A comparison of techniques for eliciting patient preferences in patients with benign prostatic hyperplasia.

Abstract

We compared the Health Utility Index (HUI), EuroQol (EQ-5D) and time trade-off methods to identify the most suitable technique for collecting preference data in a clinical trial of patients with benign prostatic hyperplasia (BPH). A total of 29 men with symptomatic BPH were interviewed by a single trained interviewer who collected demographic data and administered EQ-5D and time trade-off questionnaires. Participants self-administered the HUI and a symptom severity index, the International Prostate Symptom Score (I-PSS) questionnaire. Utility values for current patient health states obtained from the HUI, EQ-5D and time trade-off questionnaires were compared and their relationship with I-PSS data was examined using Spearman's correlation coefficients. Administration time and patient assessments of the relevance of the questions were also compared for the 3 methods. Although mean utility values for HUI, EQ-5D and 1-year time trade-off were similar, only utility values elicited using time trade-off with a 1-year time frame significantly correlated with symptom scores. The 1 and 10-year time trade-off derived values were reasonable predictors of the I-PSS with multiple correlation coefficient values of 0.379 and 0.265, respectively. All participants indicated that the HUI and EQ-5D were appropriate for assessing BPH, while approximately 10% considered time trade-off questions irrelevant. Average completion time for the HUI, time trade-off and EQ-5D questionnaires was 31, 25 and 10 minutes, respectively. Because only time trade-off resulted in utility values that significantly correlated with symptom scores, we recommend its use for estimating utility in clinical trials of BPH.