Biases in medical decision-making: a cross-medication comparison

In this paper, we investigate whether cognitive biases in medical decision-making differ across types of medications when objective risks of side effects are held constant. Using data from a survey and a stated-choice experiment, we compare hypothetical medication-taking responses across four medication choices, including vaccines and therapeutic interventions, and four combinations of trials and side effects. Our main findings suggest that individuals are generally rational and prefer medications with lower risks, but responses to risk information differ systematically by medication type. In particular, individuals are more susceptible to salient side-effect information, especially for vaccines, even when overall risk levels are identical. Examining individual-level sources of variation, we find that many of these vaccine-specific distortions are substantially reduced once we account for vaccination hesitancy and illness-related anxiety, while other correlated individual characteristics also play an important role in explaining heterogeneity in medication-taking behaviour.