An interesting and important debate has been taking place in the pages of the British Medical Journal (http://www.bmj.com) which goes to the heart of the internet promise or nightmare.

from Insights, November 1998

Open up your office and say ahhh.

Upon this gifted age, in its dark hour,
Rains from the sky a meteoric shower
Of facts ... they lie unquestioned, uncombined.
Wisdom enough to leech us of our ill
Is daily spun, but there exists no loom
To weave it into fabric.

—Edna St. Vincent Millay

An interesting and important debate has been taking place in the pages of the British Medical Journal (http://www.bmj.com) which goes to the heart of the internet promise or nightmare.

The debate is whether the medical community is ready for software which links up diagnostic information about a patient to medical knowledge. Currently, a few companies make software which does this. One of the leading companies is PKC Corporation (http://www.pkc.com).

Founded in 1982, the PKC Corporation is a clinical software company located in Burlington, Vermont that builds knowledge engineering tools for the healthcare community. The PKC product set, known as Couplers (Problem Knowledge Couplers) facilitates the automatic capture of a complete set of relevant structured data on any patient or problem. These tools instantly match the patient data with a wide range of knowledge from the medical literature, protocols and guidelines from the user organization, and specific appropriate helpful resources from the user environment. The PKC KnowledgeNet database of patient specific clinical entities (signs, symptoms, research knowledge and the relationships between them) has been developed over a fourteen year period and is constantly being updated and expanded.

The PKC software tools are based on open architecture principles, the Windows NT environment, and application communication technologies such as ActiveX and formal Lexicon services.

In addition to the work in the commercial sector, PKC has developed a number of custom modules based upon PKC knowledge engineering technology and the PKC medical knowledge base for the U.S. Department of Defense. Managing these Department of Defense programs for health risk assessment, deployment monitoring, and healthcare delivery, is the responsibility of the Washington, D.C. area facility of PKC Corporation in Falls Church, VA. These offices are fully staffed to support installations throughout the country.

The President of PKC is Dr. Lawrence Weed. Dr. Weed is Professor of Medicine Emeritus at the College of Medicine, University of Vermont, in Burlington. He received his MD degree from Columbia University College of Physicians and Surgeons. He then went on to join the University of Vermont College Medicine faculty as Professor of Medicine in 1969 after holding positions at the University of Pennsylvania, Walter Reed Hospital, Johns Hopkins Hospital, the Yale University School of Medicine, Eastern Maine General Hospital, and Case Western Reserve.

From 1969 to 1982, while professor of Medicine at the University of Vermont, Dr. Weed was Principal Investigator for an NCHSR-funded research project, the computerized Problem-Oriented Medical Information System (PROMIS), a large minicomputer-based system implementing the problem oriented medical record on several wards of the Medical Center Hospital of Vermont.

Dr. Weed, in collaboration with Mr. Richard Hertzberg, shareholder and lead programmer for PKC, founded PKC Corporation in 1982 to develop information tools that address key needs in the practice of medicine through the development and use of the problem-oriented record.

What Are Problem Knowledge Couplers?

Just what is this software Weed has worked so long in developing and now advocates so strongly for use in the medical profession?

The PKC website offers the best explanation. As noted on the site, Problem Knowledge Couplers are data capture and clinical guidance software tools that provide decision and management support to clinicians in both Microsoft Windows® and Web-based environments. Couplers are designed to give the physician the tools necessary to help patients make the best-informed decisions about their own healthcare.

Couplers provide a method for identifying the unique features of the individual patient and coupling positive findings to possible causes and to different management options based upon medical literature. When properly utilized, Couplers automatically accomplish three separate and important tasks associated with any patient interaction with a healthcare system:

1. Establishment of a complete set of structured intelligent data or "findings" about the patient.
2. Matching of distinct patient findings against the best medical literature and automatically exposing the appropriate relevant knowledge for the specific patient.
3. Bringing the appropriate resources and guidelines from the user organization to bear on the identified patient problems.

As noted on the PKC site, the primary categories of Problem Knowledge Couplers include the following:

WELLNESS AND HEALTH HISTORY SCREENING COUPLERS for identification of risk factors, positive and negative health habits, behaviors, and attitudes with response providing patient- specific education on improvements for an individual's health

DIAGNOSTIC COUPLERS to solicit needed information about a patient's chief complaint so as to aid the physician in considering all appropriate diagnostic and test options

MANAGEMENT COUPLERS to work with a patient's discovered problem to present all therapy options, accompanied by relevant citations from the medical literature appropriate to this particular patient

CUSTOM COUPLERS are produced to most appropriately respond to local practice procedures and protocols, pointing to identified referrals, medical interventions, service choices, and preventive resources.

The full Coupler set represents the complete offering of PKC's screening, diagnostic and management tools. This includes approximately 70 couplers.

One example of a coupler is Set #1 which is identified as a Complete Health Screening, Diagnostic, Management, And Behavioral Health Tool Set. This set gives the physician's office the full compliment of tools to couple patient specific data to a broad and regularly updated medical knowledge base for:

• Collecting complete, structured, standardized patient data
• Organizing personal patient/provider strategies for health promotion
• Presenting unbiased and broad patient and problem specific knowledge always linked to supporting literature citation lists
• Automating and documenting detailed patient education and encounter transcription.

Wired Doctors?

The ideas of Dr. Weed about using knowledge couplers have caused a debate within the pages of the British Medical Journal (BMJ) but in the loud media noise of political scandals and tabloid news, little of this debate has come onto the "radar screen" of the general populace. The debate centers around a fascinating article in the July 26, 1997 BMJ called "New Connections Between Medical Knowledge And Patient Care."

The article (http://www.bmj.com/cgi/content/full/315/7102/231) is well-worth reading not only for the revolution in medicine it suggests but also for the application possibilities of application in other areas such as the internet and business.

The problem is stated succinctly at the beginning of the article. "The meteoric shower of medicine's scientific achievements can overwhelm a doctor's mind. A patient has no assurance that his or her doctor is able to take into account all relevant scientific knowledge and integrate it with detailed data about the patient's own condition. Yet few doctors, patients, or policy makers recognize that modern information tools can become the loom for weaving these two bodies of knowledge into a fabric. In fact, few recognize the dimensions of the problem."

The cause of this state of affairs, notes Dr. Weed, is "misplaced faith in the unaided human mind." The requirement of modern medical practice requires tools to extend the mind's limited capacity to recall and process large numbers of relevant variables, just as medical science requires the microscope to extend our capacity to see at the microscopic level. We must abandon the arrogance of professional 'expertise' that shuns such tools. Instead, we must use the new tools routinely as they are developed for more and more diagnostic and management problems."

Again, we suggest our readers link directly to the article and read it. For the purposes of brevity, though, here are the summary points of Dr. Weeds argument in the article:

• Medicine lacks an information infrastructure to efficiently connect those who produce and archive medical knowledge to those who must apply that knowledge
• There are serious "voltage drops" along the transmission line for medical knowledge in the present healthcare system
• Good medical practice requires tools to extend the human mind's limited capacity to recall and process large numbers of relevant variables
• Knowledge should be held in tools that are kept up to date and used routinely—not in heads, which are expensive to load and faulty in the retention and processing of knowledge
• Such information tools would allow fundamental changes in our approach to medical education
• The tools would also allow a defined and consistent approach for controlling and keeping track of inputs to the healthcare system, which in turn would enable outputs to be properly interpreted and corrective feedback loops to be used routinely.

At the conclusion of the article, Dr. Weed notes that "Retrieving and processing information in medicine are operations that we have tried to perform by thinking about them at the time of action under extraordinary time constraints. Enormous damage results from this misguided effort."

As Dr. Weed notes, new information tools enable us to leave behind a world of medical practice in which providers and patients alike are victims of the "predictable and undesirable internal constraints" of the unaided human mind. "Providers have been granted unjustified power to function in the face of avoidable ignorance of crucial details about medical science and about patients themselves. New tools can release patients and providers from this ignorance as they weave scientific knowledge into the fabric of patient care."

Challenges To Weed

The concepts suggested by Dr. Weed do not go unchallenged via a number of Letters To The Editor in the BMJ from other doctors. One doctor notes that the human condition is full of decisions that aren't simple yes/no decisions. Dr.Kernick from Exeter, England notes "The next generation may be driven by an information technology/designer label/ soap opera culture, but much of society will still need a shaman/healer/medicine man/health advocate/friend and will lose him at its peril." Dr. John Edwards notes that, "Patients in most need of medical attention are least able to operate computers" and that Dr. Weed, "would benefit from spending some time in a primary care doctor's practice."

Within the subtext of these concerns is the question of whether the Weed system will replace doctors or help doctors. Perhaps a mixture of both.

However, there is agreement by some to the ideas of Lawrence Weed. Dr. David Hutchon, an obstetrician and gynaecologist at Darlington Memorial Hospital in Durham, England feels information technology has much to offer certain areas of health care. "I believe that there is an enormous place for 'expert systems' in medicine," he writes, "and I see nothing wrong with cookbook medicine when the recipe is robust and effective. Unfortunately, patients are as fearful about relying on computers for medical decisions as doctors are reluctant. But we already rely on many expert systems, such as the algorithms involved in computed tomography and ultrasonography."

Dr. Hutchon see three key areas in which information technology and artificial intelligence can lead to major improvements in health care. "One is making medical knowledge available in an easily assimilated form. Although medical knowledge has always been freely available in modern times, it has been in the user unfriendly form of large textbooks and technical journals. Secondly, information systems could build research and audit into every act of health care. Thirdly, when an aspect of healthcare management becomes robust and effective, sometimes in a small and specific area, it could be handed over to an expert system, with technicians being used, for example, to palpate the spleen when necessary."

Web Surfers As Patients?

As the debate continues on in the medical community, the analogy to other areas might be overlooked. Certainly one area is the internet where web surfers might be compared to "patients" in many ways.

One of the most exciting technologies of the internet is the increasing power and "smartness" of net-based relational databases. These range from what one might term simple "static" relational databases where little data is input into form fields to complex "smart" relational databases which "learn" based on growing amounts of data in the form of web surf/click patterns.

Examples of the first "static" type of databases are real estate (Coldwell Banker at http://www.coldwellbanker.com), car (Microsoft's Carpoint at http://carpoint.msn.com) and travel (Microsoft's Expedia at http://expedia.msn.com). On these, specific information is entered to find homes, cars or travel packages. Examples of "smart" relational databases of the second type would be Firefly (http://www.firefly.com).

In effect, rather than using PKCs stand alone "knowledge coupling" software, the internet itself becomes the software. Rather than capture medical information, the net captures click information. Both are methods which serve to define the person and match what "fits" the needs of this person.

In this way, a patient with a particular medical question is in many ways similar to a surfer with a particular house, car or travel question.

Organizations As Patients?

One of the emerging trends of the late 20th century is to see the world in a new biological-based perspective. The evolutionary biology of E.O.Wilson at Harvard and his study of ants gets increasing "air play" in magazines such as the Atlantic. The concept extends beyond the little microscopic community of ants to encompass the entire economy in theories such as Michael Rothchild's concept of "bionomics" or the economy as an ecosystem expressed in his book Bionomics.

One of the major inroads of biological thinking into modern organization is in the work of MIT's Sloan School of Management professor Peter Senge and his Center for Organizational Learning (http://www.sol-ne.org). The center was founded in 1991 at Sloan in response to widespread interest in the concepts described in Senge's best-selling book, The Fifth Discipline. These concepts were developed and built on decades of research in system dynamics, group process, action science, and the creative process, as well as years of practical consulting and workshop experience.

As noted on the center's website, the purpose of The Society for Organizational Learning is to discover, integrate, and implement theories and practices for the interdependent development of people and their institutions. The guiding principles of SoL (also from their website) are the following:

• Drive to Learn - All human beings are born with an innate, lifelong desire and ability to learn, which should be enhanced by all organizations
• Learning is Social - People learn best from and with one another, and participation in learning communities is vital to their effectiveness, well-being and happiness in any work setting
• Learning Communities - The capacities and accomplishments of organizations are inseparable from, and dependent on, the capacities of the learning communities which they foster
• Aligning with Nature - It is essential that organizations evolve to be in greater harmony with human nature and with the natural world
• Core Learning Capabilities - Organizations must develop individual and collective capabilities to understand complex, interdependent issues; engage in reflective, generative conversation; and nurture personal and shared aspirations
• Cross-Organizational Collaboration - Learning communities that connect multiple organizations can significantly enhance the capacity for profound individual and organizational change.

The Real Links Are Between Disciplines

The emerging magic of relational databases on the internet, whether medical or consumer, can only go so far. They might be able to get smarter and smarter and match up more and more fields. But the real link is not between user defined fields but rather the "fields" within various disciplines.

The various disciplines today need to be viewed like a number of patients with their own set of problems all trying to link to a large database similar to a database in medicine.

In this sense, it might be beneficial for commercial technological companies like Lawrence Weed's PKC Corporation to link to academic groups like Peter Senge's Society for Organizational Learning. Or, for that matter, for management consulting firms to link with business schools.

Of course some of this is already being done but the effort is little more than a half-hearted one. In an era of increased product differentiation and branding, what is the economic motive in working towards similarities? In this sense, Peter Senge's ideas about learning organizations are important and interesting. But realistic in a "brand" saturated world?

At the conclusion of his article in the BMJ, Dr. Weed quotes philosopher Alfred North Whitehead. "Civilization advances by extending the number of important operations which we can perform without thinking about them."

Maybe medicine, business and connection in our new internet age advance like civilization. But before we rush ahead into this new territory, perhaps we should advance a little more thinking about all of this. It is an important area even though it quietly surfaces in the relatively obscure pages of a foreign medical journal. The implications, though, go far beyond medicine.

© 1998 - John Fraim

{/viewonly}