Note: Copies of the Geriatric Anxiety Inventory may be obtained by emailing email@example.com
Validity and reliability of the Geriatric Anxiety Inventory in Parkinson's disease
Article first published online: 29 DEC 2010
© 2010 The Authors. Australasian Journal on Ageing © 2010 ACOTA
Australasian Journal on Ageing
Volume 31, Issue 1, pages 13–16, March 2012
How to Cite
Matheson, S. F., Byrne, G. J., Dissanayaka, N. N., Pachana, N. A., Mellick, G. D., O'Sullivan, J. D., Silburn, P. A., Sellbach, A. and Marsh, R. (2012), Validity and reliability of the Geriatric Anxiety Inventory in Parkinson's disease. Australasian Journal on Ageing, 31: 13–16. doi: 10.1111/j.1741-6612.2010.00487.x
[Correction added after online publication 29 Dec 2010: Author affiliation details for Nancy A Pachana]
- Issue published online: 14 MAR 2012
- Article first published online: 29 DEC 2010
- anxiety disorder;
- Geriatric Anxiety Inventory;
- Parkinson's disease;
- psychiatric status rating scales
Aim: To examine the psychometric properties of a novel anxiety rating scale, the Geriatric Anxiety Inventory (GAI) in Parkinson's disease (PD).
Method: The predictive validity of the GAI was tested against the presence of any DSM-IV anxiety disorders in 58 PD patients using receiver operating curve analysis. The concurrent validity of this scale was also studied against the state half of the Spielberger State Trait Anxiety Inventory (STAI). The internal consistency and test–retest reliability of the GAI were also examined.
Results: The GAI displayed good concurrent validity against the STAI and the DSM-IV. It also showed good internal consistency and test–retest reliability.
Conclusions: This study suggested that the GAI is an appropriate scale to use in non-demented PD patients.
The neuropsychiatric features of Parkinson's disease (PD) impact significantly on the quality of life of both patients and carers. The prevalence of anxiety symptoms in PD has been variously estimated at between 30% and 40% . The most common anxiety symptom clusters in patients with PD are panic, phobic anxiety and generalized anxiety disorders [2,3]. Anxiety symptoms and disorders frequently coexist with depressive symptoms and disorders, and patients with a combination of anxiety and depressive symptoms experience more severe PD symptomatology, treatment resistance and increased functional impairment . Critically, the overlapping somatic symptom profile between anxiety and PD results in under-recognition of anxiety in PD and thus many of these patients remain untreated. In particular, symptoms of fatigue, restlessness, poor concentration, insomnia, irritability and muscular tension occur in both PD and anxiety disorders .
There are currently no recommended scales to assess anxiety in PD . To date, only two studies have examined the psychometric properties of anxiety rating scales in PD patients. The first study evaluated the Hospital Anxiety and Depression Scale (HADS) in 205 PD patients  and the second examined the HADS together with the Spielberger State Trait Anxiety Inventory (STAI) and the Hamilton Anxiety Scale in 46 PD patients . None has compared the usefulness of these scales against ‘gold standard’ Diagnostic and Statistical Manual Edition IV (DSM-IV) diagnoses of anxiety disorders.
The main goals of our study were: (i) to examine the validity of the Geriatric Anxiety Inventory (GAI) , against DSM-IV anxiety disorder diagnoses in PD patients; (ii) to investigate the concurrent validity of the GAI against the STAI ; (iii) to study the reliability of the GAI in PD patients; and (iv) to examine the repeatability of the use of the GAI in PD.
The GAI contains 20 items and focuses on the cognitive manifestations of anxiety, limiting the number of items containing somatic symptoms that overlap with PD symptoms. It is relatively brief, with a dichotomous response set for ease of administration. The GAI is a self-report measure, but may also be read out to the patient by a clinician or carer. These properties make it appealing for use in medically ill groups such as PD patients, as well as in older people. This scale is not designed to diagnose specific anxiety disorders, but rather to measure the severity of anxiety symptoms in people with either syndromal or subsyndromal manifestations of anxiety.
Our previous work in a non-PD sample has shown the GAI to be free of confounding by age, sex and cognitive function, to have satisfactory inter-rater reliability (Spearman's rho = 0.99, kappa = 0.50, z = 16.92, P < 0.001) and to have moderate concurrent validity against a range of other anxiety scales, including the Beck Anxiety Inventory and the STAI . Moreover, the GAI has shown high internal consistency as reflected in Cronbach's alpha values of 0.91 and 0.93 in non-PD older people with and without psychiatric disturbances, respectively. The usefulness of the GAI in PD patients is evaluated in the present study.
The PD patients were recruited from Neurology outpatient clinics in Brisbane, Australia. A diagnosis of idiopathic PD according to the UK brain bank criteria  was made by neurologists with a special interest in PD. All participants were of Caucasian origin, residing in Brisbane, free of comorbid dementia (Mini-Mental State Examination (MMSE)  of >27) and able to complete the questionnaires by themselves or with an assistant. Written informed consent was provided by all PD patients and the study was approved by Human Research Ethics Committees at the participating institutes.
The GAI, the state half of the STAI and the Geriatric Depression Scale (GDS-15)  were mailed to the participants. Face-to-face interviews were conducted at patient residences by a psychiatrist or a trained research assistant within 2 weeks of the mail-out. A diagnosis of anxiety and depressive disorders according to the DSM-IV criteria was made using the Mini International Neuropsychiatric Interview plus version (MINI-Plus) . The MINI-Plus is a well-validated, structured diagnostic interview, which was designed for use in research. It includes 23 DSM-IV disorders, and examines for both current and lifetime symptoms. It has shown good concordance with the SCID (Structured Clinical Interview for DSM-IIIR diagnoses) with kappa values showing good or very good agreement between MINI and SCID diagnoses.
The Hamilton Depression Rating Scale was also applied at the interview. The severity of PD was measured by the Unified Parkinson's Disease Rating Scale  and the Hoehn and Yahr Staging scale , and functional independence was evaluated using the Schwab and England disability scale . In a convenience sub-sample of 24 participants, the GAI was administered again at the interview in order to measure its test–retest reliability. All participants were asked to complete the questionnaires while they were in their ‘on’ state with respect to their PD symptoms (that is, at a time when their symptom control was optimal) and the interviews were also conducted in their ‘on’ state.
Fifty-eight patients completed the study. Characteristics of the patients are shown in Table 1, including their age, sex and measures of PD severity. Table 1 also describes the incidence of anxiety and mood disorders within the study group, and the results of the psychiatric status rating scales administered.
|Age (mean, range) in years||66.24 (37–85)|
|Sex (male)||34 (57%)|
|Marital status (married or de facto)||47 (81%)|
|Duration of Parkinson's disease (mean, range) in years||8.15 (1–32)|
|Current anxiety disorder||16 (28%)|
|Current depressive disorder||12 (21%)|
|Comorbid anxiety and depressive disorder||10 (17%)|
|Mean (SD; range) GAI||5.03 (6.06; 0–20)|
|Mean (SD; range) STAI (state version)||35.08 (12.37; 11–69)|
|Mean (SD; range) Hamilton Depression Rating Scale||10.69 (8.23; 2–39)|
|Mean (SD; range) Geriatric Depression Scale||4.57 (3.58; 0–14)|
|Mean (SD; range) MMSE||29.10 (0.95; 27–30)|
|Mean total (SD; range) UPDRS score||40.90 (18.59; 9–103)|
|Mean total (SD; range) Swab and England scale||77% (13%; 40–100%)|
|Hoehn and Yahr staging (mild = 1, 1.5, 2; moderate = 2.5, 3; severe = 4, 5)||23(40%)/31(53%)/4(7%)|
The GAI item characteristics are presented in Table 2. The corrected item total-correlation exceeded 0.50 reflecting that the quality of the data was good. The mean (SD; range) GAI score was 5.03 (6.06; 0–20). In men (n= 34) the mean GAI score was 5.23 (6.21; 0–20) and in women (n= 23) it was 4.95 (6.00; 0–20) (Wilcoxon rank-sum test: z = 0.025, P= 0.98). Significantly higher GAI scores were observed in PD patients with anxiety disorders than with those without anxiety. The mean (SD) GAI score in patients with a current DSM-IV anxiety disorder (n= 16) was 12.25 (5.72), whereas in patients without a current anxiety disorder (n= 42) it was 2.29 (3.32) (Wilcoxon rank-sum test: z = 4.99, P < 0.001).
|Item||Mean||SD||Corrected item-total correlation|
|1 I worry a lot of the time||0.40||0.49||0.75|
|2 I find it difficult to make a decision||0.29||0.46||0.54|
|3 I often feel jumpy||0.31||0.47||0.63|
|4 I find it hard to relax||0.31||0.47||0.50|
|5 I often cannot enjoy things because of my worries||0.16||0.37||0.66|
|6 Little things bother me a lot||0.29||0.46||0.68|
|7 I often feel like I have butterflies in my stomach||0.19||0.40||0.56|
|8 I think of myself as a worrier||0.40||0.49||0.78|
|9 I can't help worrying about even trivial things||0.24||0.43||0.82|
|10 I often feel nervous||0.33||0.47||0.72|
|11 My own thoughts often make me anxious||0.36||0.48||0.75|
|12 I get an upset stomach due to my worrying||0.09||0.28||0.60|
|13 I think of myself as a nervous person||0.33||0.47||0.70|
|14 I always anticipate the worst will happen||0.24||0.43||0.76|
|15 I often feel shaky inside||0.28||0.45||0.64|
|16 I think that my worries interfere with my life||0.17||0.38||0.78|
|17 My worries often overwhelm me||0.09||0.28||0.53|
|18 I sometimes feel a great knot in my stomach||0.09||0.28||0.59|
|19 I miss out on things because I worry too much||0.24||0.43||0.68|
|20 I often feel upset||0.26||0.44||0.72|
Anxious PD patients recorded significantly lower mean (SD) Schwab and England Scale scores than the non-anxious PD patients (anxious = 68.4 (13.6); non-anxious = 80.8 (11.8); Wilcoxon rank-sum test, z = 3.018, P= 0.002). The GAI score was negatively correlated with the Schwab and England Scale (Spearman's rho =−0.51; P < 0.001) and suggested increased disability and dependency in anxious PD patients.
The test–retest reliability of the GAI (n= 24) was satisfactory (Spearman's rho = 0.99; Kappa = 0.38; z = 5.05; P < 0.001) and the Kuder-Richardson coefficient (a measure of internal consistency for dichotomous scales) was 0.95. The mean (SD; range) score on the STAI was 35.08 (12.37; 11–69) and the concurrent validity of the GAI against the STAI was high (Spearman's rho = 0.69; P < 0.001).
We undertook a receiver operating characteristic (ROC) analysis to determine the optimum GAI cut point to distinguish patients with and without any current DSM-IV anxiety disorder in this group of patients with PD (Figure 1). The area under the ROC curve was 0.91 (95% CI 0.81–0.97). The optimum cut point was 6/7 (i.e. a score of 7 or greater), which correctly classified 86% of patients with a sensitivity of 87.5% and a specificity of 85.7%. As displayed in Figure 1, the area under the ROC curve for the Spielberger questionnaire in this patient population was 0.84. Although the difference between the areas under the two curves does not reach statistical significance (χ2= 3.19; P= 0.074), Figure 1 shows that the GAI appears to offer superior performance in distinguishing patients with anxiety disorders in this study population.
This study investigated the psychometric properties of a novel anxiety rating scale, the GAI, for use in PD patients. We showed that the GAI displays good concurrent validity against the STAI and the DSM-IV, good internal consistency and satisfactory test–retest reliability. The area under ROC curves (Figure 1) suggested that both the GAI and the STAI are appropriate to use in PD patients. The GAI can be used with an optimal cut-off value of 6/7, which was generated against a DSM-IV diagnosis of anxiety disorders. In this way PD patients who score ≥7 in the GAI can be classified as anxious and the patients who score ≤6 can be classified as non-anxious. These PD-specific cut-off values for dichotomising patients into anxious and non-anxious groups are likely to be valuable for future PD research using large sample sizes, including investigations of risk factors for clinically significant anxiety in PD.
Our sample of PD patients showed a high prevalence of DSM-IV anxiety disorders (28%), and high rates of comorbid depressive illness (17%) compared with the prevalence rates observed in the general non-PD older participants (10%) [16–18]. Panic disorder, social phobia and generalized anxiety disorder predominated and this was similar to previous reporting of anxiety disorders found in PD patients [19–21]. Limitations in our sample size did not allow us to examine the validity of the GAI against specific diagnosis of DSM-IV anxiety disorders. Future studies including larger sample sizes would be useful to derive cut-off values that could be used to dichotomise PD patients according to each type of DSM-IV anxiety disorder. Additionally, our sample size did not allow us to examine the usefulness of the scales in PD patients with and without comorbid depressive disorders; this area of research also warrants future study.
Several of the commonly used anxiety scales have properties which limit their usefulness in the PD patients . Many of the motor and autonomic manifestations of PD also occur in anxiety states, which can make detection of anxiety symptoms difficult. For this reason, anxiety scales which include many somatic symptoms may be less useful as screening tools in the Parkinson's population. In addition, because of potential difficulties with concentration and performance fatigue, scales with complex items or response sets (e.g. the Beck Anxiety Inventory) may be unsuitable for this patient population. The GAI presented in our study avoids these somatic symptoms and we have shown that it is an appropriate instrument to utilise in PD patients.
There is significant variation in the anxiety symptom pattern between patients with PD because of their fluctuations in motor symptoms [22–25]. While some experience a predictable increase in anxiety in motor ‘off’ states, others have an unpredictable pattern of motor ‘off’ states and non-motor symptoms. It is also possible that increased anxiety symptoms in ‘off’ periods are a function of the emotional distress related to increased physical disability. In our study, patients were uniformly assessed while in motor ‘on’ state. Investigating psychometric properties of these anxiety rating scales at motor ‘off’ states may be a useful focus of further research.
In our study, the GAI has only been tested in non-demented PD patients (MMSE > 27). A significant proportion of PD patients experience cognitive deficits, particularly at later stages of the disease  and the psychometric properties of the GAI in PD patients with cognitive deficits also warrant future study.
In conclusion, we have shown the usefulness of a novel anxiety rating scale, the GAI, in PD and compared it with the widely used anxiety rating scale, the STAI. We found that the GAI has good reliability and validity in patients with PD who are free of clinically significant cognitive impairment. Brevity, ease of use with a dichotomous response set, and avoidance of somatic symptoms that overlap between PD and anxiety, suggest that the GAI is an appropriate anxiety rating scale to use in PD patients.
- • This paper examines the applicability of a novel anxiety scale, the GAI, to patients with PD.
- • The GAI showed good concurrent validity with the STAI and DSM-IV.
- • The GAI displayed good reliability and test–retest repeatability.
- • The GAI appears to be an appropriate anxiety rating scale to use in PD patients.
- 3Anxiety disorders in Parkinson's disease. Advances in Neurology 2005; 96: 42–55..
- 8Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press, 1970., , .
- 13Committee MotUD. Unified Parkinson's Disease Rating Scale. In: Fahn S, Marsden CD, Calne DB, Goldstein M, eds. Recent Developments in Parkinson's Disease, Vol. 2. Florham Park, NJ: MacMillan Healthcare Information, 1987: 153–164., ,
- 15Projection technique for evaluating surgery in Parkinson's disease. In: Gillingham FJ, E.M.L. D, eds. Third Symposium on Surgery in Parkinson's Disease. Edinburgh: Livingstone, 1969: 152–157., .
- 24The ups and downs of Parkinson disease: A prospective study of mood and anxiety fluctuations. Cognitive and Behavioral Neurology 2004; 17: 201–207., , et al.