The Symptom Sleuths

Larry Weed created a remote diagnostic program so effective that the military licenses everything his company produces

by Rosalyn Graham

Fifty years ago, a young Dr. Lawrence Weed, not long out of medical school, had an “aha!” moment that sowed the seeds of PKC Corp., the company he eventually founded, with headquarters in the Chace Mill and an office in Washington, D.C.

There is a spot of irony in the fact that a man who seems to have a prodigious capacity to instantly recall dates, names, conversations, symptoms, diagnoses, poems by Edna St. Vincent Millay and quotations from Plato and Karl Marx should also be the one who found a way to bolster the memory and therefore the diagnostic ability of the medical establishment.

Dr. Lawrence Weed is the founder and owner of PKC, a business in the Chace Mill in Burlington that produces a microcomputer-based medical software tool for clinical decision-making. 

The seeds for the business were sown in Weed’s fertile mind in the early 1950s, soon after he had graduated from Columbia Medical School and was doing research at Yale University school of medicine. 

Having been invited to go along on medical rounds at New Haven Hospital, he realized that the doctors, students and interns were facing an almost impossible task. 

Unlike a researcher who could focus on one or maybe two problems at one time, the hospital physicians were seeing not a simple problem but five or six problems affecting one body. One patient might be suffering everything from hypertension to marital difficulties. 

“A research scientist is expected to focus on one or two problems, but a doctor has a new problem or set of problems walk in the door every 20 minutes,” he says. 

The other thing he discovered was that there was no complete record of the patient’s problems, no evidence to back up the diagnosis that was scribbled on the patient’s chart. 

“Talk about your eureka moment!” he exclaims. “Here are the doctors’ notes, the nurses’ notes, the lab notes, all the raw material for a computerized record system.” 

It was a system that became Weed’s first step toward revolutionizing the way patient information is managed.

An invitation to go to Bangor, Maine, as director of medical education in 1956 gave Weed the ideal situation to put his Problem-Oriented Medical Record (POMR) into practice. 

“I said, “For every patient when I walk in, I should be able to see all the patient’s problems — social, psychological, physical. If problem number one is pneumonia, but problem number five is no heat in the house, we can’t fix number one without fixing number five.”

PKC is an acronym for “problem-knowledge coupler,” the name given to the software that does the work. Twenty-five of PKC’s 62 employees are medical content authors. Tim Pay is assistant director for administration of medical content, and Theresa Martell is senior medical content developer.

The next step was a realization that the physician’s search for the right diagnosis was virtually impossible without a system to organize all the available information that could influence the diagnosis. 

What was needed was “an accounting system,” Weed says, for keeping track of all the possible causes for a particular symptom. “Chest pain could have 100 different causes, and the doctor can’t keep score in his head.” 

Weed’s crusade for better records and better organization of the information led to the Problem-Oriented Medical Information System (PROMIS) project, which he implemented while he was a professor at the College of Medicine at the University of Vermont. 

He was seeing a new problem, though. “When you solve the information problem,” he says, “you create a processing problem. It’s like an open-book exam — you have all the information available, but finding it is almost impossible.” 

Then fate, technology and serendipity stepped in. Richard Hertzberg was a graduate student in computer science at a time when, as he says, “Computers were as big as a room and were seen as soul-deadening devices. People who liked to play with them had to fight that.” 

Hertzberg had gone to medical school, but had decided that computer science was a better place for him. He liked the idea that working with Weed gave him a satisfying chance to use computers to help people. He was hired to work on the computer implementation of the PROMIS project.

“When microcomputers came on the scene, Larry had a brainstorm,” Hertzberg remembers. Just as the telescope and microscope and X-ray were invented to extend the capability of the eyes, and the automobile supplemented the muscles, Weed said, why not use the computer to sort and process all the knowledge that a physician should have, but that is beyond the brain’s capacity to manage. 

The result was what they called a “knowledge coupler” — they would put the software in the computer, it would tell the physician what questions to ask for a particular symptom, and then it would “couple” the answers to the knowledge that had been stored in its brain and present possible diagnoses, from most probable to least probable.

“Larry and I thought this knowledge coupler could be cheap,” Hertzberg says. “It could go in small clinicians’ offices, not big hospitals. We founded PKC with that idea in mind.” The company incorporated in 1982.

Hertzberg laughs when he remembers the early years, especially when he compares their operation to Richard Tarrant’s IDX start-up in a second-floor garret. “We didn’t even have a second-floor garret,” he says. 

They had bought one Apple computer for the astronomical sum of $4,500 (in 1980 dollars!) and they hauled the computer back and forth between Hertzberg’s Ferrisburgh home and Weed’s home in Underhill. “I’d program for a couple of days and then take it to Larry, and he and his wife, Laura, who was also a physician, would create content for a few days and then it was back to me to embellish the program.” 

For 10 years, the Weeds, their son, Chris, and Hertzberg continued to refine their knowledge coupler and market it to a world of professionals who were vested in the existing protocols and were resistant to a tool, however logical, that was contrary to the tradition of doctors who saw patients through the lens of their specialties. 

As Weed says, “If you want to introduce airplanes, you don’t go to people who own the railroads.”

 “We are eggheads,” says Hertzberg, “with no business savvy, so even though we got lots of great press for our novel idea, we didn’t achieve any mass impact on the market until about 1992 when Howard Pierce came on board.” 

Weed says, “I knew we had a product and a philosophy. All systems are made up of tools, disciplined users, the right philosophy and leadership. I said to Richard that we’ll focus on the philosophy and the tools; we can’t try to do all four. 

Richard, whose title is director of software development, is an unbelievable tool builder, and Howard made the contact with the business community, doing so many things like having meetings in Montpelier about cost of health care.”

Pierce brought 25 years of business experience running Juba Design, an architectural and construction company in Vergennes, to the tiny, four-person business. “Larry showed my wife, Wendy, who is a nurse practitioner, what he was doing, and I got so interested I agreed to help with a couple of business things. Then he asked me if I would turn PKC into a larger organization.” Pierce is chief executive officer.

Although the original assumption was that doctors and hospitals would be the logical users of the knowledge coupler technology, the Department of Defense has proved to be the most enthusiastic customer. The program is also available on the company’s Web site for employees of participating businesses and their employees. Bill Tobey is director of medical content.

The first big leap, in the mid ’90s, according to Pierce, was the work with the Department of Defense on Persian Gulf syndrome, looking for “something” that was affecting many returning military people. “That led to their licensing everything we build.” 

Military use of the software might allow a trained medical corpsman on a submarine to ask the questions used in the information-gathering part of the process, enter the answers in to the computer, and send the results by satellite to a doctor in Bethesda. The doctor would use the diagnosis tool to decide whether a million-dollar decision must be made to return the submarine to shore or the situation could be handled by the corpsman. 

Today, stepping through the door of their Chace Mill headquarters — there is also a six-person office in Washington — is to enter a multi-floor network of offices for 62 people. Twenty employees are medical content authors who spend their days gleaning the most up-to-the-minute medical information from textbooks, journals and specialized books and articles to add to the vast knowledge base that is the basis of the knowledge coupler. 

“Everybody else is here to support them in building and maintaining the couplers,” Pierce explains. 

“We have software engineers who build software, both what our customers see and what we need to build and maintain the couplers; and then we have client support people and technical infrastructure people and project management people who work with clients like the Department of Defense,” says Pierce.

Pierce says that one of the most important things in the last five years is the possibility of turning the company toward the health care consumer, licensing its knowledge coupler tools for the employees of participating companies to use before visiting their doctors. 

“It’s been a great success,” Pierce says, “especially our work with the Vermont Business Roundtable to make it available to Vermont employers. We sent out a letter to businesses, and they are taking our tools.”

Participants include SymQuest, Green Mountain Power, Burton Snowboards, Engelberth Construction, Champlain College and Vermont Teddy Bear. “We’ve priced the service so that the smallest companies can afford to do it,” says Pierce. 

A new client is taking the tools onto a national stage. Informed (pronounced “IN-for-med”), a business recently started by IDX co-founder Richard Tarrant, will provide nurse practitioners equipped with PKC knowledge couplers as health care providers in large and small businesses.

Although they are exploring new directions and new markets, the mission remains the same, according to Pierce. “We are advocating at every chance for reform of the entire operation of the medical establishment.” 

Traditional medicine divides the body into parts, and doctors look at symptoms based on their own areas of specialty. “Instead of a heart specialist door or an orthopedic door,” says Pierce, “there should be a chest pain door or a headache door. We’ve gotten into a terrible mess, and yet when you talk to the layman, it sounds like common sense.”

It’s the same problem of meeting the challenge of asking a human brain to assimilate and apply too vast a body of knowledge that Larry Weed identified as a young doctor and researcher in the ’50s, but the commitment to fixing the problem continues today. His colleagues recognize the power and vision of the PKC founder and president, an energetic octogenarian who doesn’t like to talk about his personal life and doesn’t tell his age, although he freely admits that it is 60 years since he graduated from medical school. 

Pierce says, “Larry keeps everybody focused on the mission. He’s charismatic and he makes a good case.” Hertzberg says, “When all have lost energy, Larry has kept his; he’s our motive force.” •