What's wrong with decision analysis? Can the left brain influence the right?

Abstract

In order to gain insight into the impact that decision analysis has had on clinical practice, we presented a published report on the utility of renal biopsy for patients with idiopathic nephrotic syndrome to a group of nephrologists and residents at a teaching hospital. Although the analysis showed that the decision to biopsy or use empiric steroids without biopsy was a toss-up in terms of patient outcomes, only one of six staff nephrologists was willing to forego the biopsy strategy. Many clinicians in the group discussed the pure value of the information (e.g in making statements about prognosis) derived from the biopsy as an important factor in the choice of clinical strategies, a characteristic which was not captured by the published analysis. Also, some clinicians were uncomfortable with the entire simulation process as there were no "real patients" in the study. It appeared that clinical intuition based on pattern recognition could not be influenced by the linear logic of decision analysis. We suggest that major challenges for decision analysts include incorporating the value of information into analyses, selecting and cultivating the most appropriate clinical audience, and demonstrating the benefits of decision analysis for either the decision-making process or health outcomes. Without meeting these challenges, decision analysis may remain an esoteric field within academic medicine, which will continue to have limited impact on clinical practice.