Health care services not only are the prototypical knowledge business but also are perhaps the most complex product of our economy. Just as health care organizations have struggled to assimilate earlier generations of information technology (IT), they are likely to struggle to adapt to and use the Internet.
Health care providers and systems are staggeringly inefficient at assimilating and processing information and at converting that information to knowledge. Part of the problem is that the core knowledge base of health care, biomedical science, is expanding at a geometric rate, driven by $40 billion a year in public and private sector research and development (R&D) spending. Also, more variability and uncertainty at the point of service exists in health care than in any other service in our economy. Although this variability does not completely defy capture, standardization, and manipulation by information systems, the technical and organizational problems associated with this process are daunting.3 As if this variability were not complicating enough, more complex, highly trained, and difficult people (namely, health professionals) collide at the point of service than is true in any other service in our economy. Each health profession has its unique view of the patient's needs, its own language, and an intensely territorial view of its involvement in the care process. This has created a balkanized information architecture, in which each profession has its own data system that processes and records for payment the services it provides.
The present information environment in most health care institutions is dozens of functional computing systems (such as pharmacy, clinical laboratory, billing, and accounts receivable) running different programs written in different languages on different hardware. A depressingly large fraction of these processes are mediated by paper (medical records, prescriptions, telephone message slips, and bills) -- incontrovertible evidence of an early 1970s information environment.
Some health care organizations are adopting enterprise-wide information systems, with a single patient identifier, a single patient record, and a common application set. As J.D. Kleinke has noted, the growth and development of enterprise systems in health care has been deeply troubled. Vendors must shoulder part of the blame for promising solutions they cannot readily deliver; however, the difficulty health care organizations have had in shifting from functional to enterprise computing is, in major part, inherent in the complexity of the organizations themselves.
Indeed, it would be inaccurate to describe most health care organizations as enterprises. What they really are is collections of professions loosely and uncomfortably housed in the same physical structures. A coral reef is such a structure, much more a colony than a sentient being. As a consequence, systemic innovations are adopted very slowly. Passive resistance to change is compounded by a corrosive suspicion produced by the failure of past IT applications to materially improve productivity or processes of care.5 Clement McDonald and colleagues compared computer networks to a rain forest canopy, where arboreal creatures (physicians) can gather fruit (information on patients and clinical problems) effortlessly by moving across the canopy (data network) without having to climb each tree (separate data systems).6 (The image of troupes of monkeys screaming and throwing fruit at one another is almost irresistible.)
What the Internet promises health care managers and clinicians is a flexible, external information architecture that can reach down into the dozens, even hundreds, of health care information "silos" and extract, analyze, aggregate, and redirect data, which clinicians or managers need to make decisions. Beyond clinical uses, promising business to business Internet applications in health care include paperless transmission, verification, adjudication, and payment of medical claims; online marketing of health insurance to individuals and small businesses; paperless prescribing of, monitoring of, and payment for prescription drugs; medical product ordering and inventory management; and outsourcing of data processing and other management functions.