Clinical decision support systems come of age at 20

By John Andrews
12:00 AM

While technology has changed dramatically since the birth of the clinical decision support concept 20 years ago, its underlying principles – designed to improve patient care – have not, say observers who have followed its evolution. Thus, hospital IT systems have grown to the point where physicians now have the technological capabilities to reach goals espoused back in 1987.

Stanford professor Edward Shortliffe, MD, PhD, published a clinical decision support blueprint in the July 3, 1987 edition of Journal of the American Medical Association, defining it as “any computer program designed to help health professionals make clinical decisions.” Shortliffe contended that “regardless of the underlying computer science methodology, clinical decision support systems deal with medical data about patients or the knowledge of medicine necessary to interpret such data. They generally help in determining the nature of a patient’s disease state or formulating a plan for reaching a diagnosis or administering therapy.”

In 1990, University of Utah biomedical informatics expert T. Allan Pryor, PhD, added an ascending levels to the foundation that includes alerting, critiquing, interpreting, assisting, diagnosing and managing decision-making. By and large, clinical decision support systems have followed this template as they’ve evolved during the past two decades, says Stan Pestotnik, CEO of Salt Lake City, Utah-based TheraDoc.

“After the foundation was laid, computer technology progressed, programming languages improved and it became possible to create systems based upon these principles,” he said.

The issue comes down to how well a system supports the physician’s clinical approach through patient data access, software vendors say. The more reliable the data, the speedier the delivery and the easier it is to disseminate, the better the chance of an accurate diagnosis.

Software vendors tend to be empathetic with physician methodology, and several top executives have medical degrees. The “D2” in D2Hawkeye, for instance, stands for “doctor-driven.” Chris Kryder, MD, founded the Waltham, Mass.-based company to address what he calls “a fundamental problem” of information tracking between providers and payers.

“The clinical side of healthcare drives 80 to 90 percent of costs – doctors don’t understand the business side and payers don’t understand medicine,” Kryder said. “There is plenty of information buried in existing data streams that could be re-interpreted and re-oriented into databases accessed by both providers and payers. We wanted to develop an easy-to-use product that could demystify many of the quality and cost problems that exist.”

Tom Doerr, MD, chairman and chief medical officer of St. Louis-based Purkinje, says the company’s CareSeries product delivers evidence-based decision support that is “often admired, but seldom practiced.” Historically, he said, the process has been laborious for physicians to conduct.

“The physician will see a patient, realize there is a deficit in knowledge, go to the library to research it and find an answer,” Doerr said. “It is a process that is ridiculously impractical.”

CareSeries’ computerized decision support draws from electronic medical records and runs a rules engine that makes recommendations in the vein of “if the patient has x condition, then do y.” Purkinje has a team of clinicians working full time on the rules engine.

Data continuity between points of care is a focus for New York-based ActiveHealth, which has been working with regional health information organizations on improving electronic connectivity across the continuum, said CEO Lonny Reisman, MD.

“In today’s fragmented healthcare system, patient data is scattered among physicians, hospitals, lab companies and pharmacies,” he said. “This fragmentation is at the heart of escalating healthcare costs, medical errors and avoidable hospitalizations.”

The company’s CareEngine system aims to analyze clinical data in real time instead of in intermittent batches.

‘Virtual’ consulting team

The Isabel diagnosis support system is named for co-founder Jason Maude’s daughter, who as a young child nearly died from a misdiagnosis. Based in London and Reston, Va., Isabel seeks to provide a “virtual consulting team” to physicians by giving them access to exhaustive volumes of research while at the same time providing guidance on clinical reasoning.

Drawing from references in Jerome Groopman’s book How Doctors Think, the Isabel system attempts to steer users away from mental traps like “diagnosis momentum,” in which complications may get overlooked as a physician follows a course of symptoms. This guidance is necessary, Maude said, because even as technology has progressed, the mentality behind misdiagnosis has not.

“A 2005 survey showed that one in six people have experienced a misdiagnosis either as a patient or family member,” he said.

The Carle Foundation in Urbana, Ill. deployed the Isabel system a few months ago, and it has changed the way the medical staff think about diagnoses, said Napoleon Knight, MD, an emergency physician and vice president of medical affairs and quality at the facility.

“What’s really interesting about it is that sometimes when a patient tells you something that raises suspicions, it shows you how to dig deeper into what possibilities exist,” he said. “It makes you ask more probing questions to see if there is a path you need to follow.”