Directed by Yakov Ben-Haim, PhD, and Clifford C. Dacso, MD

This program is investigating a novel approach to decision analysis in clinical practice, and work in this area affects every other Center program. Info-gap decision theory is designed to help patients and physicians transform population-based data into personally relevant treatment answers.

Most clinical decision making is probability-based. For example, a patient might be told that a treatment has a 40% chance of success. However, because no one knows how well the patient matches the study population that was used to arrive at those odds, clinical probabilities always fall short.

A better approach to medical decision making deals directly with the gaps in the data, bridging those gaps and fitting the best data available to the patient’s needs, rather than “playing the odds.” Particularly when information is sparse, this strategy ensures that the patient and physician can have confidence in their decisions as the right choices for this individual patient, rather than aiming for an unachievable optimum goal.

Info-gap is grounded in the Nobel prizewinning work of Herbert Simon, in which he identified the value of “satisficing” in improving the quality of a choice. It allows the user to operate in conditions of severe uncertainty—such as is found in most clinical situations—and to make decisions that not only match patients to the data that are relevant to their individual needs, but also are open to unexpected best outcomes, something that “playing the odds” cannot do.

Info-gap excels in conditions where the data are sparse, uncertain, and the result of failure is catastrophe. Standard clinical decision making strategies do not protect patients against errors in the clinical data. However, because info-gap is not probability-based, it provides a strong foundation for making decisions when there is not enough information—such as in the highly individualized world of the clinic. Info-gap also allows “real-time” decision making and rapid assessment, thus allowing for targeted answers without wasted time.

Thus far, Center researchers have modeled five medical conditions and developed prototype applications for three of them. A Web-based tool that allows patients, families, and physicians to explore different treatment options and create their own customized treatment plans is entering the beta-testing phase.

September 11 through 15, 2006, the Center hosted the second international workshop in info-gap applications.