Examination of Assumptions in Using Time Tradeoff and Standard Gamble Utilities in Individuals With Spinal Cord Injury

Abstract

Lin M-R, Yu W-Y, Wang S-C. Examination of assumptions in using time tradeoff and standard gamble utilities in individuals with spinal cord injury. ObjectiveTo examine how the time tradeoff (TTO) and standard gamble (SG) utilities perform at different lengths of life expectancy and across patient characteristics such as risk attitudes and injury severity in individuals with traumatic spinal cord injury (SCI). DesignSurvey. SettingStructured telephone interviews of patients after discharge from 4 teaching hospitals. ParticipantsSubjects (N=270) who had sustained SCI were block-randomized in groups of 4 to receive 1 of 2 questionnaire versions. One version asked about the TTO and SG at 2 life expectancies of 10 and 20 years, and the other at 20 and 30 years. InterventionsNot applicable. Main Outcome MeasuresThe TTO and SG were used to elicit preferences for patients with an SCI. ResultsPatients with SCI who were younger and had more severe neurologic lesions, had higher intense risk-taking attitudes, and experienced depression had significantly lower TTO and SG scores than their counterparts. A longer life expectancy was also significantly associated with lower TTO scores. As the life expectancy increased, patients who had sustained SCI more recently (≤1y ago) were more willing to trade off life years for full health (ie, lower TTO scores). SG scores did not significantly vary according to different life expectancies. ConclusionsAmong patients with SCI, there was a nonlinear relation between the TTO and life expectancy that violated the fundamental assumption of risk neutrality toward life years. Accordingly, TTO utilities elicited for different life expectancies should not be compared in quality-of-life assessments or used in cost-utility analyses. In contrast, SG utilities remained stable at different lengths of life expectancy in SCI patients. Moreover, certain patient characteristics such as age, time lapse since the injury, neurologic severity, risk attitudes, and depressive status can account for some TTO and SG variations among patients with SCI.