Vivid bad outcome influences the decisions of older adults about treatment timing: a randomized field experiment with an abdominal aortic aneurysm analog

In balancing the risk of rupture from an asymptomatic abdominal aortic aneurysm (AAA) against the risk of perioperative mortality, data-based guidelines recommend surgical repair when the AAA diameter reaches 5.5 cm, whereas smaller AAAs should be followed with periodic surveillance. Previous work with vascular surgeon subjects and a computer-based AAA analog simulation showed that, even when constantly updated with the relevant statistics, experiencing a prior bad watchful waiting outcome shortened the time until they made the decision to operate. Using the same simulation, this field experiment enrolled healthy older volunteers (n = 107). Participants were randomly assigned to experience either a bad outcome demonstration with an expanding balloon that bursts (experimental) or an expanding, nonbursting balloon (control). Participants then made decisions about how many times to allow the balloon to expand before opting-out of the simulation. The main outcome measure was the amount of time participants continued watchful waiting before opting-out. A Cox-regression analysis assessed the likelihood of opting-out after each expansion while controlling for censoring and important covariates, including baseline anxiety, uncertainty attitudes, and risk preferences. The bad outcome demonstration group ended the simulation significantly earlier than did the control subjects (Hazard ratio: 1.98; 95% CI: 1.05–3.74). These results extend previous findings from vascular surgeons to older adults at higher risk for AAA. The preceding bad outcome influenced subsequent decisions, even when statistical risk information was readily available. The influence of recent experience on medical decision making by patients with life-threatening conditions may be under-appreciated.