Care Landscape: Reinforcing the Need for CPRS
As health care costs in the United States approached $1 trillion in the early 1990s, businesses, consumers, and payers began to call for mechanisms to manage the escalating costs (IOM, 1993). This issue was so much at the forefront of U.S. business and political priorities, that the 1992 presidential election was dominated
by discussions of health care costs and debate on effective ways to control them.
Although efforts at full-scale health care reform led by the federal government gained momentum, peaked, and then disintegrated, market-driven changes in health care delivery were already under way and steadily moving forward. Managed care plans continue to gain increasing shares of the market (Hoechst Marion Rousell, Inc., 1995). The need to examine and manage the health needs of a population has dramatically increased the demand for information systems that capture clinical data. The health care information systems industry has shifted its attention from financial systems to clinical systems, particularly CPR systems. Two core tenets of managed care that impact the demand for CPR systems have been the central role of primary care and the emergence of integrated delivery systems.
The Evolving Role of Primary Care
Managed care has redefined the primary care provider as the principal provider of care and, in many instances, a gatekeeper for access to certain diagnostic tests and specialty care. To efficiently carry out these roles, the clinician needs ready access to both clinical and administrative data. At the same time, guidelines are being promoted to reduce the variances among clinical practices. When guidelines are integrated into a CPR system, they can streamline the steps necessary to adhere to them and simultaneously document that compliance. CPR systems that integrate clinical guidelines in the order-entry process have the best acceptance among providers (Sittig and Stead, 1994; Sullivan and Mitchell, 1995; CPRI, 1996c, 1997). However, despite the abundance of guidelines, few can be implemented in a computer as written (Tierney et al., 1995). Software tools are needed to help author guidelines that are internally consistent, include precise definitions of eligibility criteria, and accommodate a variety of patient conditions.
One of the major goals of managed care is to provide more of the care outside of the hospital. However, the impatient and the outpatient settings differ substantially. There are differences in the temporal nature of information, the responsibilities of each member of the health care team, the need for a communications infrastructure to facilitate coordination of care, and other logistical concerns which impact the detailed design of information systems. Consequently, vendors of information system products for hospitals find that there is a steep learning curve to understanding the information needs of physicians in the ambulatory care setting. Deliberate analysis of the information needs and work flow requirements in ambulatory care will help system developers design information systems that increase the efficiency and effectiveness of clinical practice throughout the continuum.
The Integrated Delivery System
Another fundamental concept that heightens the need for CPR systems is that of an integrated delivery system (IDS). An IDS is composed of health care providers, service providers, and facilities organized to provide a continuum of health care services to a defined population. These systems of health care were created in response to payers' desire to contract with single entities that provide comprehensive health care services for their clients. To manage the delivery of care in an IDS, a health system must have efficient and accurate ways of capturing, managing, and analyzing clinical data collected at all the different sites where care is provided.
In addition, payers and regulators are requesting ''report cards" on quality, outcomes, and costs of care provided by the integrated delivery system. For example, the National Committee for Quality Assurance (NCQA) developed the Health Plan Employer Data and Information Set (HEDIS) as a standard report card to help employers evaluate different health plans. Initially focused more on administrative data, the evolving HEDIS criteria are increasingly targeting clinical processes and outcomes. Gathering the data to prepare these reports can be immensely time-consuming and costly when they are manually abstracted from paper records, but with a CPR, reporting on aggregate data can be a byproduct of capturing data electronically.
NCQA advised health plans to "move to fully implement the information framework, including the automated patient record" in order to meet the clinical reporting requirements of forthcoming regulations (NCQA, 1997). As outcomes reporting requirements become more sophisticated and deal with complex, multifaceted diseases, it will be essential to have electronic access to the record and tools to efficiently analyze practice patterns and patient outcomes. NCQA will develop HEDIS measures that assume health plans and provider organizations use CPRs by 2002.
In short, since the release of the first edition, the demand for clinical data has become a business imperative. The organizational complexities of these large, diverse, and geographically dispersed health systems add new challenges—and new opportunities—for developers and implementers of CPR systems.
Computer processing power doubles in performance and halves in cost about every 2 years. These past 6 years have been no exception. Although it is not our intent to review all the developments in information technology (because technology has not been the critical impediment to adoption of CPR systems), three striking technological shifts have occurred that favorably impact the foundation for CPR systems. We briefly describe these developments below.
The Internet and the World Wide Web: Widespread Connectivity
Perhaps the single biggest technological change since the first edition of the report is the rising importance of the Internet. Although the Internet dates back to 1969 when the first node of ARPANET was installed at the University of California at Los Angeles, several changes have coalesced to bring together a paradigm shift that now touches all segments of society. The World Wide Web (WWW) has transformed a research network into the fabric of a new information age. Internet service providers have made access convenient and reasonably inexpensive. As much as 50 percent of the U.S. population will have access to the Internet by early in the next century (National Research Council, 1996). It is precisely this kind of ubiquitous connectivity that enables IDS organizations to share data throughout their geographically dispersed clinical delivery sites, and even to reach the consumer or patient at home. To the extent that health care becomes dependent on access to computer networks, however, policymakers need to pay special attention to the needs of the medically underserved population to ensure that lack of network access does not further impede their access to care.
Connectivity is not the only requirement for transmitting patient data to remote sites. Confidentiality and security safeguards need to be developed and enforced. Fortunately, the requirements of business to protect electronic commerce over the Internet will drive technical solutions and policy standards, which health care applications can leverage. Technology dedicated solely to health care applications risk being orphaned due to the lack of a mass market. The market drive of consumerism must pave the information infrastructure for health care applications
World Wide Web Browsers: A Universal View on the Internet
Probably the most important tool that led to the domestication of the Internet was the development of software that made it easy to connect to, search, browse, and download information from anywhere on the network as if it were located on the user's personal computer. Commonly called browser software (e.g., Netscape Navigator™, Microsoft Internet Explorer™), these programs give a graphical, intuitive, and common interface to functions that locate and interact with remote data on the Internet without the user having any technical knowledge of how it is done. Browser user interfaces have become so commonplace that they are being adopted as the interface to desktop computers.
Another fundamental breakthrough associated with WWW browsers is that the software runs on almost any computer (Cimino, 1995). One of the critical problems that had been plaguing computer users since the invention of electronic computing has been the general inability of programs written for one machine to run on another machine or to use data generated by another program. Efficient and cost-effective use of computers suffered due to the incompatibilities caused
by a lack of standards. The WWW defined standards for document formats (hypertext markup language, HTML) and transfer protocols (hypertext transfer protocol, HTTP). Compliance with these standards permitted independent developers to write programs for heterogeneous computers and operating systems, yet have them all seamlessly access information on the WWW.
The success of the WWW demonstrates the market-expanding potential of adopted standards. It also demonstrates the remarkable leverage provided through natural entrepreneurial forces complying with industry standards. The benefits to the consumer, including clinicians, are tremendous.
A variant, but powerful spin-off, of the Internet is the Intranet, a controlled-access version of the Internet. Like the Internet, Intranets use industry-standard document formats, data exchange protocols, and browsers. The prolific development of new tools and products for the Internet can be directly applied to Intranets. An estimated 80 percent of the Internet products are purchased for use on an Intranet. The Intranet, however, typically has better bandwidth, security, and administration because it is controlled and operated by the private enterprise. These advantages are particularly important for health care applications, and most health care organizations are exploring the use of Intranets as a component of their information strategy.
Industry is at the cusp of another major change—network-centric computing. The predominant paradigm for computing today uses personal computers (PCs) to perform most of the computational work using software that resides on local disks. The hardware, software, and support costs for this type of operation require large budgets and staffs.
Network-centric computers are diskless computers that attach to a network and load software stored on a server computer. The capability to manage hardware and software more centrally reduces the acquisition and operating costs for the enterprise. Although this new architecture has not been used extensively in production, it has the potential to revamp distributed computing.
Having noted the significant changes in the health care environment and advances in information technology, we next discuss the current state of the CPR.
The Current State of CPRs
As discussed in Chapter 4 of the report, the development and diffusion of a new technology are separable but interrelated events. In discussing the state of CPR development and implementation in the United States today, it would be
desirable to examine two questions. First, to what extent have CPRs, as defined by the study committee, been developed? Second, to what extent have the available CPRs been adopted? Unfortunately, a comprehensive review of the industry does not exist, and it would soon be out of date if it did. Alternatively, one could rephrase the questions from the perspectives of providers and vendors. From the perspective of system purchasers, do the CPR systems that are available in the market meet the needs of health care institutions? And from the perspective of vendors, is the market ready to buy CPR systems? Although in neither case is the answer a resounding "yes," promising signs can be reported.
The CPR Market
One must be careful when describing a "CPR market" not to tether the concept to a single, static idea whose incarnation can be purchased "off-the-shelf." In some sense, it is precisely the static nature of the paper-based record that has been such a great burden to the practice of medicine. Instead, a CPR system is a constantly evolving concept whose value and function is expected to grow with the constantly changing demands of the health care environment and the improving technology upon which the system is built. The level of CPR development activity has definitely increased significantly since 1991. The number of commercial systems addressing various attributes of CPRs has increased and most major health care information technology vendors now offer CPR-related products. However, comprehensive information system products that seamlessly integrate data and coordinate processes across the entire continuum of health care services do not exist. Most health care information system vendors, whether their products were formerly based in the impatient or the outpatient side, are working to extend their products to cover the needs of integrated delivery systems. Developers generally start from the data end of the system and work toward the human side where the clinician interacts directly with the system. Hence, one way to trace the evolving functionality is to look at the transformation of data into information that clinicians use to make decisions. Although the path is not necessarily sequential, five hallmarks of this transformation are enumerated below:
- Integrated view of patient data,
- Access to knowledge resources,
- Physician order entry and clinician data entry,
- Integrated communications support, and
- Clinical decision support.
Integrated View of Patient Data
This is one of the earliest benefits of CPRs—improving access to all patient data whenever and wherever clinical decisions are made independent of where
the data was originally acquired. Observational studies of clinicians' information needs provide details on the kinds of information physicians require to make decisions regarding the care of their patients (Tang et al., 1994). Up to 81 percent of the time, physicians could not find all the available patient information desired to make patient care decisions during an outpatient encounter. Limited by a format that has not substantively changed in close to a century, the paper record is ill-suited to the information demands of modern clinical practice. Most CPR system vendors offer products that combine data from various sources and present an integrated view to clinicians.
Access to Knowledge Resources
Providers often need other information in addition to patient data and their own personal knowledge (Covell et al., 1985). The guidelines, rules, and regulations in the managed care environment have intensified this need to access medical and administrative knowledge at the time decisions are made. CPR systems sometimes provide methods for organizations to incorporate access to local knowledge resources, but generally, this knowledge access is passive. That is, the user searches for the needed information electronically but has to abstract the pertinent content and enter any relevant orders manually into the clinical system. Ideally, access to knowledge resources should be integrated with clinical decision support in ways that directly influence physicians' ordering behavior, as described below.
Physician Order Entry and Clinician Data Entry
Physician orders initiate clinical interventions. Proactively influencing physicians' orders is the most efficient way to influence patient outcomes (McDonald, 1976, 1984). Systems that physicians use routinely to enter orders, whether in the impatient setting or outpatient setting, can produce significant effects on quality and costs of care (Barnett, 1984; Tierney et al., 1987, 1990, 1993; Sittig and Stead, 1994). In addition, when clinical data are entered and maintained by the clinicians responsible for care, the accuracy and quality of data are high. Very few commercial systems, however, are used by physicians to write all their orders. Human-computer interface issues and perceived benefit substantially affect the success of this function. Some of the reasons are discussed later under remaining barriers.
Integrated Communications Support
With an increasing emphasis on outpatient care, coordinating the activities of health care professionals from multiple organizations at different sites, including the home, becomes more important. No longer confined to an acute care facility
where proximity between the health care team members and the patient was the rule, the patient and care team in the outpatient setting meet more by appointment than by clinical demand. Relying on paper-based mail is inefficient and fallible. Clinicians need integrated communications support for effective functioning of the multidisciplinary outpatient health care team (Tang et al., 1996). A communications infrastructure that is linked to the shared patient record facilitates overall coordination of care and timely response to changing patient conditions.
Clinical Decision Support
Few, if any, commercial systems provide a high level of proactive decision support. Drug interaction checking and simple abnormal laboratory-test result alerts are available, but the richness of applying a broad range of knowledge to influence physicians' orders is still under development (Johnson, 1994).
Only when both patient data and clinical knowledge reside in the system in machine-understandable format can the system provide additional support to the clinician making decisions. For example, encoded medical knowledge about the meaning and significance of changing laboratory-test results would allow a system to provide alerts, an active function, in addition to the passive data retrieval function. Similarly, if the system could match the patient context with relevant clinical guidelines, it could present ordering options consistent with the appropriate guidelines. The clinician is responsible for the definitive decision, but the system can actively provide options and explanations that improve the clinician's efficiency and compliance with accepted guidelines of practice.
In summary, since the first edition, there has been a significant increase in development efforts on CPR systems. We remain optimistic that the remaining years of the decade-long challenge will deliver on the promises of computer-supported decision making.
At the Fork
In 1991, the Committee on Improving the Patient Record reported that the most advanced CPR systems were found in several academic medical centers or teaching hospitals affiliated with universities as well as in the Department of Veterans Affairs and the Department of Defense. Not surprisingly, today the most advanced CPR systems implemented in the United States can be found in the same places. Of the seven institutions that have been recognized as part of the Computer-based Patient Record Institute Davies CPR Recognition Program (described below), four are academic medical centers and one is the Department of Veterans Affairs. One of the distinguishing factors of institutions that have been recognized as having advanced CPRs is the clear organizational leadership and commitment to CPR system implementation that helped make their efforts successful (CPRI, 1995c, 1996c, 1997).
A comic saying attributed to Yogi Berra states: "When you come to the fork in the road, take it." In contrast to the early CPR system adopters, most health care institutions are standing at the fork, trying to decide whether or not to begin implementing a CPR system. Those who have made the decision to invest in a CPR system are grappling with the complicated issue of how to do so. Many organizations implement the CPR system in some, but not all areas. Others implement a partial system and depend on a combination of paper and electronic documentation. All adopters of CPR systems must address how to integrate the components of the CPR and how to integrate the CPR with other institutional information systems. The challenge of implementing such an expansive, robust system is daunting, but the option of continuing to manage the clinical and administrative data of an IDS on paper is increasingly becoming a nonviable alternative.
Recent Activity to Advance CPRs in the United States
Against the backdrop of technical and nontechnical changes, interest in and incentives to develop CPR systems have increased.
The National Library of Medicine (NLM) has been at the forefront in stimulating research on the effective use of CPR systems and networked access to shared data (Lindberg, 1995). NLM conducts intramural research that is directly applicable to technological and infrastructural needs of CPR system development and effectively uses its extramural research grants and contracts to apply the results of academic research to health care. Through its High-Performance Computing and Communications contracts, NLM has been a leader in facilitating the use of information technology by health care professionals of rural, urban, community, and statewide networks around the country. NLM's extensive work on the Unified Medical Language System (UMLS) has been a major contribution in the medical terminology arena. NLM and the Agency for Health Care Policy Research (AHCPR) are sponsoring a large-scale vocabulary test to assess the "extent to which a combination of existing health-related classifications and vocabularies covers vocabulary needed in information systems supporting health care, public health, and health services research" (Humphreys et al., 1996). NLM is further extending the reach of shared computer-based patient data and systems through its telemedicine program.
Due to AHCPR's role in developing scientifically based clinical guidelines, it has long recognized the importance of standard data definitions and capturing clinical data in structured form. AHCPR has also played an active role in facilitating standards development. Widespread use of CPR systems would not only facilitate the collection of aggregate data in support of guidelines development, but also make effective the dissemination and use of clinical guidelines in clinical practice.
In addition to the efforts undertaken by NLM and AHCPR, other federal
agencies, including the Department of Defense, and the National Institute of Standards and Testing have internal and extramural activities in CPR systems. However, a cohesive federal policy to speed the development of a health information infrastructure and the diffusion of CPRs has not emerged in the United States (Shortliffe et al., 1996).
One focal point for accelerating the development of the infrastructure for CPR systems (e.g., confidentiality policies, standards, evaluation criteria) in the private sector has been the Computer-based Patient Records Institute (CPRI). CPRI was formed in response to the recommendation of the Institute of Medicine (IOM) study committee to promote and facilitate the development, implementation, and dissemination of the CPR. Led by many of the initial supporters of the IOM committee, CPRI was incorporated in January 1992, as an association of organizations representing the various stakeholders in health care. In its first year, 22 organizations became members of CPRI. Over the subsequent years, the organization has grown to more than 70 members. CPRI was charged by the IOM committee with the following objectives:
- Support the effective, efficient use of computer-based patient records.
- Educate change agents and stakeholders about the value of CPRs in improving patient care.
- Foster the CPR as the primary vehicle for collecting patient data.
- Promote the development and use of standards for CPR security and data content, structures, and vocabulary.
Despite efforts in the public and private sectors, however, significant barriers impeded the development and use of CPR systems in the United States. Many of the remaining critical barriers to CPR system development and routine use concern problems that are most effectively dealt with by cooperative, focused activity. We describe some of these barriers below.
Remaining Barriers or Challenges
Technology has continued to move forward at a rapid pace. By comparison, the human and organizational sides of the challenges have remained relatively stagnant. In 1991, the committee stated that informational, organizational, and behavioral barriers must be addressed to advance CPR systems, and that these barriers overshadowed the technical barriers. Below, we elaborate on some of the critical barriers to CPR development and diffusion.
Definition of the CPR
Although work on a common definition of a CPR and CPR system is under way by various groups, a universal understanding of the concepts embodied in a
CPR does not exist. Without a clear understanding, users have a difficult time selecting systems that will meet their needs and vendors have difficulty supplying such systems. CPRI described a CPR as "electronically stored information about an individual's lifetime health status and health care." It replaces the paper medical record as the primary record of care, meeting all clinical, legal, and administrative requirements. A CPR system provides reminders and alerts, linkages with knowledge sources for decision support, and data for outcomes research and improved management of health care delivery. It is worth repeating a point made earlier that a CPR system is an evolving concept that responds to the dynamic nature of the health care environment and takes advantage of technological advances.
Beyond the first set of definitions, however, few details have been worked out and agreed upon. For example, there is no common data model for the CPR, no common set of data elements, no common vocabulary, and no common set of scenarios that are supported. These requirements are fundamental if developers are to create a person-centered CPR that links care across different sites, specialties, and circumstances.
Many systems still follow the traditional organization and characteristics of the paper-based system and have simply automated that system. Narrative documentation, for example, is far more prevalent than structured text. Even though most new CPRs support a multimedia record, new data forms have not been smoothly integrated into the record, and little has been done to evaluate their true worth. Finally, the concept of incorporating patient-derived information (e.g., health status) as part of the patient record has not been implemented to any significant degree.
Meeting User Needs
In order for clinicians to rely on data in the system, they must be the direct users of the system. The emphasis on clinical data is a fundamental change from the previous era of hospital information systems, where clerks were the main users because the primary motivation was to capture charges and generate bills.
Analysis of the common questions concerning patient information that physicians ask (e.g., what evidence supports the diagnosis, has a patient ever had a specific test, and has there been any follow-up because of a particular laboratory test result?) provides insight into the difficulties clinicians have had finding the answers in the paper-based chart. Although current computer-based tools installed in health care institutions can typically help clinicians retrieve laboratory test results, they are not designed to answer many of the common questions clinicians ask about patient data. CPR system developers will need to address these needs to satisfy the new clinician users. The maker of the Swiss Army Knife describes the key to inventions as follows: "Make it useful. Very useful. Conveniently useful…." Likewise, key to gaining clinician user acceptance is providing efficient
tools that help clinicians retrieve and understand data relevant to their decision making tasks.
Clinicians also need answers to many questions regarding medical knowledge (Covell et al., 1985). Diagnostic decision support tools are available as stand-alone microcomputer programs (e.g., QMR, Illiad, DXplain) or are available over the Internet (e.g., DXplain; Miller et al., 1986; Barnett et al., 1987; Feldman and Barnett, 1991; Berner et al., 1996). However, their greatest use would occur when proven diagnostic decision support tools are integrated with CPR systems.
Among the more challenging issues confronting CPR system developers is the issue of effective user-computer interfaces. Physicians must be the users of the system, performing data entry (e.g., orders, progress notes) as well as information retrieval, if they are to realize the benefits of interactive, on-line decision support. Progress must be made in understanding the "cognitive processes involved in human-computer interactions in order to design interfaces that are more intuitive and more acceptable" (Tang and Patel, 1994). Cognitive issues are also relevant to designing the presentation of clinical information in ways that facilitate rapid assimilation and analysis. Ultimately, good solutions to the human-computer interface will require changes not only in how the system looks but also in how humans interact with the system. What information the provider needs and what tasks the provider is performing will influence what is presented as well as how it is presented. Templates that reflect providers' work flows will aid them in using the system efficiently. Defaults that represent common, desired selections, for example, not only improve efficiency but also help increase compliance with practice-defined guidelines (CPRI, 1996c, 1997).
Although user acceptance was a major barrier in the past, more and more users are demanding that organizations implement computer-based means for accessing and managing patient data. This is not to say that the cultural and organizational challenges associated with major change are not substantial barriers to overcome, but the users' awareness of the potential benefits will become an increasing catalyst for change.
The change in users and uses has a dramatic impact on the applications desired and the kinds of user interfaces required. The industry must now focus on a completely different set of users and must define and address their information needs. Proactive and informed user participation will be necessary for implementations to succeed. The attention to clinical data and clinical users described in this report is even more important today than it was 6 years ago.
Rather than existing as a monolithic system, contemporary health care information systems are made up of multiple-component systems manufactured by multiple vendors, owned by multiple entities. To share data, which is required in
an IDS, industry-adopted standards must be defined for interfaces between components (AMIA, 1994; Hammond, 1994).
Health systems must have a unique health identifier (UHI) to accurately and reliably link all the data on a single individual. Until last year, striving for a national UHI had been an elusive goal. Public Law 104-191 (The Insurance Portability and Accountability Act of 1996) calls for the secretary of Health and Human Services ''to adopt standards for unique health identifiers, confidentiality policies, and terminology." To build support for addressing these pivotal issues, in November 1996, CPRI convened a national Summit of over 80 senior executives representing health care providers, health systems, government agencies, information systems vendors, businesses, regulators, and quality assurance professionals. The organizations participating in the Summit developed consensus recommendations regarding a unique health identifier, confidentiality legislation and policies, and standard health terminology. CPRI was charged with follow-up actions to work with government agencies and industry groups to implement the recommendations.
The need for standards governing the content, vocabulary, and format of data remains a high priority. So far, most of the progress has been made in the development of messaging standards. Standards for data exchange have continued to be expanded and are increasingly being adopted by provider organizations and vendors. Standards now exist for exchanging clinical data (Health Level Seven [HL-7], 1994), images (ACR/NEMA), clinical observations (ASTM Committee E-31), bedside instrument data (IEEE, 1995), prescription data (NCPDP, 1992), and administrative data associated with claims (Accredited Standards Committee X12N). It is important to note, however, that even with standards that define message formats, implementation of messaging standards may differ among vendors. The next step is to call on industry to provide standard implementations for messaging standards such as HL-7. That would be a major step toward "plug-and-play" capability.
Although there has been progress in developing individual coding standards for data elements, none has emerged as a comprehensive standard. More than 150 different coding sets defining terms for use with the medical record have been created. Among these coding sets are ICD9-CM, ICD10, SNOMED III, CPT, NANDA, Read Classification, LOINC, and MEDRA, to name a few. Not only findings, but also results of laboratory tests must have standard terms (e.g., results of a Pap smear or mammogram). Additional coding sets are being defined by specialty groups that recognize the need for a common vocabulary. The NLM has expanded its UMLS to map many of these coding sets into a common set. Names for every element that must be exchanged among systems should be standardized. For the CPR, a clinically rich vocabulary that accurately describes patient problems and findings is mandatory. In addition, physicians must feel comfortable with the standard vocabulary if they are to use it for entering data. CPRI has published an evaluation of existing comprehensive codes (Chute et al., 1996).
In addition to standards for patient data, medical knowledge must be encoded in CPR systems with decision support capability. The Arden syntax has been used to exchange medical knowledge encoded as simple rules (Hripcsak et al., 1990). Further work must be undertaken to represent medical knowledge in standard, transferable form.
Leadership at the federal level is required to ensure that standards necessary to preserve and enhance health care in the United States are developed. Until standards exist for uniquely identifying individuals and coding and exchanging health data, the value from capturing and aggregating data will go unrealized and each organization will be its own pioneer.
Legal and Social Issues
Security, privacy, and confidentiality concerns have become major barriers to widespread implementation of CPR systems and sharing data. There is, as yet, no agreement on what must be done to establish the balance between appropriate use of health care data and the individual patient's rights to privacy (Detmer and Steen, 1996). The issue of who owns the data in a CPR is still being debated. Of equal importance to preserving patient privacy and confidentiality is the necessity of preserving institutional privacy. No institution will be willing to share data if those data can be used to provide a business advantage for a competitor. Again, the human factors outweigh the technical solutions in dealing with this issue (Barrows and Clayton, 1996).
Privacy and confidentiality are concepts that involve people, policies, and legislation. Information security technology plays an enabling and facilitating role by helping organizations prevent unauthorized access to confidential information. In addition, properly designed and monitored audit trails can enhance user accountability by detecting and recording unauthorized access to confidential information. CPRI has produced position papers on user authentication and access to patient data and provided substantive guidelines on security policies, security education programs, job descriptions for information security managers, model confidentiality policies, and security functionality requirements for CPR systems (CPRI, 1995a,b, 1996b,d). CPRI and the American Health Information Management Association have been instrumental in developing model policies and legislation regarding confidentiality and privacy. Public Law 104-191 establishes legal sanctions for wrongful disclosure of individually identifiable health information. It also calls on the secretary of Health and Human Services to provide detailed recommendations on privacy of health data and procedures and rules for authorized disclosure of such information. The recently revitalized National Committee on Vital and Health Statistics (NCVHS) advises the secretary on this and other standards related to health information. Federal legislation is necessary to overcome many of the inadequacies and inconsistencies between the state regulations and laws that are described in this report.
Whereas stringent security measures should be applied to protect the confidentiality of patient information, it is also in the patient's best interest for the CPR to be accessible for appropriate, legitimate uses by authorized users. The measures used to protect the confidentiality of patient data must not be so onerous that clinicians taking care of a patient do not have ready access to the patient's health data. This balance must be carefully crafted in the enabling legislation and policy standards, and enforceable through system security functions. In addition to clinicians, researchers, quality assessment professionals, and health care managers need access to aggregate data to continuously improve health and the delivery of health care. Most of the time, these secondary uses of data can be satisfied without access to individually identifiable information.
Costs and Benefits
Certainly, the cost of developing a CPR system remains a significant barrier, but this barrier has been offset at least in part by three trends. First, as CPR-related technologies continue to advance, greater performance can be obtained for equivalent costs. Second, as integrated delivery systems become more prevalent, the demand for CPR and related information systems will increase, thus reducing the risk to developers. Third, some federal funds have been made available (e.g., through NLM and AHCPR) to address specific development challenges associated with CPRs. The stability and level of federal funding in the future, however, are uncertain.
Much of the enthusiasm for the CPR is based on the belief that a CPR system will reduce the cost and improve the quality of care through the existence of better-informed health care providers and patients, the elimination of duplicate testing, and better coordination of treatment by more than one health care provider. Since the first edition, additional data have been published on the positive impact of CPR systems on the cost and quality of health care. As existing clinical systems continued to accumulate data, new studies have demonstrated improvements in the quality of care or reductions in the cost of care (Classen et al., 1991, 1992; Evans et al., 1992; Schoenbaum and Barnett, 1992; Tierney et al., 1993; Grandia et al., 1995). However, most of the carefully done empirical studies presented in the literature have examined the benefits of systems that are not commercially available. Like other enabling technologies (e.g., electronic mail, cellular telephones, computers), it may be difficult to identify and reliably quantify the direct and indirect benefits attributable to the CPR system. It is hard to quantify what is better, when so much may not even have been possible without the enabling technology.
Although discrete benefits of CPR system features have been quantified in selected environments, the comprehensive capital and operating costs have not been fully articulated. It is known, however, that the costs are substantial. One estimate for a capital budget (including capitalized labor costs) for an integrated
health care information system with $1 billion to $2 billion in revenue is $75 million to $275 million (Council on Competitiveness, 1996). At a time when health care organizations need to reduce their costs, allocating capital to information systems is still a challenge. Although it is tempting to propose that large empirical studies be conducted to produce definitive cost-benefit data, it may be unrealistic to expect that such studies can be undertaken, or that they could truly represent the diverse health systems in the United States.
Many health care organizations are aware of the benefits of CPR systems, but have questions on how to implement and use this new technology. Recognizing that experience with these systems is not widespread, CPRI developed a program to help disseminate the knowledge of those who have previous experience with CPR systems and to provide a tool to help organizations undertake new projects and initiatives. Modeled after the Malcolm Baldrige Quality Award Program, CPRI developed a program to promote quality in CPR system projects. The program is named after Nicholas E. Davies, a member of the IOM Patient Record committee who was tragically killed in an airplane crash just as the committee was concluding its work. The Nicholas E. Davies Annual CPR Recognition Symposia provide recognition for organizations demonstrating exemplary measurable impact of CPR systems on health care at their organizations (CPRI, 1995c, 1996c, 1997).
The Objectives of the Davies CPR Systems Recognition Program are to: (a) promote the vision of CPR systems through concrete examples, (b) provide visibility and recognition for excellence in CPR system implementation and demonstrated results, (c) provide a forum for discussing critical success factors and lessons learned, and (d) provide criteria for evaluating CPR system projects (CPRI, 1996a). A comprehensive set of evaluation criteria was developed in an iterative fashion by CPRI and external experts. Emphasizing the fact that successful CPR system projects require a comprehensive, multidisciplinary team approach led by senior management, the evaluation categories are divided into four categories: (1) management, (2) system functionality, (3) technology, and (4) impact on quality, costs, and access. These categories reflect the notion that a CPR system implementation project is more than a selection or development of an information system product, it is the conceptualization of a new record keeping system that efficiently and effectively supports the delivery of health care and health promotion by the health care team. Existing organizational structures, departmental cultures, and an inward focus may have to be transformed to efficiently deliver high-quality care and effectively compete for managed care contracts. The CPR system is an essential tool to accomplish these broader goals.
Due to the barriers described above, leadership of and commitment to CPR system projects must come from the top of an organization. The organization's
Board of Directors must fund and support the implementation and use of CPR systems as a business imperative. Often, chief executive officers must act on personal conviction, existing evidence of benefits, and the desire to capitalize on a strategic differentiator—the value of data. In the new managed care market, a CPR system is not only an essential tool for health professionals, it is a business imperative. Time will tell whether investing in clinical systems provides a strategic advantage or is a minimum requirement to "stay in the game." Neither outcome favors continuing the status quo. It takes a combination of leadership conviction, organizational fiscal strength, and medical informatics and systems expertise to implement CPR systems effectively. Where such leadership exists, CPR systems are more likely to be acquired and successfully implemented.
A careful reading of The Computer-Based Patient Record by all stakeholders in health care is still fruitful. We believe that the original vision described by the IOM Committee on Improving the Patient Record was correct and remains timely and essential to the success of the new health care delivery system. Widespread use of CPRs would serve both private- and public-sector objectives to transform health care delivery in the United States. Equally important, early evidence indicates that the introduction and use of robust CPRs will enhance the health of citizens and reduce the costs of care, and in so doing, the use of CPRs will strengthen the nation's productivity.
Technological advances aside, progress toward CPRs as envisioned in this report has been slower than anticipated. The IOM committee expressed its strong belief that the early phase of CPRI's activities should be federally initiated and funded. A major coordinated national effort with federal funding and strong advisory support from the private sector is needed to accelerate the pace of change in the United States. Health care is a public good and many of the barriers to widespread implementation of CPR systems require national mandates, policy changes, or, in some cases, new legislation. Leadership in government and the private sector must be galvanized to make sweeping changes where possible (e.g., a national UHI, confidentiality legislation) and to instigate, motivate, and provide incentives to accelerate development of solutions to other impediments (e.g., terminology standards).
Data are the currency of quality management and the endowment for continuous quality improvement of patient care. Only by capturing primary clinical data from health care providers in a way that they can be applied to health care decisions for individuals and to policy decisions for populations can the United States achieve its goal of providing high-quality, affordable health care for all. A computer-based patient record is essential to accomplishing that goal.
We are deeply indebted to Octo Barnett who refocused our attention on the important points we were trying to communicate. The message is far stronger because of his constructive comments. We thank Don Detmer and Elaine Steen for their gracious invitation to contribute to this second edition.
Accredited Standards Committee X12N. Series of EDI standards available from Data Interchange Standards Association, Inc., 1800 Diagonal Road, Suite 200, Alexandria, VA 22314-2852.
ACR/NEMA Standards Publication. Digital Imaging and Communications in Medicine (DICOM). PS 3.1–3.13. National Electrical Manufactuers Association, 2102 L Street N.W., Washington, D.C. 1992–1995.
AMIA (American Medical Informatics Association) Board of Directors. 1994. Standards for Medical Identifiers, Codes, and Messages Needed to Create an Efficient Computer-Stored Medical Record. Journal of the American Medical Informatics Association 1:1–7.
ASTM Committee E-31 on Healthcare Informatics. Group of standards available from ASTM, 100 Barr Harbor Drive, West Conshohocken, PA 19428-2959.
Barnett, G. O. 1984. The application of computer-based medical-record systems in ambulatory practice. New England Journal of Medicine 310:1643–1650.
Barnett, G. O., J. J. Cimino, J. A. Hupp, and E. P. Hoffer. 1987. DXplain: An evolving diagnostic decision support system. Journal of the American Medical Association 258:67–74.
Barrows, Jr., R. C., and P. D. Clayton. 1996. Privacy, confidentiality, and electronic medical records. Journal of the American Medical Informatics Association 3:139–148.
Berner, E. S., J. R. Jackson, and J. Algina. 1996. Relationships among performance scores of four diagnostic decision support systems. Journal of the American Medical Informatics Association 3:208–215.
Chute, C. G., S. P. Cohen, K. E. Campbell, D. E. Oliver, and J. R. Campbell. 1996. The content coverage of clinical classifications. Journal of the American Medical Informatics Association 3:224–231.
Cimino, J. J., S. A. Socratous, and P. D. Clayton. 1995. Internet as clinical information system: Application development using the World Wide Web. Journal of the American Medical Association 2:273–284.
Classen, D. C., R. S. Evans, S. L. Pestotnik, S. D. Horn, R. L. Menlove, and J. P. Burke. 1992. The timing of prophylactic administration of antibiotics and the risk of surgical-wound infection. New England Journal of Medicine 326:281–286.
Classen, D. C., S. L. Pestotnik, R. S. Evans, and J. P. Burke. 1991. Computerized surveillance of adverse drug events in hospital patients [published erratum appears in Journal of the American Medical Association 267(14):1922, 1992] [see comments]. Journal of the American Medical Association 266:2847–2851.
Council on Competitiveness. 1996. Highway to Health: Transforming U.S. Health Care in the Information Age. Washington, D.C.: Council on Competitiveness.
Covell, D. G., G. C. Uman, and P. R. Manning. 1985. Information needs in office practice: Are they being met? Annals of Internal Medicine 103:596–599.
CPRI (Computer-based Patient Record Institute). 1995a. Guidelines for Establishing Information Security Policies at Organizations Using Computer-Based Patient Record Systems. Schaumburg, III.: CPRI.
CPRI. 1995b. Guidelines for Information Security Education Programs at Organizations Using Computer-Based Patient Record Systems. Schaumburg, III.: CPRI.
CPRI. 1995c. Proceedings of the First Annual Nicholas E. Davies CPR Recognition Symposium. Ed. E. B. Steen. Schaumburg, III.: CPRI.
CPRI. 1996a. CPR Systems Evaluation Work Group: CPR Project Evaluation Criteria (version 2.1). The Nicholas E. Davies Recognition Program. Schaumburg, III.: CPRI.
CPRI. 1996b. Guidelines for Managing Information Security Programs at Organizations Using Computer-Based Patient Record Systems. Schaumburg, III.: CPRI.
CPRI. 1996c. Proceedings of the Second Annual Nicholas E. Davies CPR Recognition Symposium. Ed. E. B. Steen. Schaumburg, III.: CPRI.
CPRI. 1996d. Sample Confidentiality Statements and Agreements for Organizations. Schaumburg, III.: CPRI.
CPRI. 1997. Proceedings of the Third Annual Nicholas E. Davies CPR Recognition Symposium. Ed. J. J. Teich. Schaumburg, III.: CPRI.
Detmer, D. E., and E. B. Steen. 1996. Shoring up protection of personal health data. Issues in Science and Technology 12:73–78.
Evans, R. S., J. P. Burke, D. C. Classen, R. M. Gardner, R. L. Menlove, K. M. Goodrich, L. E. Stevens, and S. L. Pestotnik. 1992. Computerized identification of patients at high risk for hospital-acquired infections. American Journal of Infectious Control 20:4–10.
Feldman, M. J., and G. O. Barnett. 1991. An approach to evaluating the accuracy of DXplain. Computational Methods and Programs in Biomedicine 35:261–266.
Grandia, L. D., T. A. Pryor, D. F. Willson, R. M. Gardner, P. J. Haug, S. M. Huff, B. R. Farr, and S. H. Lam. 1995. Building a computer-based patient record in an evolving integrated health system. Pp. 5–33 in Proceedings of the First Annual Nicholas E. Davies CPR Recognition Symposium, ed. E. B. Steen. Schaumburg, III.: CPRI.
Hammond, W. E. 1984. The Status of Healthcare Standards in the United States. Pp. 87–92 in Information Systems with Fading Boundaries, ed. W. E. Hammond, A. R. Bakker, and M. J. Ball. New York: Elsevier Science.
Health Level Seven (HL-7), Version 2.2. 1994. Health Level Seven, 3300 Washtenaw Ave., Suite 227, Ann Arbor, MI 48104-4250. 1994.
Hoechst Marion Roussel, Inc. 1995. HMO-PPO Digest.
Hripcsak, G., P. D. Clayton, T. A. Pryor, P. Haug, O. B. Wigertz, and J. Van der Lei. 1990. The Arden syntax for medical logic modules. Pp. 200–204 in Proceedings of the Fourteenth Symposium on Computer Applications in Medical Care. Washington, D.C.:IEEE Computer Society Press.
Humphreys, B. L., W. T. Hole, A. T. McCray, and J. M. Fitzmaurice. 1996. Planned NLM/AHCPR large-scale vocabulary test: Using UMLS technology to determine the extent to which controlled vocabularies cover terminology needed for health care and public health. Journal of the American Medical Informatics Association 3:281–287.
IEEE (Institute of Electrical and Electronics Engineers). 1995. Standards available from IEEE Standards Department, 445 Hoes Lane, PO Box 1331, Piscataway, NJ 08855-1331.
IOM (Institute of Medicine). 1993. Assessing Health Care Reform. Washington, D.C.: National Academy Press.
Johnson, M. E., K. B. Langton, R. B. Haynes, and A. Mathieu. 1994. Effects of computer-based clinical decision support systems on clinician performance and patient outcome: A critical appraisal. Annals of Internal Medicine 120:135–142.
Lindberg, D. A. B. 1995. The High-Performance Computing and Communications Program, the national information infrastructure, and health care. Journal of the American Medical Informatics Association 2:156–159.
McDonald, C. J. 1976. Protocol-based computer reminders, the quality of care, and the non-perfectibility of man. New England Journal of Medicine 295:1351–1355.
McDonald, C. J., S. L. Hui, D. M. Smith, et al. 1984. Reminders to physicians from an introspective computer medical record: A two-year randomized trial. Annals of Internal Medicine 100:130–138.
Miller, R., F. E. Masarie, and J. D. Myers. 1986. Quick medical reference (QMR) for diagnostic assistance. MD Computing 3:34–48.
NCPDP (National Council for Prescription Drug Programs). 1992. Standards available from NCPDP, 4201 North 24th St, Suite 365, Phoenix, AZ 85016.
NRC (National Research Council). 1996. The Unpredictable Uncertainty: Information Infrastructure Through 2000. Washington, D.C.: National Academy Press.
National Committee for Quality Assurance. 1997. A Road Map for Information Systems: Evolving Systems to Support Performance Measurement. National Committee for Quality Assurance.
Schoenbaum, S. C., and G. O. Barnett. 1992. Automated ambulatory medical records systems: An orphan technology. International Journal of Technical Assessment in Health Care 8:598–609.
Shortliffe, E. H., H. L. Bleich, C. G. Caine, D. R. Masys, and D. W. Simborg. 1996. The federal role in the health information infrastructure: A debate of the pros and cons of government intervention. Journal of the American Medical Informatics Association 3:249–257.
Sittig, D. F., and W. W. Stead. 1994. Computer-based physician order entry: The state-of-the-art. Journal of the American Medical Informatics Association 1:108–123.
Sullivan, F., and E. Mitchell. 1995. Has general practitioner computing made a difference to patient care? A systematic review of published reports. British Medical Journal 311:848–852.
Tang, P. C., and V. L. Patel. 1994. Major issues in user interface design for health professional workstations: Summary and recommendations. International Journal of Biomedical Computing 34:139–148.
Tang, P. C., D. Fafchamps, and E. H. Shortliffe. 1994. Traditional medical records as a source of clinical data in the outpatient setting. Pp. 575–579 in Proceedings of the Eighteenth Symposium on Computer Applications in Medical Care, ed. J. G. Ozbolt. Washington, D.C.
Tang, P. C., M. A. Jaworski, C. A. Fellencer, N. Kreider, M. P. LaRosa, and W. C. Marquardt. 1996. Clinician information activities in diverse ambulatory care practices. Pp. 12–16 in Proceedings of the 1996 AMIA Annual Fall Symposium, ed. J. J. Cimino. Washington, D.C.: American Medical Informatics Association.
Tierney, W. M., C. J. McDonald, D. K. Martin, and M. P. Rogers. 1987. Computerized display of past test results. Effect on outpatient testing. Annals of Internal Medicine 107:569–574.
Tierney, W. M., M. E. Miller, and C. J. McDonald. 1990. The effect on test ordering of informing physicians of the charges for outpatient diagnostic tests [see comments]. New England Journal of Medicine 322:1499–1504.
Tierney, W. M., M. E. Miller, J. M. Overhage, and C. J. McDonald. 1993. Physician inpatient order writing on microcomputer workstations: Effects on resource utilization. Journal of the American Medical Association 269:379–383.
Tierney, W. M., J. M. Overhage, B. Y. Takesue, L. E. Harris, M. D. Murray , D. L. Vargo, and C. J. McDonald. 1995. Computerizing guidelines to improve care and patient outcomes: The example of heart failure. Journal of the American Medical Informatics Association 2:316–322.