Assessment of Unconscious Decision Aids Applied to Complex Patient-Centered Medical Decisions

Abstract

"Background: To improve patient health, recent research urges for medical decision aids that are designed to enhance the effectiveness of specific medically related decisions. Many such decisions involve complex information, and decision aids that independently use deliberative (analytical and slower) or intuitive (more affective and automatic) cognitive processes for such decisions result in suboptimal decisions. Unconscious thought can arguably use both intuitive and deliberative (slow and analytic) processes, and this combination may further benefit complex patient (or practitioner) decisions as medical decision aids. Indeed, mounting research demonstrates that individuals render better decisions generally if they are distracted from thinking consciously about complex information after it is presented (but can think unconsciously), relative to thinking about that information consciously or not at all. Objective: The current research tested whether the benefits of unconscious thought processes can be replicated using an Internet platform for a patient medical decision involving complex information. This research also explored the possibility that judgments reported after a period of unconscious thought are actually the result of a short period of conscious deliberation occurring during the decision report phase. Methods: A total of 173 participants in a Web-based experiment received information about four medical treatments, the best (worst) associated with mostly positive (negative) side-effects/attributes and the others with equal positive-negative ratios. Next, participants were either distracted for 3 minutes (unconscious thought), instructed to think about the information for 3 minutes (conscious thought), or moved directly to the decision task (immediate decision). Finally, participants reported their choice of, and attitudes toward, the treatments while experiencing high, low, or no cognitive load, which varied their ability to think consciously while reporting judgments. Cognitive load was manipulated by having participants memorize semi-random (high), line structured (low), or no dot patterns and recall these intermittently with their decision reports. Overall then, participants were randomly assigned to the conditions of a 3 (thought condition) by 3 (cognitive-load level) between-subjects design. Results: A logistic regression analysis indicated that the odds of participants choosing the best treatment were 2.25 times higher in the unconscious-thought condition compared to the immediate-decision condition (b=.81, Wald=4.32, P=.04, 95% CI 1.048-4.836), and 2.39 times greater compared to the conscious-thought condition (b=.87, Wald=4.87, P=.027, 95% CI 1.103-5.186). No difference was observed between the conscious-thought condition compared to the immediate-decision condition, and cognitive load manipulations did not affect choices or alter the above finding. Conclusions: This research demonstrates a plausible benefit of unconscious thinking as a decision aid for complex medical decisions, and represents the first use of unconscious thought processes as a patient-centered medical decision aid. Further, the quality of decisions reached unconsciously does not appear to be affected by the amount of cognitive load participants experienced.