Designing a new scale to measure anxiety symptoms in P arkinson's disease: item selection based on canonical correlation analysis

BACKGROUND AND PURPOSE: The lack of appropriate measures has hindered the research on anxiety syndromes in Parkinson's disease (PD). The objective of the present cross-sectional, international study was to identify shared elements and grouping of components from anxiety scales as a basis for designing a new scale for use in PD. METHODS: For this purpose, 342 consecutive PD patients were assessed by means of the Mini International Neuropsychiatric Inventory (depression and anxiety sections), the Clinical Global Impression of severity of the anxiety symptoms, the Hamilton Anxiety Rating Scale (HARS), the Neuropsychiatric Inventory (section E), the Beck Anxiety Inventory (BAI) and the Anxiety subscale of the Hospital Anxiety and Depression Scale (HADS-A). RESULTS: As the HADS-A showed a weak correlation with the HARS and BAI, it was not considered for more analyses. HARS and BAI exploratory factor analysis identified nine factors (62% of the variance), with only two of them combining items from both scales. Therefore, a canonical correlation model (a method to identify relations between components of two groups of variables) was built and it showed four factors grouping items from both scales: the first factor corresponded to 'generalized anxiety'; the second factor included muscular, sensory and autonomic 'non-specific somatic symptoms'; the third factor was dominated by 'respiratory symptoms'; and the fourth factor included 'cardiovascular symptoms'. CONCLUSIONS: BAI is heavily focused on panic symptoms, whilst HARS is more focused towards generalized anxiety symptoms. The new scale should include additional components in order to assess both episodic and persistent anxiety as well as items for evaluation of avoidance behaviour.