Letter to the EditorThe accuracy of the Unified Parkinson's Disease Rating Scale (UPDRS—Section 1) as a screening measure for depression
Introduction
Depression is a common feature in Parkinson's disease (PD), with prevalence rates varying between 7% and 70% [1]. To assist in the recognition of non-motor symptoms such as depression, clinicians are often guided by instruments such as the Unified Parkinson's Disease Rating Scale (UPDRS) [2]. This study sought to assess whether the UPDRS item relating to depression would accurately identify patients who required further screening by comparing it to three depression measures that have been validated for use in the PD population: the Beck Depression Inventory-II (BDI-II), Geriatric Depression Scale (GDS) and Hospital Anxiety and Depression Scale (HADS) [3], [4], [5]. Because there is no consensus regarding which measure should be used to screen depression in PD. We also examined the level of agreement among these scales using the optimal suggested cut-offs for this patient group.
Section snippets
Methods
The study gained approval from the local ethics committee and written consent was obtained from patients. Inclusion criteria were: no evidence of dementia (Mini Mental Status Exam (MMSE) of ⩾25) confirmed diagnosis of PD and <80 years of age. Fifty-nine patients who met the inclusion criteria volunteered to take part. Depression was assessed using the following measures: GDS (30 items rated 0/1 with a maximum score of 30), BDI-II (21 items rated 0–3 with a maximum score 63), and HADS (7 items
Statistical analysis
The relationship between the different measures of depression was assessed using Pearson's correlations. We then examined how well the UPDRS predicted possible and probable depression, as defined by the BDI-II, GDS or HADS score exceeding the respective cutoff, through percentage agreement and Receiver Operating Characteristic (ROC) analyses.
Results
Overall 29/59 patients (49%) had scores indicative of possible depression using the BDI-II, GDS or HADS. Correlations were moderately strong between the UPDRS and BDI-II: r=.61, p<.001; GDS: r=.63, p<.001; and HADS: r=.42, p<.01. The BDI-II and GDS were highly correlated: r=.77, p<.001. Correlations between the HADS and BDI-II and GDS were also positive: r=.53, p<.001 and r=.66, p<.001.
Table 1 shows crosstabulations between the UPDRS as diagnostic screen and the BDI-II, GDS, and HADS. The UPDRS
Discussion
Overall, 29 of 59 cases were identified as having possible depression by the BDI-II, GDS or HADS. The UPDRS failed to identify 34% of these cases as having any depressive symptoms whatsoever, which is an unacceptably high Type II error rate. ROC analyses showed that the UPDRS had only moderate accuracy overall for predicting possible depression, as measured by BDI-II and GDS, with average AUC=.70. There was a high level of agreement between the BDI-II and GDS. By contrast, the HADS identified
Conclusion
The UPDRS in its present form has limited utility as a screening instrument for possible depression. Administration of more comprehensive measures such as the BDI-II or GDS is advisable.
Acknowledgements
This project has been supported by grants from the Canterbury Medical Research Foundation. Audrey McKinlay was sponsored by a scholarship from the Foundation for Science Research and Technology and by a Claude McCarthy Fellowship.
References (5)
Beyond the iron mask: towards better recognition and treatment of depression associated with Parkinson's disease
Mov Disord
(2002)- et al.