Context Matters: Emotional Sensitivity to Probabilities and the Bias for Action in Cancer Treatment Decisions

Abstract

Background: Past studies have shown a commission bias for cancer treatment, a tendency to choose active treatment even when watchful waiting is less risky. This bias suggests motivations for action beyond mortality statistics, but recent evidence suggests that individuals differ in their emotional sensitivity to probabilities (ESP), the tendency to calibrate emotional reactions to probability. The current study aims to examine the role of ESP in the commission bias, specifically whether those higher in ESP are more likely to choose watchful waiting when risk probabilities align with that choice.

Methods: Participants (N = 1,055) read a scenario describing a hypothetical cancer diagnosis and chose between surgery and watchful waiting, with random assignment between versions where the mortality rate was either lower for surgery or for watchful waiting. We modeled choice using the Possibility Probability Questionnaire (PPQ), a measure of ESP, and several other individual differences in a logistic regression.

Results: We observed a commission bias as in past studies with most participants choosing surgery both when surgery was optimal (71%) and when watchful waiting was optimal (58%). An ESP × Condition interaction indicated that the predictive role of ESP depended on condition. Those higher in ESP were more likely to choose surgery when probabilities favored surgery, β = 0.57, P < 0.001, but when probabilities favored watchful waiting, ESP had a near-zero relationship with choice, β = 0.05, P < 0.99.

Conclusions: The role of ESP in decision making is context specific. Higher levels of ESP predict choosing action when that action is warranted but do not predict a shift away from surgery when watchful waiting offers better chances of survival. ESP does not overcome the commission bias.

Highlights: Past studies have identified a "commission bias," a tendency to choose active treatment over watchful waiting, even when mortality rate is lower for waiting.Evaluation of risk probabilities is related to individual differences in emotional sensitivity to probabilities (ESP) and has been shown to predict reactions to and decisions about health risk situations.ESP appears to be selectively factored into decision making. ESP was a robust predictor of choosing surgery when probability information supported surgery but did not predict decisions when probability information supported watchful waiting.Those who are most emotionally attuned to probabilities are just as susceptible to the commission bias as those who are less attuned.

Keywords: action bias; cancer; commission bias; decision-making; risk perception; sensitivity to probabilities.

Figure 1. Relationship of PPQ-ESP scores and probability of electing surgery for surgery optimal and watchful waiting optimal conditions.

This figure shows the change in the expected probability of electing surgery and accompanying 95% confidence intervals of these estimates as the observed PPQ-ESP scores increase. The plot shows the middle 95% of the empirical cumulative distribution (from the 2.5th to the 97.5th percentile) defined separately for each condition, and as expected, it falls within about 2 standard deviations above and below the PPQ-ESP mean (centered at 0). All other individual difference measures are fixed at their respective means. The points at the probabilities of 0 and 1 are the observed distribution of treatment choice. While log odds are linearly related to PPQ-ESP, probabilities are not because they are bounded, resulting in curvature and asymmetric confidence intervals. Practically, this means that the change in probability per unit change in PPQ-ESP is not constant across the PPQ-ESP level. Together, the 2 panels illustrate the pronounced difference in the relationship between PPQ-ESP and the probability of electing surgery for the 2 different conditions. When surgery has the lowest mortality rate and is considered optimal, a 1-standard deviation increase in PPQ-ESP relative to the mean adds 0.10 to the probability of electing surgery. However, when waiting is optimal and has the lowest mortality rate, 0.01 is added to the probability of surgery, which is now the nonoptimal option with the higher mortality rate. ESP, emotional sensitivity to probabilities; PPQ, Possibility Probability Questionnaire.