OBJECTIVES: To explore whether risk tolerance changes when varying the order in which multiple risks are presented in a threshold technique (TT) survey.

METHODS: Adult HCC patients completed a web-based TT survey in which they were first asked to choose between two therapies that differed only by mode of administration (oral pills [tablets] or intravenous [IV] infusion and then presented with seven threshold series exploring their risk tolerance for seven clinically relevant adverse events (AEs): hypertension, decreased appetite, hand-foot skin reaction, diarrhea, ascites, proteinuria, and peripheral edema. The threshold series were presented in random order to each respondent, depending on their initial choice of preferred mode. The average maximum risk respondents would accept for each risk was estimated using interval regressions. Dummy-coded variables, indicating whether the risk was seen first or last in the threshold series, also were included in the regressions.

RESULTS: The survey was completed by 150 patients with a physician-confirmed HCC diagnosis. The regressions did not consistently show evidence of significant differences in risk tolerance based on whether a respondent was shown a risk first or last in the threshold exercises. In the TT series where tablet was preferred, seeing a risk first was only significant (p<0.05) for one risk, and seeing a risk last was only significant for two risks (all coefficients decreased average maximum risk). In the IV TT series, seeing a risk first was only significant for one risk and it increased the average maximum risk.

CONCLUSIONS: When multiple risks are presented in a TT survey, it is good practice to randomize the order in which the risks are presented. We did not find evidence of consistent differences in risk tolerance that depended on the order a risk was presented. A specifically designed experiment is needed to confirm our empirical finding.