Stability of patients' preferences for chemotherapy: The impact of experience

Background. Studies have shown that utilities for a particular treatment, e licited by means of a hypothetical treatment scenario, may remain stable wi thin the same patients when examined before, during, and after experiencing that treatment (within-group stability). However, other studies have found that utilities for a particular health state may differ between patient gr oups who are and who are not experiencing the particular health state (betw een group differences). Objective. The authors evaluated this apparent cont radiction in the case of adjuvant chemotherapy for breast cancer. A related purpose was to examine whether a chemotherapy scenario adequately reflects the patients' own experiences with chemotherapy. Method. Forty-three patie nts with early-stage breast cancer evaluated their actually experienced hea lth state and a chemotherapy scenario before, during, and after undergoing adjuvant chemotherapy (chemotherapy group). A control group of 51 patients for whom chemotherapy was not part of the treatment plan was interviewed at similar points in time. Utlities were elicited by means of a visual analog scale (VAS), a chained time trade-off (TTO), and a chained standard gamble (SG). Results. The utilities for the chemotherapy scenario remained relati vely stable over time in the 2 patient groups. Furthermore, the chemotherap y scenario was evaluated more positively by patients in the chemotherapy gr oup than by control patients (e.g., utilities before chemotherapy: VAS 0.69 vs. 0.50, TTO 0.88 vs. 0.50, SG 0.92 vs. 0.58, all Ps < 0.01). Finally, pa tients in the chemotherapy group evaluated their actually experienced healt h states during chemotherapy higher than the chemotherapy scenario that was assessed at the same time (VAS 0.79 vs. 0.69, TTO 0.93 vs. 0.87, SG 0.97 v s. 0.96, all Ps < 0.05). Conclusions. Both within-group stability and betwe en-group differences were found. A possible explanation for within-group st ability may be that the chemotherapy scenario did not fully correspond to t he patients' actual experiences with chemotherapy ("noncorresponding descri ption"). Therefore, preferences did not change even when the patients' own clinical health status had changed. The between-group differences may be ex plained by "anticipated adaptation." Both explanations may work together to explain why utilities remain stable within the same patients but differ be tween different patient groups.