Last Updated: May 3, 1994 _________________________________________________________________________ Health Care and the NII DRAFT FOR PUBLIC COMMENT PART I: What Is the Application Arena? Description of a Health Care Information Infrastructure Implementation of wide-area, comprehensive, integrated, networked information systems is a logical response to the challenges faced by the Nation's health care delivery system. These challenges arise from several sources: dissatisfaction over rising health expenditures, in both private and public health care programs; concern over the personal health security issues of access and continuity of insurance coverage, and serious questions about the uneven quality and appropriateness of health care [reference 1,2]. These challenges are driving the health system to a cost-conscious, competitive, market-based, managed care environment. In such an environment, information systems linked to the National Information Infrastructure (NII) are destined to play a central role. The applications of the NII have significant potential for cutting unnecessary medical costs and improving health care access and quality. With the NII in place, consumers, physicians, other practitioners, hospitals, payers, and managers could readily obtain the information needed to make informed choices about treatments, providers, institutions, and health plans. With standards for defining, collecting, communicating, and storing administrative and clinical patient care data, scientific studies could point the way to medically effective and cost-effective care. National networks would enable all persons and health care providers to access the most recent information about particular medical technologies, clinical treatments, and provider performance. Patient outcome information could be linked to medical treatment data in a variety of settings so that all interested parties could obtain a better understanding of what works in the practice of medicine in the community and where it works best. In addition to improving clinical processes, the NII can simplify and speed up administrative processes within the health system, eliminating much duplication of paperwork and making uniform the data definitions required to make health care claims. As a result, electronic claims and payment transfers could occur rapidly over national networks and administrative costs would be significantly lowered. However, there is much infrastructure to build. A Vision of the Future SCENARIO 1: In a rural area, a child awakens with severe coughing, fever, and a rash on her chest. Her mother dials the interactive telecommunication connection to access medical care support and describes her child. The nurse at the other end asks for the mother to connect special probes that monitor the child's temperature, blood pressure, pulse. She then listens through an electronic stethoscope to the child's breathing. She examines the rash through the high resolution telecommunications viewer. After consulting information through the NII about recent health events reported in the community, such as the incidence of measles, bacterial and viral infections, she recommends action to the mother. Such action could be (1) stay on the connection and the physician will be right with her, (2) remain at home and continue to monitor the child and report in, (3) come in for an appointment with the doctor, or (4) head immediately to a designated emergency room. A valid medical encounter record is documented by this system and sent to the family's longitudinal medical service file, to the community's information repository, to the family for verification, and then to the family's health plan for payment. SCENARIO 2: A state public health official examines the state's health profile based on encounter records (with the identifiers removed) from health plans serving the state's communities. The records are retrieved from a statewide information network which is part of the NII. She is alerted by the information system to a statistically significant high incidence of children treated for respiratory disorders in a community. This leads her to call up the laboratory information from a sample of these children (the identifiers are removed but the information has been linked). In one of the cases, the laboratory results confirmed a diagnosis of whooping cough (pertussis). Immediately, she queries the immunization records and finds that some children do not appear to have been vaccinated. She then calls the community's health department to verify the data in the system. Finding it accurate, she queries the information system about the vaccine inventory in that community and, discovering it to be short of pertussis vaccine, calls four other communities with ample supplies to request that half of their vaccine be shipped to the first community. After notifying the first community of her actions and receiving their plan to resolve the problem, she returns to her examination of the state's health profile. SCENARIO 3: In the hospital of a major medical university in the state, Dr. Jones visits a virtual reality learning center to review procedures for a surgical removal of a portion of the prostate (prostatectomy). As she sits in the virtual reality clinical education room, she takes the electronic scalpel and feels the sensation of cutting into the patient, the texture of the skin, the hardness of the prostate as she is guided to making the proper incisions. The simulation program that guides her uses an electronic model human object obtained via the NII from a national library of reference models in conjunction with clinical measurement readings from the actual patient who will undergo the prostatectomy. Two floors up, Dr. Smith is performing a cataract surgery operation using robotics assistance. Although Dr. Smith is past middle age and has slight tremors in his hands, the robotics device with microsurgical vision enhancements eliminates the effects of his tremors. This supporting device allows his surgical productivity to continue for many years, increasing the life-long value of his medical training and years of experience. A Picture of Today Other sectors of the U.S. economy, some even less data- and information-intensive than the health sector, have for many years centered their operations around computerized systems. Banks, airlines, stock markets, and even salvage yards use computers to communicate, maintain inventory control, allocate costs, bill, and manage their major activities in an integrated, seamless manner. All these industries have experienced operating efficiencies, improved products and services, and, most important, greater customer satisfaction. These same benefits can be acquired for health care. The health sector, however, has lagged far behind the other sectors of our economy in applying information and communication technologies. Most hospitals and clinics have computers but relegate them to perform isolated, relatively small segments of the organizations' clinical operations. In these settings, the computer's widest use is for billing purposes and for patient admission, discharge and transfer functions, not for clinical purposes. Few hospitals and clinics link all caregivers together over local- or wide-area networks. As a result, patient care information is re-entered numerous times, information of value is not widely shared, and the paper outputs of these systems are manually collated in what is called a patient record. In this paper form, the patient record does not provide the basis for efficient clinical management, quality control, cost allocation, accurate billing, or clinical or health services research. Often the paper record and the information it contains is simply not available to the clinician when needed. The course of the patient through the health system is obscured by lack of documentation of the decisions, consultations, and sequence of interventions that he (or she) experiences. Thus, it is difficult to trace longitudinally the course of an individual patient, impossible to aggregate the data across a large number of similar patients, and improbable that all useful medical knowledge can be gleaned from the ongoing treatment of patients. Without reliable, comparative, performance feedback to the provider of health care, it is not likely that improvements in the quality of care or the efficiency of operation can be effected. Reliable feedback requires uniform vocabulary and coding standards for health care conditions, diagnoses, and procedures. Further, without an active communications interface among providers of care, it is difficult to make available -- especially in underserved urban and rural areas -- the benefit of the rapidly developing and evolving body of knowledge arising from biomedical and health services research. What is the Public Interest in Promoting the Application? Health care spending is high and growing. In 1994, the American public will spend $1 trillion on health care, nearly 15 percent of its Gross Domestic Product (GDP). National health care expenditures have risen by 10.5 percent per year for the past 8 years -- more than double the rate of increase in the consumer price index [reference 3]. Insufficient knowledge exists for informed decision making. Health and medical decision-making processes are flawed by a lack of knowledge and by financial considerations. The man (or woman) on the street has less knowledge about medical treatment alternatives for a specific condition than he (or she) has about any other service he buys. Therefore, he is more heavily dependent upon experts in the health care industry who often have no financial incentive to refrain from ordering every service, regardless of cost, if there is the hope of a benefit, however small. People do not pay the full price of the health care they consume. There would be no problem with the rapid rate of growth of national health expenditures and its portion of the GDP if it adequately represented consumer preferences expressed in the marketplace. After all, how much would the GDP have grown if not for the large increases in national health expenditures? Growth by itself is not bad. There is more than a suspicion, however, that when people pay 25 percent or less out of pocket for medical care at the time of choice, with insurance or public coffers paying the rest of the cost, there is a tendency to consume additional medical services. The value to the consumer of many of these additional services is less than the cost of the resources to produce them. How Can the NII Help? While the NII cannot change the U.S. health care system's financial incentives directly, it can support research into cost containment efforts and payment initiatives targeted at incentives to lower costs. For example, it could supply information to help appraise which payment systems in use are the most cost-effective. Further, the NII can provide information that increases knowledge about the medical effectiveness of alternative treatments and make it available to the providers and consumers of health care. The NII also can make available information consumers need to become more cost-conscious purchasers of health care services. The NII can provide an infrastructure that supports personal health improvement and medical technology assessment. Finally, the United States is one of the world's leading manufacturers of medical technology. With increased emphasis on cost-effective technology, there is a greater need for information about how well alternative technologies work when applied (1) in an ideal setting such as an academic medical center and (2) in the average community. The goal is to generate knowledge about which treatments and technologies work best for specific clinical conditions and under what circumstances, to have this knowledge available at the point of service (care), and to have medical decisions made jointly by caregivers and their patients. The NII can help attain this goal by supporting the analysis of large quantities of patient care and administrative data, by protecting its confidentiality, by assisting in the dissemination of information based on these data, and by adding value through the evaluation of the information gained from these data and converting it into useful knowledge. It is well recognized that there is substantial unexplained geographical variation in medical practices. The findings of unexplained differences in decisions about the best treatment for similar patients with the same condition elevates concern about the quality of care being delivered. Analyzing of patient care data from communities and providing feedback about these findings to the caregivers and consumers can both reduce inappropriate care and increase beneficial care. It can also improve continuous, life-long learning for health care providers who have difficulty keeping up with the flood of biomedical literature and clinical practice guidelines. By providing information access at home, schools, and the workplace, the NII can play an important role in improving public knowledge and decisionmaking about health, thereby reducing the significant information gap between consumers and clinicians and improving clinical outcomes. National and community networks that allow consumers to obtain information about their own health care conditions and to obtain professional medical advice in their homes can empower patients to take better care of them selves. What is the Evidence of the Benefits? Although the health care industry has been slow to adopt information and communication technologies in routine patient care, there are studies pointing to where the greatest benefits may be achieved. However, because these studies most often are conducted in single sites, both the size of the benefits to be achieved in multi-site and community settings and the costs of obtaining these benefits are unknown. Some examples of the studies follow: - In randomized controlled trials conducted at Wishard Memorial Hospital in Indianapolis by the Regenstrief Institute for Health Care at Indiana University, experimental groups of attending physicians that wrote their orders on microcomputer workstations were shown (1) prior test results for their patients, (2) computer predictions of abnormal results if another test was ordered, and (3) test prices at the time test orders were placed. These three trials showed reductions in outpatient test-ordering costs. A further randomized controlled trial tested the effect of physicians writing all their inpatient orders on workstations with screen information that encouraged cost-effective ordering. This intervention resulted in charges per admission reduced by $887 and hospital stays shortened by .89 days. However, achieving these savings did require more physician time per patient (33 minutes more over a 10-hour observation period, or 5.5 minutes per patient).[reference 4] - Including the costs, as well as the benefits, of implementing electronic data interchange for administrative health care activities (such as consumer enrollment, eligibility checking, billing, and claims payment), the Workgroup for Electronic Data Interchange (WEDI) believes "the cumulative net savings over the next six years (to the year 2000) is estimated to total over $42 billion."[reference 5] - An Automated Antibiotic Consultant software program was introduced into the HELP (Health Evaluation through Logical Processing) system at Latter Day Saints Hospital in Salt Lake City and evaluated. The Consultant was used when it was necessary to select an antibiotic therapy before the results of bacterial culture and susceptibility were known. The Automated Consultant suggested an appropriate antibiotic 94 percent of the time. The Consultant can support improved physicians' decisions under such conditions of uncertainty whether accessed on site or through a medical network.[reference 6] - The HELP system was also instrumental in determining the optimal timing (0 to 2 hours) of antibiotics before surgery to minimize the risk of post-surgical infection. The uniformity of data produced by such systems is advantageous not only for obtaining accurate patient care data, but also for conducting medical effectiveness studies.[reference 7] - A case study supported by the U.S. Public Health Service's Agency for Health Care Policy and Research (AHCPR) of a comprehensive hospital information system in one hospital in California, compared with systems in two other hospitals, was recently completed. The original system was installed in 1975. Now, almost 20 years later, the hospital continues to outperform comparison hospitals in all financial indicators. It experiences shorter patient stays (adjusted for mix and lower costs per admission.[reference 8] - A National Cancer Institute cancer treatment information system, Physician Data Query (PDQ), provides physicians and patients with information about state-of-the-art therapy and clinical treatment trials for each cancer and its stages. In PDQ, new literature and cancer prevention and treatment trials are continually reviewed by panels of clinical cancer experts, synthesized monthly, and updated. PDQ has experienced rapidly increasing use since its implementation and provides necessary information about the most recent cancer treatments and research findings worldwide using CancerFax and CancerNet. - The Comprehensive Health Enhancement Support System (CHESS) developed at the University of Wisconsin runs on a personal computer and offers a range of information, social and emotional support, and problem-solving tools for people in health crises. CHESS is typically placed in homes, but can also be installed in health care settings and community sites. CHESS currently offers modules for early stage breast cancer, HIV/AIDS, sexual assault, academic failure, adult children of alcoholics, and stress management. Such personal health information systems may grow to achieve user familiarity and acceptance, heavy use, quality of life improvements, and reductions in total costs of care. PART II: Where Are We Now? In many health care settings, patient information is handwritten in paper records and stored manually. Some of this is due to "state quill pen laws" that require handwritten pen and ink signatures on paper medical records. The current health information system does not adequately support patient care, medical effectiveness and cost effectiveness, and the public health of the community. This lack of support is often a result of incompletely recording the patient's signs, symptoms, and conditions; coding the patient's medical diagnoses to maximize billed charges instead of accurately describing the patient's ailment and the treatment given; and storing this information in ways that hinder both retrieval and making comparisons among patients with similar complaints. Consumers have insufficient information to make informed choices among the health insurance plans, health institutions, and providers available to them. Providers of care have insufficient means to keep abreast of all the information generated in their fields of specialty. Moreover, they often are unable to marshall all relevant information on a patient when making medical decisions. Health organization administrators are hampered in their ability to merge administrative and clinical information to make rational choices concerning resource allocations, quality of care, and product and service pricing. Payors of care have insufficient information to determine what package of benefits by which providers of care yield the best value for their clients. Further, public health officials should have the ability to more rapidly detect sharp increases in the incidence of influenza, specific bacterial infections, and other public health problems and to act quickly in health crises to inform the community. Public health policymakers often have insufficient information for offering solutions to health care problems. As a result, public health decisions are made without the advantage of timely, relevant information using technology that could reduce the costs of health care and improve patient outcomes and the health status of populations. The value of data on patient treatment and outcomes -- especially automated, uniformly defined, linked, and anonymously aggregated data -- is increasingly recognized and demanded throughout the health care sector. These data are needed for clinical, quality assurance, utilization review, business planning, administrative, and public health purposes. For example, computerized ambulatory patient care data are scarce and not uniform in definition, coding, or content. Computerized hospital clinical care data are collected on hospitalized patients in a small number of settings, but often are not stored for long in retrievable form after the patient is discharged. As valid methods for assessing the quality of care increase, so will the value of community patient care data. When the benefits from this information are shown to exceed the costs of producing it, society must find a way to pay for the resources necessary to produce it. Confidentiality and privacy are important concerns. Society must deal with perhaps its most vital information issue: assuring the privacy, confidentiality, and security of health care data about identifiable individuals. Even though patient care data can lead to important information for health care providers and their patients, it also has potential for personal harm if disclosed inappropriately. For example, these data may be required for emergency medical treatment or telemedicine applications in rural areas. As the data are transferred across wide areas, the system that transfers it must provide security against unauthorized access and disclosure, maintain the integrity of the data, and confirm the originators and requesters of the data. Quite possibly, most of the uses of patient care data may not require that the individuals be identified. When patient identification is necessary, the legal system must provide severe penalties for inappropriate uses of confidential patient care data. Although many States have their own privacy laws, many others do not. Moreover, uses of patient care data are not controlled uniformly from State to State. This problem must be addressed by national legislation. Selected Private Activities Private-sector activities discussed here include coordinated activities and projects in standards development, computer-based patient records, telemedicine, and community health data repositories. Standards. In the private sector of the United States, the development of medical information standards is coordinated through the American National Standards Institute (ANSI) Healthcare Informatics Standards Planning Panel (HISPP). AHCPR, in cooperation with the Food and Drug Administration, supports the meetings and administration of HISPP. The HISPP is also the official link between U.S. and European standards developing organizations. Other countries (particularly the European countries) are making notable progress with central development of medical information standards. Pursuing international cooperation in the development of these standards could prove beneficial for standards development in the United States. Administrative health data standards are being developed by the ANSI-accredited standards committee X12, through its subcommittee X12N. The Workgroup for Electronic Data Interchange is a private sector advisory body that has provided much leadership for these standards, which are essential for electronic exchange of health insurance business information. Progress has been faster for administrative health data standards than for clinical health data standards. The development of standards for the electronic interchange of clinical data is becoming more coordinated under the Message Standards Developers Subcommittee of ANSI HISPP. These standards will permit standardized data flows among departments of a hospital, for example, and among hospitals, physicians, and other medical organization entities. Some examples of standards for the coding of medical diagnosis and procedures are the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) and Current Procedure Terminology, 4th Revision (CPT-4). The ICD-9 coding system was developed by the World Health Organization and modified in the United States by the Health Care Financing Administration and the National Center for Health Statistics to produce the ICD-9-CM. The American Medical Association maintains the CPT-4 coding system. These codes, however, are used more for billing purposes than for their clinical information. They do not code the signs, symptoms, and conditions of the patient upon which the diagnosis is made. The Systemized Nomenclature of Medicine (SNOMED), 3rd Edition, does code for signs and symptoms of diseases as well as for disturbances in biochemical and enzyme factors of interest to the clinical pathologist and for anatomy, pathology, and etiology upon which the diagnosis is made. However, clinical codes that better reflect the conditions of the patient and the treatment received are needed. Further, much work is needed to develop standard validated measures of the patient outcomes of medical care. Although coordination of standard development efforts continues to improve, the pace of clinical data standards development has much room for improvement, and there are many gaps. What is the reason for the slow development of clinical data standards? Because no one firm in the private sector can generate sufficient gains from developing clinical data standards, the development effort is voluntary, and it suffers from a lack of resources. Further, it takes time to reach consensus, and the benefits to the public are diffuse. If the public good is sufficiently large, a proper role for the Federal Government is to accelerate standards development and collaboration in the private and public sectors. Computer-based patient records. Created in 1991, the Computer-based Patient Record Institute (CPRI) promotes and coordinates the development of CPR systems in the United States. The CPRI is composed of representatives from physician, hospital, computer system, vendor, managed care, university, and other national groups. The CPRI is working for the ubiquitous use of CPR systems in medical care, with workgroups in four areas: Codes and Structure; CPR Systems Evaluation; Confidentiality, Privacy, and Legislation; and Professional and Public Education. Several private sector-projects are attempting to build a computer-based patient record and to export common definitions and reporting systems to all their sites nationwide. Two of these efforts have begun at Kaiser Permanente in California and at the Mayo Foundation in Minnesota. Although there is strong interest in finding the best information systems for particular health-care organizations, there is no common method for evaluating existing systems. Additionally, there are differences in opinion regarding which costs and benefits to include in such evaluations, and how to place a value on outputs and resources that are not priced in the market place. In addition, more knowledge is needed about the technical, legal, social, and economic barriers to the development and deployment of computer-based patient record systems. Telemedicine. In early 1994, Ameritech, Inc., demonstrated a system that enables the family of a person with a serious chronic illness to access medication at regular time intervals and to obtain direct consultation with the family physician via personal computer, television cameras, and special hook-ups. This project can reduce visits to the hospital emergency room and physician's office, while improving patient compliance with drug therapy. Another example of a personal health information system has been described above (see CHESS). The private sector also supports several telemedicine projects such as those in Texas and Georgia. Although payment for telemedicine services has been nearly impossible to obtain from insurance programs, these projects continue to support patient care. Additionally, they are informing health care policymakers and potential entrepreneurs about the potential for cost savings and increasing access particularly in rural areas -- and about the social and legal concerns that must be addressed when providing medical care across State borders. Most of the potential savings are projected to come from reduced transportation expenses required to bring physician and patient together in the absence of telemedicine. Community Data Repositories. Stimulated by grants from the John A. Hartford Foundation, Community Health Management Information System projects have been initiated in the States of Washington, Iowa, Vermont, New York, Minnesota, and Ohio and the city of Memphis. These projects aim to extract patient, provider, and service data from claims and encounters and store them in a shared community data repository. As desired by the community, the repository may be enhanced to include condition-specific data and patient-centered surveys. In addition, many other health information networks are in planning stages, although without a common model for health information networks to guide them. The goal of these projects is to begin the development and implementation of local health data networks and data repositories. The expected benefits are improved health business transactions and community health data repositories of standardized health data about the process and outcomes of health care delivery. Demonstrations such as these can show the advantages of collecting standardized health data about the community. The potential for NII to link national and community networks with homes, offices, and health institutions to facilitate improved and cost-effective health care should be demonstrated and evaluated. This information would increase understanding about the sources of efficiency and the size of the costs relative to the benefits. Armed with this knowledge, private and public ventures, when targeted to obtain demonstrated cost savings, would be less risky. Selected Public Activities Department of Health and Human Services The National Library of Medicine (NLM) contributes to the NII on many different levels. Foremost, NLM produces data bases and information services that provide access to the scientific knowledge underpinning biomedical research and health care. NLM's on-line data bases and data banks are the most widely used medical information resources in the world. The expanding Internet provides an enhanced vehicle for delivering NLM services -- from MEDLINE to GenBank -- including on-line access to clinical practice guidelines that combine images and full text; the developing "Visible Human" digital library of image data from photographic, computed tomography, and magnetic resonance imaging of complete male and female bodies for a range of educational and health care applications; and remote execution of sophisticated gene sequence matching algorithms. Since the mid-1980's, the number of clinicians with direct access to MEDLINE has increased dramatically. More than 34,000 individual health professionals now search NLM's computer system, up from just a few hundred less than a decade ago. "Physicians report that in situations involving individual patients, rapid access to the biomedical literature via MEDLINE is at times critical to sound patient care and favorably influences patient outcomes.[reference 9] The Unified Medical Language System (UMLS) Project of NLM focuses on linking terms and codes in patient records to evidence-based knowledge such as that in practice guidelines and the scientific literature. Through the development of the Metathesaurus that connects the various computerized coding schemes and controlled vocabularies and an Information Sources Map that will support automated selection of data bases containing information relevant to particular information needs, the UMLS is providing tools for successful navigation among the growing number of health care information resources and for capturing and encoding patient data. The NLM's High Performance Computing and Communications (HPCC) program funds NII applications research in the areas of test-bed networks linking health care organizations, telemedicine, and development of computer-based patient record systems, coordinating and in some cases co-funding projects with other agencies such as AHCPR and the Department of Defense, Advance Research Projects Agency. Supporting the development of computer-based patient record systems is also a priority for NLM's National Information Center on Health Services Research and Health Care Technology (NICHSR), since data collected as a by-product of current health care delivery could greatly enhance the quality and timeliness of health services research, including outcomes studies. To ensure that academic medical centers, hospitals, and members of the National Network of Libraries of Medicine can access and make effective use of the NII, NLM and the National Science Foundation are co-sponsoring a Medical Connections program to link these institutions to the Internet. The NLM's programs to support the development of enterprise-wide Integrated Advanced Information Management Systems (IAIMS) and to provide Medical Informatics training to health and information professionals create a receptive environment for effective use of the NII. The NLM's current Director serves as the Director of HPCC's National Coordination Office, which is housed at the NLM and reports to the President's Office of Science and Technology Policy. AHCPR, through its Medical Treatment Effectiveness Program (MEDTEP), adds to a knowledge base of medical effectiveness research findings and clinical practice guidelines that inform practitioners and their patients about what works best and for whom. AHCPR is collaborating with NLM to make guidelines available in a form compatible with NII applications. In addition, AHCPR supports the development and evaluation of computer-based patient record systems and attendant clinical computer decision support systems. AHCPR assesses their impact on the medical and cost effectiveness of health care and their potential to generate uniform, accurate patient care information for medical effectiveness researchers. Through its extramural program, AHCPR has for years supported the study of computer-based patient records and clinical decision support systems. Through its Office of Science and Data Development, AHCPR promotes the coordination of the developers of patient care data standards and the analysis of confidentiality and privacy issues concerning researcher access to patient care data. The Food and Drug Administration (FDA), through its Submission Management and Review Tracking (SMART) program, is developing a consistent approach to the electronic submission and review of drug, biologic, and medical device applications. As part of this project, FDA is working with its regulated industries, the clinical community, the World Health Organization, and the European Union to enhance and standardize nomenclature for coding clinical safety data. FDA also is pursuing nomenclature and data format standards for other data elements required in product submissions, including toxicology, pathology, and patient information. These standards, coordinated with the private sector, will improve the uniformity of patient care data. The Health Resources and Services Administration, through its Office of Rural Health Policy, is undertaking support of telemedicine project application and evaluation in rural areas. These applications should stimulate the outreach of NII health care applications to rural America. The Health Care Financing Administration (HCFA), the Nation's largest health insurer, has the most comprehensive health care data base in the world, which supports the management of the Medicare and Medicaid programs. Much of this data is managed in a large network, linking Medicare enrollment and benefits information to both the contractor claims payment sites and providers. HCFA electronically receives more than 90 percent of the institutional providers' claims and more than 65 percent of the individual providers' claims, which collectively require HCFA to process over 500 million claims annually. HCFA agents use clinical data to assure quality and appropriateness of care provided to beneficiaries. HCFA shares this information with other organizations to facilitate research and formulate health care policy. HCFA has been an active participant in the development of data standards, and was the first among the health insurance groups to deploy electronic health insurance claims standards. HCFA is an active member in ANSI-Accredited Standards Committees and ANSI's Healthcare Standards Planning Panel. HCFA also supports research on telemedicine projects in rural areas and on payment methodology for telemedicine consultations. In 1985, HCFA, in cooperation with the National Center for Health Statistics (NCHS), developed a process for updating and maintaining ICD-9-CM through the Coordination and Maintenance Committee. HCFA is responsible for maintaining Volume 3, Procedures, while NCHS is responsible for Volumes 1 and 2, Diagnoses. Proposed revisions to the coding system are received by the appropriate agency, researched, developed in a standardized format, and publicly presented for informational purposes. Final approval for any coding changes comes from the Administrator of HCFA and the Director of NCHS. The changes are made annually. To address the shortcomings of the ICD-9-CM procedure codes, HCFA investigated developing a new procedure coding system and contracted with 3M, Health Information Systems, Inc., to develop a prototype. Six chapters thus far have been developed, with the cardiovascular chapter being the most complete. The National Cancer Institute, in addition to the PDQ program described above, is supporting the development of technology that could make use of a national broadband network capability for digital mammography. If successful, this technology could electronically transmit breast images for the purpose of obtaining rapid and expert radiologic consultation. This future application of teleradiology requires significant improvement of digital imaging modalities to acquire primary digital images, film digitizers to translate conventional film images into digital form, digital data networks, image-processing algorithms, computer workstations, and other imaging technology. Within the Public Health Service, Office of the Assistant Secretary for Health, the Office of Disease Prevention and Health Promotion is developing a community services workstation. With support from other PHS agencies and the Department of Defense, Advanced Research Projects Agency, a prototype of the workstation is being created to illustrate how information about health, education, and welfare service availability may inform a community's population. Department of Commerce The National Telecommunications and Information Administration (NTIA) was allocated $26 million for FY 1994 to support the development of the NII. The purpose of the NII is to interconnect the nation's businesses, residences, schools, health care facilities, and other public information and social service providers through broadband interactive telecommunications networks. During FY 1994, NTIA intends to issue grants in the area of health care for pilot demonstrations that are designed to develop, demonstrate, and promote applications of information technology that will educate, restrain health care costs, improve quality, and increase access to health care, with the potential for wide-scale deployment and interconnection over NII networks. Also within the Department of Commerce, the National Institute for Standards and Technology (NIST) has been active in several standards development efforts related to health care, including (1) Health Level 7 and (2) the Health Care Special Interest Group of the Open Systems Interconnection (OSI) Working Group. The purpose the OSI Working Group is to encourage broad scale adoption of medical information standards developed in the United States, leading to information systems that will exchange data with each other and operate on a variety of computer operating systems. The NIST also has a cooperative research and development agreement (CRADA) with private industry in medical information systems/architecture standards, which draws upon NIST's experience with electronic data interchange. Department of Veterans Affairs The Department of Veterans Affairs (VA) operates 171 medical centers located in all 50 States and Puerto Rico. All 171 facilities are supported with a hospital information system called the Decentralized Hospital Computer Program (DHCP). DHCP has approximately 60 public domain software modules supporting functions such as admissions, pharmacy, laboratories, medicine, order entry, health summary, engineering, purchasing, and finance. Both the Department of Defense and the Indian Health Service, as well as some State institutions, have used this public domain software, modifying it to meet their particular needs. Recent clinical releases, such as Discharge Summary, Progress Notes, Allergy Tracking, and Problem List are part of an incremental approach to automating the patient record. The Problem List serves as a point of integration, organizing patient information by clinical discipline and associating patient treatments and outcomes with problems. It also ties the automated patient record to appropriate billing procedures. The Problem List works with VA's new Clinical Lexicon; a tool that permits the user to enter a clinical term using natural language. The system then maps the term to all of the applicable coding schemes (e.g., ICD, DSM, SNOMED, CPT) and stores the term so that it can be retrieved and used by a wide variety of legitimate users of clinical data. All VA facilities have been interconnected for the past 8 years with a digital communications network. VA is enhancing its data transport utility called Patient Data Exchange (PDX) so that VA health care facilities can exchange health summaries containing relevant clinical data across the VA network. At the Washington and Baltimore Medical Centers, VA has installed an integrated imaging project to store medical images as an integrated part of the electronic patient record. These systems have been in daily use for several years and store images from such applications as pathology slides, gastrointestinal, bronchoscopy, cardiac catheterization, echocardiography, and radiology examinations. At the Baltimore Medical Center, the VA has built a standards-based gateway between a hospital in formation system and a commercial Picture Archiving and Communications System. With these capabilities, VA could be a useful test bed to determine the effectiveness of electronic data exchange between facilities and health care providers. As a result of the common software at many Federal medical care facilities, this test bed could easily be extended to other agencies as appropriate. An example could be the effectiveness of making AHCPR-sponsored clinical practice guidelines available to clinicians on VA wards using connections to Internet. Department of Defense The Department of Defense, Office of Health Affairs has supported the development of a Comprehensive Health Care System leading to six modules that are being deployed world-wide. The Department was also charged by Congress to deliver a plan for developing and deploying a computer-based patient record in military settings in early 1994. The military's Medical Diagnosis Imaging Support System has developed and demonstrated projects that involve filmless radiology departments in hospitals, electronic transmissions of digital images via satellite, and the use of the most recent standards in digital image transmission. Department of Agriculture Within the Department of Agriculture, the Rural Electrification Administration (REA) plays a key role in the rural aspect of the NII. The Distance Learning and Medical Link Grant Program (DLMLGP) demonstrates the ability of rural communities to utilize existing or proposed telecommunications systems to achieve sustainable, cost-effective distance learning or medical-link networks. Rural schools, libraries, hospitals, health care clinics, and related organizations that operate rural educational or health care facilities are eligible. Implemented in FY 1993, the DLMLGP has selected 28 rural projects (9 medical projects) for funding, and is currently reviewing applications submitted under the second round of funding. REA funds equipment used for distance learning classrooms such as encoding and decoding devices, specialized cameras and video monitors, video switchers, microphone mixers computers, and local area networking equipment. For medical link projects, REA funds equipment used in physician consultation, teleradiology, and educating rural health care providers. Some of this specialized equipment includes teleradiology workstations, X-ray scanners, digital microscopes, and all of the above distance learning equipment. Applicants to the DLMLGP work closely with local telecommunications providers creating a demand for bandwidth capacity and switching technology in remote areas. In this respect, the DLMLGP accelerates the development of rural networks, one of the most challenging areas of the NII. National Aeronautics and Space Administration The National Aeronautics and Space Administration (NASA) has been a pioneer in the field of telemedicine since the early days of manned space flight. NASA's interest in telemedicine is to understand its application to medical care in space for future long duration platforms, such as a space station, and to minimize risk to astronauts and increases probability of mission success. NASA's early efforts to monitor the health of its crew members have helped promote dramatic changes in the way medical monitoring in terrestrial medical transport is conducted in the U.S. NASA has been involved in several telemedicine projects over the past 30 years. Space Technology Applied to Rural Papago Advanced Health Care Program (STARPAHC), brought medical care to remote areas of the Papago Indian Reservation in Arizona in the 1970s. The Space bridge to Armenia provided satellite consultation to a disaster area in 1988. Currently, NASA is involved in a joint effort with Russia, the Space bridge to Moscow, to link several U.S. medical centers with a hospital in Moscow. NASA's experience in telemedicine and communications technologies has helped promote the practice of telemedicine across the globe. The Armenia experience demonstrated that interactive consultation by remote specialists can provide valuable assistance to onsite physicians and favorably influence clinical decisions in the aftermath of disasters. PART III: Where Do We Want to Be? To obtain the benefits called for by the vision of the health care information system of the future, an advanced NII should support the development and evaluation of information technology applications that can improve patient care, both directly and indirectly. These achievements would improve the health status of communities and reduce their costs of health care. Applications that bring both higher benefits and lower costs should be carefully evaluated. Patient care data describing patient's signs, symptoms, and conditions; treatment; and outcomes should be generated at the point of care delivery by the providers of health care. These data should be defined uniformly across all points of care, automated, and made available through the NII for direct patient care; public health policy development at community, state, and national levels; and research purposes. This sharing should occur only under conditions of confidentiality, privacy, and responsibility that are acceptable to society. Patient care data and other information necessary in the direct care of the patient should be promptly available to providers of care at the site of care. Clinical decision support systems should incorporate research findings based on studies of these data and on other studies. The purpose is to give providers of care information about drug interaction alerts, allergy alerts, preventive screening reminders, and other prompts that improve the delivery of health care. Personal health information should be widely available on the NII and be accessed through personal computers and telephone links, cable television, or other links to community and nationwide networks. This linkage will permit people to obtain health care information, computer-assistance for analyzing health problems, and advice from medical professionals and from people with similar health conditions. The result of improving personal self-care and wellness should be more power in the hands of the people to influence their health and a more appropriate use of health care resources. Public health surveillance and epidemiologic studies based on patient care data and social indicators should be available to inform public policy and to guide the provision of public health services. Information about the patient outcomes of care produced by health care providers and health care plans should be available to guide consumers in making health plan choices and to feed back information to providers of care about the patient outcomes their peers are achieving. This information will benefit consumers in their homes, schools, and workplaces. Providers of care in physicians' offices, other ambulatory sites of care, and hospitals will have access to data about specific patients and the information, if necessary, to guide decisions about treatment alternatives and their expected outcomes. Health care managers and policymakers in health plans, public health departments, national health policy positions and other settings will be able to develop an overall picture of health care utilization to assess the allocation of health resources and whether private and public health needs are being met. Achieving these benefits requires the development of several components of a health information infrastructure. These components are: - Medical information standards for the - nomenclature, coding, and structure; - content of specific data sets; and - electronic data interchange of patient care data. They are necessary to achieve the uniformity of definition and meaning of the patient care data used in the care of the patient and in generating information about the outcomes of care. The standards will improve the sharing of patient care data across different computer information systems. The slowness of there development of these standards hinders the cost-effectiveness of clinical decision support systems in institutional and provider settings. - Unique personal identification for accurate links across databases used for patient care. Although the social impact and confidentiality issues are the most important for society, technical issues still remain. Patient information must be uniquely identified and linked across databases used for patient care. Some options are thumbprints, retinal eye scan images, DNA blood typing, or personal identification numbers in digitized form. If personal identification numbers are used, they could be social security numbers (SSN) or identifiers unique to health care. The costs and security of different techniques to assure unique personal identity, plus confidentiality and privacy of patient care information and any information to which it may legitimately be linked, need to be investigated. - Model development for health care information, reference requirements, and a reference architecture to define and relate patient care data and medical information and the clinical and administrative functions they serve. A concept model should be developed that serves as the guiding framework that shows the purpose, dimensions, and minimum characteristics of health information networks, computer-based patient record systems, and other concepts. The concept models may pertain to specific domains, such as hospitals, clinics, and local networks. These models, requirements, and architecture will provide a common framework that will allow software vendors and system designers to build software tools that can work together. If they work together, these tools can fill out, or build, the health care application architecture (the common framework). By supporting the design, building, and implementation of systems that can interact with each other, this framework, and tool development, will support improved patient care and build a path for the movement of existing systems to patient-centered systems. - Federal confidentiality and privacy laws that supersede a patchwork quilt of State privacy laws. They will allow society to gain the benefits of rapid automated information transfers across States through information technology, while protecting patient care data from disclosure. They should provide penalties for inappropriate linking, use, or disclosure of patient care data and define inappropriate use. - Health data repositories to maintain and assure the uniformity and confidentiality of patient care data and to provide access to the appropriate users of these data. These repositories might be distributed among local communities or located regionally across the United States. At the extreme, there could be one central or national data repository. Safeguarding the confidentiality of patient-identifiable data wherever it is stored is essential and must be a prime responsibility of the depository management. - Computer-based patient record system development to capture patient data at the point of care and make it available electronically upon request of the provider for patient care. This development should extend computer-based patient record systems so they support both clinical and administrative decisionmaking. - Health Care Computer Laboratory (test bed) development to determine the technical usefulness of data standards and data exchanges that support specific functions. Findings from these pilot test sites should guide modifications to data standards, models, and architectures to make them suitable for commercial applications. - Pilot tests and evaluation of health information technology in patient care settings such as the home, physician's office, hospital, and community. These pilot tests and evaluations should include rural as well as urban settings, and consumer as well as provider settings. They should reveal where the most beneficial applications are likely to be. - Community trials for applications that have been proven successful in single site settings. These trials would be linked in broad-area studies to assess their scalability (i.e., their costs and performances at different volume levels and configurations) and their success in achieving quantifiable savings that can be duplicated. - Specific studies should evaluate the economic and medical feasibility of patient care data transfers between primary care physicians and specialists across geographic distances, of the use of personal home information systems to promote wellness and efficient use of medical services, and of administrative electronic data systems to improve the efficiency of medical claims handling and payment. These transfers include telemedicine transmissions such as medical and patient images, consumer health information and decision analyses, and consultations with experts and patients with similar conditions. Additional studies should evaluate the potential for libraries of information on standard representations of medical conditions to be accessed by providers to improve their understanding of patient conditions, disease entities, and healthy body functioning which, in turn, should improve patient outcomes of care. Studies of alternative means for professional education and training using the NII should also be undertaken. PART IV: How Are We Going to Get There? Issues and Questions to be addressed - How can the federal government facilitate a public-private partnership for the development of health care information standards? What should the priorities be among: -- Medical information standards for the nomenclature, coding and structure, content of specific data sets, and electronic exchange of patient care data, information, and medical knowledge. -- Standards for electronic signatures, especially for the validation of physicians' orders and other information within health information systems. -- Standards for personal identification methods. With regard to personal identification, how best should the federal government examine the benefits and costs of alternative schemes, including (1) improving the Social Security Number, (2) creating and implementing a new unique health identification number for the population, and (3) adopting any other identification system, such as a private-sector numbering system, retinal scans, or fingerprint readers? - When developing standards, data sets, and formats for regular reporting of patient care data used in the programs of federal agencies, how should the agencies work cooperatively with the private-sector standard developers to achieve common standards and to accelerate the private sector's development of standards to meet the programs' needs? - How should federal confidentiality and privacy legislation be developed that would apply civil and criminal penalties for inappropriate use of personally identifiable patient data, including inappropriate disclosure and redisclosure of that data, and discriminatory decisions based on linking such data with other personal data bases? -- Should this legislation supersede State privacy laws to achieve nationwide uniformity in the way the uses of personally identifiable patient data are restricted? -- How should information system security requirements be addressed? - How should information about the health care information system technologies being developed, placed in use, and evaluated by federal agencies be coordinated? -- Due to the diverse missions, technologies, service delivery programs, and research and demonstration programs across federal agencies, should there be a federal work group established for coordination to reduce unnecessary duplication by sharing program information that could improve NII health care applications? - Should the federal government support a public-private partnership to develop integrated health care information systems in hospitals, physicians' offices, and other health care settings? -- Should this partnership develop models for health care information and reference requirements and architectures for supporting the functions that require patient care data? -- Should this partnership develop tools for software developers and system designers? -- Should this partnership support the development of the computer-based patient record for capturing patient care data at the point of care for clinical and administrative decisionmaking. -- Should this support be based on a study of the expected benefits and costs of these systems and target to those areas where the benefits are most strongly expected to exceed costs? -- Should this support include a portion of the capital cost of these systems? -- Should this support be coordinated with existing telemedicine projects aimed at increasing health care access for rural populations? -- Should this support include investigating the training necessary for these systems to be adopted by health professionals and their staffs? - Should the federal government undertake a program of research, demonstration, and evaluation of integrated health care information systems to identify the specific applications for which the benefits exceed the costs? -- Should this program participate with the private sector in the funding of pilot tests of these systems in single-site pilot tests, multiple-site trials, and community-wide demonstrations? -- Should this program be coordinated with standards developing efforts to test preliminary standards and provide feedback to the developers for revisions to the preliminary standards? -- Should this program support demonstrations that show the benefits and costs of linking clinical decision support systems in hospital and physician office settings to national sources of medical knowledge? -- Should this program support demonstrations showing the value of personal health information systems that provide access to health care knowledge to individuals in their homes, in their workplaces, and in public libraries? -- Should this program investigate the costs and benefits of nationwide networks for carrying out administrative health care functions, including enrollment, eligibility checking, claims processing, and electronic funds transfer? -- Should this program investigate the potential administrative and clinical cost savings, the changes in health care service utilization behavior, and the change in consumer satisfaction from providing access to financial and medical effectiveness information to community populations and their health care providers? -- Should this program investigate the expected changes in the process of health care and in the patient outcomes of care that result from use of clinical decision support systems in relevant health care settings that are linked to national information sources? - Should the federal government support a public-private partnership to develop Federal strategies for repositories of health care data? -- Should this partnership fund workshops and projects that support the development of important community issues, new concepts and policies concerning health care data that will be needed for addressing control, access, and accountability questions in the NII? -- Should a research program investigate the value of community health care data repositories for efficient and effective deliver of health care services and for improving patient outcomes in the community? -- Should Federal support include partial funding of community health care data repository demonstrations? -- Should Federal support include funding evaluations of the community uses of their own data, determining what are the benefits to the community, who else receives benefits from such repositories, and what are the costs of achieving those benefits? - Should the federal government support State efforts to use the NII to promote and evaluate the public health of their populations and the administrative efficiency of their health care programs? -- Should there be Federal support for States to demonstrate the benefits to state and local public health functions of community data repositories linked at the state level? References [1] Wennberg, J. and Gittelsohn, A. (1982). Variations in medical care among small areas. Scientific American, 246,120-134. [2] Chassin, M., Kosecoff, J., Park, R. E., Winslow, C. M., Kahn, K. L., Merrick, N. J., Keesey, J., Fink, A., Solomon, D. H., and Brook, R. H. (1987). Does inappropriate use explain geographic variations in the use of health care services? A study of three procedures. Journal of the American Medical Association, 258, 2533-2537. [3] Health Care Financing Administration, Office of the Actuary, October 21, 1993, personal communication. [4] Tierney, W. T., Miller, M. E., Overhage, J. M., and McDonald, C. M. (1993). Physician inpatient order writing on microcomputer workstations: effects on resource utilization. Journal of the American Medical Association, 269, 379-383. [5] Workgroup for Electronic Data Interchange Report (October 1993). 1993 WEDI Report, Executive Summary, p. iii. [6] Evans S., Pestotnik S. L., Classen D. C., and Burke J.P. (1993). Development of an automated antibiotic consultant. M.D. Computing, 10, 17-22. [7] Evans R. S., Reed M. G., Burke J. P., Pestotnik S. L., Larsen R. A., Classen D. C., and Clayton P. D. (1987). A computerized approach to monitor prophylactic antibiotics. Proceedings of the Eleventh Symposium on Computer Applications in Medical Care (pp. 241-245). Silver Spring, MD: IEEE Computer Society Press. [8] Kunitz and Associates, Inc. Final Report (Revised March 28, 1994). Agency for Health Care Policy and Research, Contract 282-91-0062. [9] Lindberg DAB, Seigel E. R., Rapp B. A., Wallingford K. T., and Wilson S. R. (1993). Use of MEDLINE by Physicians for clinical Problem Solving. Journal of the American Medical Association, 269,3124-3129. .