Computerizing clinical decision support: lessons

Web-enabled clinical support software developed for oncologists provides sophisticated rules-based treatment recommendations based on established oncology clinical guidelines, according to a speaker at the HIMSS conference this morning.

"Rising costs coupled with declining reimbursements and demands for greater efficiency in healthcare are all driving the development of the system," said Dr. Mitchell Morris, senior vice president and chief information officer for the M.D. Anderson Cancer Center (MDACC) at the University of Texas, Houston. "Information technology is seen as a solution, although we are still in the hunter-gatherer stage."

Working with partner iKnowMed, a Palo Alto, CA, software developing company, the MDACC team set out to design a decision support tool that matched patients to evidence-based therapy, while at the same time capturing outcomes information to assess compliance and treatment effectiveness.

The system called KnowChart Navigator faced many early obstacles, among them overcome were physician concerns. Some clinicians fear that what they perceive as "cookbook medicine" might reduce their independence and medical creativity. Many feel no urgency to change, Morris said. They believe their patients' problems are too complex to justify reliance on a software program, and they fear their position of patient advocate could be negatively affected.

Nevertheless, initial results after going live with KnowChart Navigator in May, 1999, at the M.D. Anderson facility in Orlando indicate that compliance with guidelines hit 78% out of a targeted 80%. Physician tend to feel more amiable toward the system once they realize these are guidelines, not rules, Morris said.

So far, guidelines for most major cancer diagnoses are complete and more than 70 care paths have been implemented for the nearly 25,000 patients treated. Case managers are now coordinating care, using the Internet delivery feature of the program.

"We're seeing decreased length of stay with no extra complications or decrease in satisfaction," Morris said. "We're also seeing a reduction in resource utilization."

The following figures, for instance, came from a gynecology patient study:

- median length of stay decreased from six days to four days;
- laboratory costs decreased by 74%;
- medication costs dropped by 35%,;and
- total hospital costs were reduced by 20%.

"There have been no readmissions for complications, and patient satisfaction remained high," Morris said.