Age and Ageing 1999; 28: 35–38 䉷 1999, British Geriatrics Society

Use of the GDS-15 geriatric depression

scale as a screening instrument for
depressive symptomatology in patients
with Parkinson’s disease and their
carers in the community
JOLYON MEARA, EURIAN MITCHELMORE, PETER HOBSON
University Department of Geriatric Medicine, Glan Clwyd Hospital, Rhyl, North Wales LL18 5UJ, UK

Address correspondence to: J. Meara. Fax: (+ 44) 1745 534668

Abstract
Objectives: to assess the level of depressive symptomatology in a community based group of patients with
Parkinson’s disease (PD) and their carers and to investigate the patient characteristics that might predict carer
distress.
Methods: the GDS-15 geriatric depression scale was used to measure self-rated depressive symptoms in a group of
132 subjects with clinically probable PD randomly selected from a community-based disease register. Disease
severity was assessed by the Webster scale and cognitive function by the CAMCOG test. Carers of the patients, who
in this study were all spouses, were also asked to complete the GDS-15.
Results: a total of 64% of our group of patients and 34% of carers scored within the ‘depressed’ range on the GDS-
15. Patients with high levels of depressive symptoms tended to have more severe disease, disease of longer duration
and more impaired cognitive function. The GDS score of the carer was best predicted by the GDS score of the
patient being cared for. Less than 10% of patients and carers were being treated with antidepressant medication.
Conclusions: this community-based study confirms the high level of depressive symptoms in PD suggested by
hospital- and clinic-based studies. Depression in patients appears to be related to disease severity and cognitive
impairment. An important determinant of carer distress and mood disorder, as reflected by the GDS score, appears
to be the level of depression expressed by the patient being cared for. Despite high levels of depressive symptoms
in both patients and carers, very few subjects were in receipt of antidepressant drug therapy.

Keywords: carer, depression, GDS-15, depression scale, Parkinson’s disease

depression of 33% [3] and 26% [4]. The prevalence of

Introduction
The most common neuropsychiatric complication of
Parkinson’s disease (PD) is depression [1]. Prevalence
figures for depression in PD vary widely and, due to
methodological differences between studies, are diffi-
cult to compare. Studies using standardized criteria for
both depression and PD in hospital-based patient
populations generally report a prevalence for major
depression of around 40% [1], although much lower
figures have also been reported [2]. The prevalence of
depression in community-based populations of sub-
jects with PD remains to be established, although two
recently published studies have reported figures for
major affective disorder, as defined by the criteria of
fourth edition of the American Psychiatric Association’s
Diagnostic and Statistical Manual of Mental Disorders
(DSM-IV), may also be lower in the community than in
hospital-based studies. Two community-based studies
have reported frequencies for major depression based
on DSM-IV criteria of around 8% [4] and 3% [5] of
subjects.
Because depressive symptoms in PD are so common
and potentially reversible by pharmacological and non-
pharmacological means, it is important that depression
is recognized and treated. Antidepressant drug therapy
is effective in treating depression in PD [6], although

35
J. Meara et al.
the older tricyclic antidepressants are poorly tolerated
[7]. The newer antidepressant drugs such as the
selective serotonin re-uptake inhibitors have still to
be evaluated in PD, although one study has demon-
strated the effectiveness of sertraline in this condition
[8]. Cognitive therapy is effective in the treatment of
depression but has not been evaluated in PD.
The self-report screening instruments available to
detect depressive symptomatology would seem suita-
ble for use in community-based studies. The shortened
version of the self-administered Geriatric Depression
Scale [9], the GDS-15, has been validated in elderly
community dwellers and elderly patients in other
settings [10, 11]. A cut-off point of ⱖ5 on the GDS-15
has a sensitivity around 90% and a specificity of 70%
when compared with validated psychiatric diagnosis
[10, 12].
Most patients with PD requiring some care are
looked after by relatives, usually a spouse, at home.
Little is known about how depression in a patient with
PD affects the level of distress experienced by a carer.
One study found that distress in carers was best
predicted by motor impairment of the patient with PD
[13], although a more recent study, in hospital
inpatients, reported that carer distress was correlated
strongly with the level of depression, rather than
physical disability, in the patient for whom they were
caring [14]. The latter study also demonstrated for the
first time that carers of subjects with PD showed higher
levels of distress and strain than age-matched spouses
of subjects in good health [14]. Whether a similar
relationship exists between carer distress and patient
depression in community samples is unknown.
The detection and active treatment of depression in
patients with PD may be important in treating the
distress of their carers. In this study we have used the
GDS-15 to screen for depressive symptoms in a
community-based sample of PD patients and their
carers to establish the frequency of depression in these
two groups and positive predictors of carer distress.
Methods
A random sample of 132 subjects with clinically
probable PD [15] without severe dementia was
drawn from a community-based register for parkinson-
ism that has been established in a geographically
defined area of North Wales. Cases of parkinsonism
were identified on the basis of receipt from general
practitioner of a prescription of anti-parkinsonian
medication. All general practices in the study area
used computerized prescribing systems and could
generate a list of all subjects receiving treatment for
PD. Patients ascertained in this way were visited at
home and a history was taken and neurological
examination performed. The presence or absence of
true parkinsonism and the likely type of parkinsonism
36
present was determined by this assessment and the
examination of any available medical records. Diag-
nosis was based on the best currently available
diagnostic criteria [15].
Patients with PD were invited to complete the GDS-
15 and the CAMCOG assessment of cognitive function
[16] as part of a semi-structured interview in their own
homes or institutional setting. Patients who scored <70
on the CAMCOG, indicating severe dementia, were not
included in this study. If requested, assistance in
completing these assessments was given to those
with sensory or motor difficulties. Demographic details
and medical histories of the PD patients were also
collected at the time of interview. All subjects were
screened for depressive symptomatology by use of
the GDS-15, using a cut-off value of ⱖ5 to indicate
clinically important depressive symptoms. Formal
psychiatric diagnosis was not attempted.
The severity of PD was measured by the Webster
rating scale, taking a score of 0–10 as indicating mild
disease, 10–20 as moderate disease and 20–30 as
advanced disease [17].
Where a patient had a spouse living with them who
was the main carer, the spouse was also invited to
complete the GDS-15. Of the 132 subjects selected for
this study 79 had a spouse as main carer.
The data were analysed using the SPSS for Windows
release 6.0 statistical computer package.
Results
The demographic details of the patients and their
disease severity ratings are given in Table 1. The overall
mean score on the GDS-15 for the PD group was 5.6
(range 0–13). A total of 84 (64%) of PD patients had
significant depressive symptoms using the GDS-15. No
significant differences were observed between PD
patients with and without significant depressive
symptoms in terms of age, sex, age of onset of PD
or previous reported history of depressive illness.
Differences were observed between the two groups in
relation to disease severity, duration of disease and
Table 1. Demographic details and disease characteristics of
subjects with Parkinson’s disease (n =132)
Mean (and SD) Range %
.......................................................................................................
Age (years) 73.4 (9.5) 46–89 –
Gender (% male) – – 55
Parkinson disease
Age at onset (years) 65.3 (11.6) 26–88 –
Duration (years) 7.8 (6.9) 1–40 –
Score
Webster 16.9 (5.5) 6–30 –
GDS-15 5.6 (3.1) 0–14 –
% treated for depression – – 7
 Use of the GDS-15 in patients with Parkinson’s disease

Table 2. Sub-group comparisons of disease characteristics, Table 4. Correlation (Spearman rank correlation coefficient)
Webster score and patient and carer GDS-15 scores in between patient variables (GDS-15 score, Webster score,
patients with Parkinson’s disease CAMCOG score) and carer GDS-15 score
Patient GDS-15 score Patient score
............................................. .....................................................................

<5 ⱖ5 Carer score GDS-15 Webster CAMCOG

.......................................................................................................
No. of patients
No. of carers
Mean age of patients (years)
Mean value (and SD) for patients
GDS-15
Webster score
Disease duration (years)
CAMCOG score
Mean (and SD) GDS-15 for carers
score on the CAMCOG assessment of cognitive
function (P < 0.05 see Table 2).
Demographic details and GDS-15 scores of carers
are shown in Table 3. There were significantly more
women caring for a disabled partner in the depressed
PD group (P < 0.05). The mean GDS-15 score for the
carers was 3.6. A total of 28 (35%) scored ⱖ5 on the
GDS-15 and 6% were receiving antidepressant medica-
tion from their general practitioner. A between-group
analysis of carers of depressed and non-depressed PD
patients demonstrated significantly more depressive
symptomatology in those caring for depressed patients
(P < 0.05, see Table 2).
The correlation between characteristics of the PD
patient being cared for and the carer’s GDS-15 score is
shown in Table 4. The only patient variable that
correlated strongly with the carer’s GDS-15 score was
the patient’s GDS-15 score. While neither disease
severity nor disease duration correlated with the
carer’s GDS-15 score, there was a trend in the data to
suggest that impaired cognitive function in the patient
(indicated by a low CAMCOG score) was associated
with higher levels of mood disturbance in the carer.
Discussion
This community-based study confirms that PD is
associated with high levels of depressive symptomatol-
ogy, with 64% of the patient group scoring in the
Table 3. Mean ages and GDS-15 scores of
carers [n = 79; 30 men (38%), 49 women
(62%)]
Mean (SD)
.......................................................................
Age (years) 69.8 (9.5)
GDS-15 score 3.6 (3.3)
a
Significant at P < 0.05 Mann–Whitney U-test.
b
not significant.
.......................................................................................................
a
48 84 GDS-15 0.39 0.088 ¹0.133
33 46
71.8 74.1b a
Significant P < 0.01.
a
2.5 (1.3) 7.5 (2.2)
14.4 (4.8) 18.4 (5.3)a
5.8 (4.1) 8.9 (7.8)a
‘depressed’ range of the GDS-15. The impression
83.8 (12.6) 78.7 (14.6)a gained from patient interviews was that, whereas
2.1 (2.1) 4.8 (3.6)a depressive symptoms were very common in the
patient group, the occurrence of major affective
disorder was relatively rare. Depression in patients
was associated with severity of disease, as measured by
the Webster scale, and disease duration—both of
which clearly interact. This has been found by most
[18–20] but not all other investigators [21]. It is likely
that the relationship between disease severity, disease
duration and mood disorder is complex and this study
could not address the issue of whether change in
disability is more important than absolute disability or
whether an increase in disability occurring soon after
diagnosis and again in advanced disease is a more
powerful predictor of mood in the patient than
absolute levels of disability [19].
We also found that depressive symptoms were
common in carers, with over one-third of carers
scoring in the depressed range on the GDS-15.
The only patient variable that correlated with the
presence of depressive symptoms in the carer was the
GDS-15 score. This supports the findings of the only
other recent study of depression in carers of patients
with PD recruited mainly from hospital sources [14].
However, there are several points that must be
borne in mind when considering our findings. Our
conclusions are limited by the method of ascertain-
ment of cases of PD, the use of a self-rating scale to
determine depressive symptoms and the cross-sec-
tional design of the study. Our group of subjects with
PD was restricted to those medically diagnosed and in
receipt of anti-parkinsonian medication. Possibly
50% of symptomatic cases of PD may be medically
undiagnosed [22, 23] and some medically known
subjects will not be on drug treatment. Although only
patients classified as having clinically probable PD
were studied, it is probable that up to 20% of this group
may have had parkinsonism resulting from a cause
other than PD.
The GDS-15 performs well in comparison to formal
Range
psychiatric classification of depression, with good
44–87 sensitivity and reasonable specificity, although the
0–14 performance of the GDS-15 in the presence of
cognitive impairment is less reliable [24]. In this
study, patients with marked impairment on the
CAMCOG test of cognitive function were excluded.

37
J. Meara et al.
The sensitivity of self-rating scales for depression such
as the GDS-15 in medically ill subjects is likely to result
in a reasonably large proportion of false-positive
results. In our patient group perhaps 30% of the
group scoring in the depressed range on the GDS-15
would not be classified as depressed by formal
psychiatric diagnosis. Nonetheless, the GDS-15 is still
detecting high levels of depressive symptoms in these
patients. Since treatment validation studies of the
results of screening for depression with instruments
such as the GDS-15 in medically ill subjects have not
been undertaken, the proportion of our patients with
high levels of depressive symptoms who would benefit
from antidepressant treatment is unknown. Despite
the high levels of depressive symptoms that were
detected in this study and evidence of the effectiveness
of antidepressant drug treatment in PD [6, 8], only 7%
of patients and 6% of carers were being treated with
antidepressants.
Key points
• In a community sample of patients with Parkinson’s
disease and their carers, marked depressive symp-
toms detected by self-rated scales were found to be
present in two-thirds of patients and one-third of
carers.
• Carer distress was predicted by the degree of
depressive symptomatology in the patient.
• Despite high levels of depressive symptoms, <10%
of subjects in each group were in receipt of
antidepressant drug treatment.
Acknowledgement
This study was partly funded by a Parkinson’s Disease
Society Welfare Grant that supported E.M.
References
1. Cummings JL. Depression and Parkinson’s disease: a review. Am J
Psychiat 1992; 4: 443–54.
2. Brown RG, MacCarthy B. Psychiatric morbidity in patients with
Parkinson’s disease. Psychol Med 1990; 20: 77–87.
3. Tison F, Dartigues JF, Auriacombe S, Letenneur L, Boller F,
Alperovitch A. Dementia in Parkinson’s disease: a population based
study in ambulatory and institutionalized individuals. Neurology
1995; 45: 705–8.
4. Tandberg E, Larsen JP, Aarsland D, Cummings JL. The occurrence
of depression in Parkinson’s disease. A community based study. Arch
Neurol 1996; 53: 175–9.
5. Hantz P, Caradoc-Davies G, Caradoc-Davies T, Weatherall M, Dixon G.
Depression in Parkinson’s disease. Am J Psychiatry 1994; 151: 1010–4.
6. Andersen J, Aabro E, Gulmann N, Hjelmsted A, Pedersen HE. Anti-
depressive treatment in Parkinson’s disease. Acta Neurol Scand 1980;
62: 210–9.
38
7. Laitinen L. Desipramine in treatment of Parkinson’s disease. Acta
Neurol Scand 1969; 45: 109–13.
8. Meara RJ, Bhowmick BK, Hobson JP. An open uncontrolled study
of the use of sertraline in the treatment of depression in Parkinson’s
disease. J Serotonin Res 1996; 4: 243–9.
9. Yesavage JA, Brink TL. Development and validation of a geriatric
depression screening scale: a preliminary report. J Psychiatr Res
1983; 17: 1: 37–49.
10. D’Ath P, Katona P, Mullan E, Evans S, Katona C. Screening,
detection and management of depression in elderly primary care
attenders. I: The acceptability and performance of the 15 item
Geriatric Depression Scale (GDS15) and the development of short
versions. Fam Pract 1994; 11: 3: 260–6.
11. Lesher EL, Berryhill JS. Validation of the Geriatric Depression
Scale—Short Form among inpatients. J Clin Psychol 1994; 50: 2:
256–60.
12. Jackson R, Baldwin B. Detecting depression in elderly medically
ill patients: the use of the Geriatric Depression Scale compared with
medical and nursing observations. Age Ageing 1993; 22: 349–53.
13. Calder SA, Ebmeier KP, Stewart L, Crawford JR, Besson JA. The
prediction of stress in carers: the role of behaviour, reported self-care
and dementia in patients with idiopathic Parkinson’s disease. Int J
Geriat Psychiatry 1991; 6: 737–42.
14. Miller E, Berrios GE, Politynska BE. Caring for someone with
Parkinson’s disease: factors that contribute to distress. Int J Geriat
Psychiatry 1996; 11: 263–8.
15. Hughes AJ, Daniel SE, Kilford L, Lees AJ. The accuracy of clinical
diagnosis of idiopathic Parkinson’s disease: a clinicopathological
study. J Neurol Neurosurg Psychiatry 1992; 55: 181–4.
16. Huppert FA, Brayne C, Gill C. CAMCOG—a concise neuropsy-
chological test to assist dementia diagnosis: socio-demographic
determinants in an elderly population sample. Br J Clin Psychol
1995; 34: 529–41.
17. Webster DD. Clinical analysis of the disability in Parkinson’s
disease. Mod Treatment 1968; 5: 257–82.
18. Mindham RHS, Marsden CD, Parkes JD. Psychiatric symptoms
during L-dopa therapy for Parkinson’s disease and their relationship
to physical disability. Psychol Med 1976; 6: 23–33.
19. Brown RG, MacCarthy B, Gotham AM, Der GJ, Marsden CD.
Depression and disability in Parkinson’s disease: a follow-up study of
132 cases. Psychol Med 1988; 18: 49–55.
20. Starkstein SE, Preziosi TJ, Bolduc PL, Robinson RG. Depression in
Parkinson’s disease. J Nerv Mental Dis 1990; 178: 27–31.
21. Santamaria J, Tolosa E, Valles A. Parkinson’s disease with
depression: a possible sub-group of idiopathic parkinsonism.
Neurology 1986; 49: 1130–3.
22. Morgante L, Rocca WA, Di Rosa AE et al. Prevalence of
Parkinson’s disease and other types of parkinsonism: a door-to-door
survey in three Sicilian municipalities. Neurology 1992; 42: 1901–7.
23. Meara RJ, Bisarya S, Hobson JP. Screening in primary health care
for undiagnosed tremor in an elderly population in Wales. J Epidemiol
Community Health 1997; 51: 5: 574–5.
24. Montorio I, Izal M. The Geriatric Depression Scale: a review of its
development and utility. Int Psychogeriatr 1996; 8: 1: 103–12.
Received 7 October 1997; accepted 20 November 1997