OBJECTIVES: To develop a statistical model to predict EQ-5D 3-level (EQ-5D-3L) utilities as a function of patient demographics and PD severity, as measured by the Unified Parkinson’s Disease (PD) Rating Scale (UPDRS) subscales. The aim was to develop a predictive equation for utilities, suitable for use in an economic model to conduct cost-utility analysis (CUA).

METHODS: Patient-level data were obtained from the National Institute of Neurological Disorders and Stroke (NINDS) Exploratory Trials in PD Long-Term Study 1 (NET-PD LS-1), a multicenter Phase 3 study of creatine in patients on dopaminergic therapy within 5 years of diagnosis (N=1,741; 6 years follow-up). The EQ-5D-3L index scores were calculated using the UK preference weights. The mean utility values and UPDRS scores were comparable between the two treatment arms in the trial, and thus patient-level data were pooled for analysis, as the treatment effect was not statistically significant. The data were analyzed using a mixed-effect model with repeated measures. Candidate predictors were informed by a previous SLR conducted to identify published studies that reported the association between utilities and PD severity (Chandler 2018).

RESULTS: The average decline in utilities per year was 0.018 and mean utilities at baseline, year 3, and year 6 were 0.81, 0.76, and 0.70, respectively. The significant predictors of utility values included gender and UPDRS I, II, III, and IV. Age was excluded from the multivariate model as it was not statistically significant after adjusting for UPDRS scores. The statistical model performed well in validation analyses—average predicted EQ-5D-3L utilities were compared with the average observed scores for each year post-baseline and were within +/-0.01 at all visits.

CONCLUSIONS: The predictive equation for utilities captures the impact of non-motor and motor-related aspects of the disease as all four UPDRS subscales were identified as significant predictors.