OBJECTIVES: Utility data is often needed in health economic modeling, but not always available in clinical studies in Parkinson’s disease (PD). Therefore, we have recently presented a prediction algorithm to estimate utilities from a clinical rating scale. We now report effects of major complications in PD on this prediction algorithm. METHODS: Data from an ongoing prospective cost study of the German Competence Network of Parkinson Syndromes was used (n=115). Our prior utility prediction algorithm was exclusively based on subscales II-IV of the Unified Parkinson’s Disease Rating Scale (UPDRS). Now, we applied multivariate regression analysis to investigate the independent effect of sociodemographic factors (age, sex), neuropsychological disturbances (depression, hallucinations, dementia) and motor complications (falls, dyskinesia, fluctuations) on utilities measured by the EuroQoL (EQ-5D). Depression was evaluated by Beck’s Depression Inventory (BDI) and dementia by Mini Mental State Examination. RESULTS: The study population had a mean age of 67 years and was predominantly male (67%). Mean EQ-5D was 0.74. Among all complications investigated, only depression reached statistical significance (p = 0.03). UPDRS variables (subscores from part II, III and IV) remained in the model (see equation): EQ-5D = (103.23 - 1.29*UPDRS II - 0.30*UPDRS III – 1.23*UPDRS IV – 0.55*BDI)/100. After adjusting for UPDRS, depression reduced the utility by 0.55 score units per BDI point. However, the inclusion of BDI only marginally improved the explained variance (adjusted R-square increased from 0.51 to 0.52). CONCLUSIONS: Whereas sociodemographic factors, motor complications and neuropsychological disturbances do not exert independent effects on utilities when controlling for UPDRS, depression has an independent effect on EQ-5D. Although depression is not explicitly included in the UPDRS score of parts II-IV, its inclusion did not substantially reduce the unexplained variance. It remains to be investigated which factors further explain the remaining variance.