A Time Trade-Off Method to Elicit Preferences of Spouses About the Influnce of Their Partners' Potential Prostate Cancer Health States on Them

Background: Medical cost-effectiveness analyses (CEA) traditionally treat patients as isolated individuals and neglect the effects of improvement in patients' health on the welfare of their family members. Recent analyses of the theoretical foundations of CEA suggest that cost-effectiveness analyses may better reflect the full costs and benefits of medical interventions if they incorporate these family effects. However, there is dearth of methods for measuring these spillover effects using preference-based approaches. Objective: To develop and apply a time trade-off (TTO) technique to measure the impact on the quality of life of spouses due to potential prostate cancer-related health states of their partners. Methods: In an ongoing study, we used standard TTO questions to measure utilities of male patients at the biopsy clinic who are at risk of developing prostate cancer for various prostate cancer health states. In an add-on pilot study, we developed modified TTO questions that ask a spouse of a patient to trade-off her own life in order to reduce the burden that she herself expects to experience if her partner develops one of the prostate cancer related health states. We review the theoretical justification for this question and carefully frame the question so as to reduce measurement error and also to avoid possible double counting with the effect on the partner's health. Results: We collected data from 26 spouses about their preferences for their own health, which is influenced by the patient's health condition. We find that the average quality of life weights for the spouses when the patients are healthy, impotent, incontinent, metastatic or dead are 0.81, 0.66, 0.68, 0.50 and 0.28 respectively. We find evidence of good face, convergent, divergent and concurrent validity with these time-trade-off questions. Furthermore, we matched the preferences of the spouses with those of the patients on the same health states and found that these preferences were either uncorrelated or negatively correlated, implying that the spouses were not acting as proxies for the patient. Conclusions: Preliminary results show that this new time-trade-off method produces valid measurements of the spillover effect on family member's QOL