by William Rozycki
The most comprehensive electronic medical records system in the United States may be found in Indianapolis hospitals. When a patient comes into the emergency room of any hospital in the city, given the individual's consent, physicians can tap into the Regenstrief Medical Records System (RMRS) using a computer network line to learn important information about the patient's medical history. Blood type, drug allergies, previous illnesses, medical complaints, and prior blood chemistry tests are all available almost instantaneously. In some cases, the time saved by this knowledge can make the difference between life and death. Much of the credit for the RMRS goes to Clement McDonald, M.D., director of the Regenstrief Institute.
McDonald, now Distinguished Professor of Medicine at the Indiana University School of Medicine, had already begun his work as a pioneer in computer applications in medicine when he came to IU in 1972. "During my senior year in medical school," McDonald recalls, "I helped a professor with a project on the computer diagnosis of abdominal surgery problems. That was 1965, the era of mainframes, when data were punched on cards and stored on magnetic coils, a time well before the personal computer and the floppy disk. There was no packaged software; programming was a time-consuming and daunting task. The project was not a booming success, but it started me on my path of interest."
Clement McDonald, Distinguished Professor, Indiana University School of Medicine, and director of the Regenstrief Institute, (front), and Paul Dexter, clinical assistant professor of medicine, Indiana University School of Medicine, and research scientist of the Regenstrief Institute. Dexter is working with McDonald on implementing the computer-based order entry system and studying how the computer can help physicians provide better care for their patients by reminding them about needed therapies. --credit
After graduation from medical school and internship at the Harvard service of Boston City Hospital, McDonald delayed his training to more deeply explore the links between medical knowledge and computing. He entered Northwestern University and in 1968 earned a master's degree in biomedical engineering. The National Institutes of Health then offered him a fellowship to develop the first clinical laboratory computer system for its clinical center in Bethesda, Maryland. Next came a residency in internal medicine at Cook County Hospital and the University of Wisconsin. When McDonald completed that program, he received an invitation from Indiana University and the Regenstrief Institute, an independently endowed, nonprofit research institute affiliated with the IU School of Medicine and Wishard Memorial Hospital. In 1972, the institute was looking for someone who could lead research to determine how computer data storage and computation could enhance health care management and treatment.
The interest in computers was a natural extension of the vision and goals of Samuel Regenstrief. Regenstrief was a nine-year-old Viennese citizen when he immigrated to America in 1919, and he subsequently became a leading industrialist and endowed the Regenstrief Institute. Regenstrief's interest in applying technology to health care set the stage for the development of the RMRS. A nationally recognized authority on industrial production and management techniques, Regenstrief had risen to become a vice president for Philco, when he left the corporation in the 1950s. At that time, he established his own firm, convinced that postwar America wanted and needed automatic dishwashers. His company, based in Connersville, Indiana, went on to produce about 40 percent of the world's home dishwashing machines. Regenstrief, who died in 1988, believed that rational methods learned from scientific and industrial engineering could improve health care delivery and treatment. Long before his death, his Regenstrief Foundation had funded the institute that carries his name. The Regenstrief Institute's goals, then as now, include supporting the study of biomedical science, providing rational methods for using information and marshalling medical resources, and fostering the use of quantitative measures in medical decision and policy making.
The fit between McDonald and the Regenstrief Institute was perfect. After joining the IU medical faculty, McDonald coupled various teaching duties with the development of computer-stored medical records for the institute. The project began by recording the charts of thirty-five patients in the diabetes clinic at what was then Marion County General Hospital. Over the past quarter century, it has grown to carry data on 1.4 million patients and now holds, among other things, the records of 200 million clinical observations, 15 million prescriptions, and over a quarter million electrocardiogram tracings.
As he developed the system, McDonald and his collaborators became self-taught experts on two quite different aspects of biomedical computing: the design of an efficient and usable database and the problems that arise in linking different health care organizations, professions, and disciplines in a single medical records system.
In 1969, Samuel Nathan Regenstrief (1910-1988) formed the Regenstrief Institute, a joint enterprise of Regenstrief Foundation, Inc., the Indiana University School of Medicine, and the Health and Hospital Corporation of Marion County. The founders hoped that the techniques of biomedical science, computer science, and industrial engineering could be applied directly to health care. In 1975, the Regenstrief Health Center, with a contribution of $2 million from Samuel and Myrtie Regenstrief, was opened. The health center provides the outpatient clinics for Wishard Memorial Hospital and acts as a laboratory for the Regenstrief Institute to test its research in health care delivery systems. --credit
The first task in the early 1970s was to design the program. "There was no commercially available medical database software then," McDonald relates. "We wrote our own code. Every time we made changes to the structure of the records, we found we had to make changes to the program that stored the files and to the program that reported the information." Consulting with computer science faculty members at Purdue University, he learned the principles of modern database structure. "We learned to externalize the details of a program's functions in a separate file. The storage and reporting programs access that file and make adjustments. After that, we no longer had to change the basic program code each time the externals changed."
The more enduring task has been to capture data from many providers in a structured and usable form. "As a patient encounters health care providers," McDonald says, "he or she leaves a trail of medical information at many sites: the private physician's office, the hospital, the nursing home. Each of these may use different computer systems, and each provider may contract with different laboratories and use different pharmacies or radiology services." To compound the problem, each organization may have its own descriptions and terminology for procedures, diagnoses, and outcomes. McDonald gradually became convinced that the solution to the problem of merging data from these heterogeneous sources was to establish industry-wide standards in terminology.
As the founding editor of an international journal called M.D. Computing, McDonald has been in a position to advocate the adoption of such standards. Over the years, as his editorial advocacy galvanized health organizations into action, McDonald originated many standards. He organized and edited the clinical messaging parts of the American National Standards Institute's Health Level 7 standards from its inception in 1987. He also chaired both the Healthcare Informatics Standards Planning Panel and ANSI's Message Standards Developers subcommittee from 1992 to 1995. He has been a member since 1995 of the Commission on Nomenclature, Properties, and Units under the auspices of the International Union of Pure and Applied Chemistry. His work led to the adoption of the first widely used clinical message standard.
McDonald then turned his attention to another problem hampering the progress of medical databases, the lack of universal codes to simplify data entry. Since laboratory reports were invariably linked with clinical messages, McDonald set out to standardize the nomenclature and assign codes for laboratory orders, tests, and test variables. The result of these efforts was "Logical Observation Identifier Names and Codes" or LOINC, a database that assigns universal identifying codes for laboratory tests, vital signs, critical care measurements, obstetrical ultrasound measurements, and other elements of clinical practice. The adoption of LOINC allows Indianapolis health providers to easily encode data for computer processing. These standards are now freely available on the Web, and their adoption elsewhere makes electronic record keeping easier for all. McDonald reports with gratification, "LOINC is now in use by the largest commercial laboratories in the United States, by the big insurers like Aetna, and by health providers like Kaiser Permanente." The coding system has been adopted by the Veterans Administration and the entire Department of Defense, the health care systems of the Canadian provinces of Ontario and British Columbia, and the national health systems of Switzerland and New Zealand. In addition, the National Library of Medicine has mapped LOINC into the Unified Medical Language System. LOINC gains ground daily.
The mission of the Community of Scholars of the Regenstrief Institute for Health Care is to conduct research to improve health care by improving the capture, analysis, content, and delivery of the information needed by patients, their health care providers, and policy makers to measure the effect of the application of this research on the efficiency and quality of health care. --credit
All of these developments in database design, message standardization, and universal coding allow for an unparalleled merging of data within the Indianapolis health care system. Yet while emergency rooms can tap into the Regenstrief system to obtain information for individual patients under highly restricted policies, other uses are not allowed for reasons of confidentiality and proprietorship. "With the exception of emergency care in hospitals, we are simply the vault keepers," McDonald explains. "Each institution has the key to its own storage area, and only that institution can access it." Security for the system is high; all transmitted information is encrypted, and individual passwords, access logins, and electronic signatures provide user-level security. With Wishard Hospital serving as the network hub, the RMRS network today links all the major health care providers of Indianapolis from the university medical complex to neighborhood clinics.
Maintenance of patients' medical records using the RMRS can be invaluable to individual hospitals for a variety of reasons. While paper medical charts may be illegible, delayed, or even lost as they shuttle among multiple care sites, electronic medical records share none of these problems and can be simultaneously viewed by multiple health care workers. The RMRS allows health care providers to easily aggregate such data as patient characteristics, treatments, and disease types. The computer makes it easy to analyze those treatments that lead to improvement.
McDonald has demonstrated prescience in a variety of areas. As he merged data from multiple Indianapolis hospitals to improve patient care, McDonald avoided the collection of purely administrative data, understanding that it would be politically and economically charged. "Attempts in other parts of the country to make a comprehensive shared system started out with administrative data, with plans to add clinical data later," he says. "None have been successful. We keep focused on clinical data only."
In the future, McDonald hopes technological innovation will improve data entry in the health professions. A current difficulty faced by RMRS is the amount of time required to enter medical chart data into the computer. "Someday, maybe in ten or twenty years, we'll see physicians carrying around a small gadget that they will speak into, perhaps the size of a pager, summarizing patient observations, diagnoses, treatments, prescriptions. The messages will go wirelessly right into a computer database. Then, we'll truly see comprehensive electronic record keeping."
Asked about the promise and the reality of computers in medicine, McDonald pauses for a moment and then answers: "Computers are good for some things, like storing medical records and offering treatment guidelines and prompts. But the early hopes that computers would someday replace physicians in diagnosing disease is just not likely. I think attempts at developing that sort of thing are futile. Humans take in so much more than a computer does. We'll always need a human making those decisions."
For more information on the Web:
Return to the Table of Contents