Why Are They Wrong? Exploring Surrogates’ Accuracy When Predicting Patient Treatment Preferences

Abstract

Surrogate decision making is common in medical settings, particularly regarding treatment decisions for patients at end-of-life. When making decisions on behalf of patients, surrogates are most often encouraged to use the substituted judgment standard and make the decision that the patient would make if he or she were able to express a choice. A significant body of research, using patient-surrogate pairs and hypothetical vignettes, has documented that surrogates tend to inaccurately predict patient preferences when making these decisions (Shalowitz et al., 2006). This results in treatment decisions that do not match what the patient would have selected. Various factors that may influence surrogates’ predictive accuracy have been proposed, but no empirical examination of the relation between such factors and accuracy has yet been conducted. This exploratory study investigated whether certain intrapersonal and interpersonal variables predicted surrogates’ accuracy, using patient-surrogate dyads and an online survey methodology employing hypothetical treatment decision vignettes. One-hundred-and-fifty-four adults over age 45 were recruited from Amazon Mechanical Turk and the West Virginia University faculty and alumni listserv to complete an online survey as the “patient” and indicate the treatment option they would prefer in a variety of treatment decision scenarios. Patients then shared a link to a similar online survey with an adult from their life who would likely serve as their surrogate decision maker and make decisions on their behalf in the event of incapacitation. Therefore, 154 patient-surrogate pairs, or 308 total participants, were recruited. Surrogates completed the decision scenarios using the substituted judgment standard, indicating the treatment preference that they thought the patient would prefer. All participants completed self-report measures of the selected interpersonal and intrapersonal variables. Patients and surrogates were compensated with an Amazon gift card for their participation. Consistent with prior research, the current study found that surrogates are generally poor predictors of patient preferences when selecting interventions on decision vignettes. In terms of intrapersonal variables, surrogates’ accuracy when predicting preferences was positively correlated with levels of empathy and religiosity. Surrogates’ levels of trust in the medical profession were also positively correlated with accuracy. When considering their partners’ values, surrogates who reported greater perceived similarity to their patients regarding end-of-life values and who perceived their partners as more reluctant to burden others tended to be more accurate. Pairs who reported high discrepancies in perceived similarity on end-of-life values (e.g., the patient reported low similarity while their surrogate reported high similarity, and vice versa)