C-4 UNDERSTANDING WHY SHARED DECISION MAKING MAY DECREASE UTILIZATION

Monday, October 20, 2008: 2:15 PM
Grand Ballroom D (Hyatt Regency Penns Landing)
Liana Fraenkel, MD, MPH, Yale University, New Haven, CT, Ellen Peters, Ph.D., Decision Research, Eugene, OR and Paul R. Falzer, PhD, Yale School of Medicine, West Haven, CT
Purpose: Some trials have found that shared decision making (SDM) tools decrease health services utilization. These changes are believed to result from a shift towards treatment plans which are more concordant with patients’ preferences. An alternative explanation is suggested by a recent study describing that volition (perceived control over engaging in a risky behavior) increases risk perception. The objective of this study was to examine whether the SDM model might inadvertently decrease utilization by emphasizing volition and consequently increasing risk perception.

Methods: We presented patients with a video describing the availability of a new medication. The new medication was described as being very effective (for a patient-specific condition) covered by insurance, and without toxicity except for the risk of an extremely rare, but dreaded adverse effect. After watching the video, we examined risk perception (willingness to take the medication, worry, perceived chance, anticipated responsibility and regret on 11-point numeric scales) under two conditions: 1) Physician makes the decision and writes the prescription 2) Physician leaves the decision up to the patient. Differences between the two conditions were dichotomized. Pilot testing demonstrated no ordering effects. The data were analyzed using logistic regression analysis. We report goodness of fit coefficients derived from Negelkerke’s pseudo R2 to indicate the percent of variance in patient willingness that is accounted for by each predictor, after age, gender, education, and health status are taken into account.

Results: 139 subjects were recruited; mean age 59 (range 21-88); 62% female; 76% Caucasian; 40% college graduates. Willingness to take the medication was significantly lower in the patient scenario compared to the physician scenario (mean willingness = 4.0 and 5.1 respectively, p<0.001). Anticipated responsibility was significantly higher in the patient scenario (p<0.001). A substantial proportion of the variance in willingness was explained by increased worry, perceived chance, anticipated regret, and responsibility (Table 1).

Conclusion: Patients are less willing to accept treatment and significantly more concerned about the risk of toxicity when asked to assume responsibility for making a decision. These findings highlight an important potential unintended consequence of promoting patient participation in medical decision making, and may explain why SDM interventions result in decreased utilization of risky interventions. Table 1. Variance in Patient Willingness

R2

Worry

0.456

Perceived chance

0.199

Anticipated regret

0.147

Anticipated responsibility

0.368