Clinical Gerontologist

ISSN: 0731-7115 (Print) 1545-2301 (Online) Journal homepage: http://www.tandfonline.com/loi/wcli20

Clinical and Psychometric Validation of the

Geriatric Depression Scale (GDS) for Portuguese
Elders

Margarida T. S. Pocinho PhD , Carlos Farate MD PhD , Carlos A. Dias MD

PhD , Tina T. Lee MD & Jerome A. Yesavage MD

To cite this article: Margarida T. S. Pocinho PhD , Carlos Farate MD PhD , Carlos A. Dias MD
PhD , Tina T. Lee MD & Jerome A. Yesavage MD (2009) Clinical and Psychometric Validation of
the Geriatric Depression Scale (GDS) for Portuguese Elders, Clinical Gerontologist, 32:2, 223-236,
DOI: 10.1080/07317110802678680

To link to this article: https://doi.org/10.1080/07317110802678680

Published online: 14 Feb 2009.

Submit your article to this journal

Article views: 305

View related articles

Citing articles: 5 View citing articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=wcli20

Download by: [b-on: Biblioteca do conhecimento online UEvora] Date: 16 December 2017, At: 18:48
 Clinical Gerontologist, 32:223–236, 2009
Copyright © Taylor & Francis Group, LLC
ISSN: 0731-7115 print/1545-2301 online
DOI: 10.1080/07317110802678680

Clinical and Psychometric Validation

1545-2301
0731-7115
WCLI
Clinical Gerontologist,
Gerontologist Vol. 32, No. 2, Jan 2009: pp. 0–0

of the Geriatric Depression Scale (GDS)

for Portuguese Elders
Downloaded by [b-on: Biblioteca do conhecimento online UEvora] at 18:48 16 December 2017

MARGARIDA T. S. POCINHO, PhD

Geriatric
M. T. S. Pocinho
Depression
et al.Scale for Portuguese Elders

Superior School of Health Technologies (ESTES), Coimbra, Portugal

CARLOS FARATE, MD, PhD and CARLOS A. DIAS, MD, PhD

Superior Institute Miguel Torga (ISMT), Coimbra, Portugal

TINA T. LEE, MD and JEROME A. YESAVAGE, MD

Veterans Affairs Palo Alto Health Care System, Palo Alto, California, USA

The main purpose of this study is the adaptation and psychometric

validation of the Geriatric Depression Scale (GDS) for the Portuguese
population. The analyses of temporal (a (27 items) = .995) and internal
(.91) consistencies, inter-rater reliability (Kappa = .87), and concur-
rent (.83) and factorial (3 factor) validities demonstrate that GDS is
highly reliable as a screening instrument for geriatric depression. This
conclusion was further validated by both discriminant function
analysis and diagnostic value testing, which showed a sensitivity of
100%, a specificity of 83%, a positive predictive value of 93%, and a
negative predictive value of 100%. The study has also evaluated GDS
applicability to clinical assessment and therapeutic intervention with
elders suffering from moderate feelings of loneliness linked to recur-
ring suicidal ideation.

KEYWORDS Geriatric Depression Scale (GDS), reliability, factor

analysis, Portuguese elders

This research is supported by grant AG 17824 from the National Institutes of Health, by
the Medical Research Service of the Veterans Affairs Palo Alto Health Care System, and by
the Department of Veterans Affairs Sierra-Pacific Mental Illness Research, Education, and
Clinical Center (MIRECC).
Address correspondence to Margarida T. S. Pocinho, Department of Exact Sciences, Biology,
and Engineering, Superior School of Health Technologies (ESTES), Coimbra, Portugal. E-mail:
margarida_pocinho@estescoimbra.pt

223
 224 M. T. S. Pocinho et al.

INTRODUCTION

Depression is a serious mental and public health problem. It is a common

and highly comorbid disorder, affecting individual functioning and societal
welfare. People with depression often withdraw socially, perform poorly at
their usual tasks at work and home, and often think about suicide. Depres-
sion causes large economic losses because of increased morbidity, mortality,
and treatment costs.
Downloaded by [b-on: Biblioteca do conhecimento online UEvora] at 18:48 16 December 2017

Even though depression can be treated successfully, seriously depressed

individuals very often do not receive appropriate treatment. One reason is
that screening depression in the general population, and especially among
elderly people, is a challenging task. Koenig & George (1998) found that most
cases of depression are neither diagnosed nor treated during hospitalization.
Many clinicians believe that depression is situational and will resolve soon
after the patient is discharged from the hospital, despite epidemiological
evidence to the contrary.
The “Improving Mental Health Information Project” based in Scotland
has concluded that psychiatric diagnoses were made on only 1% of all acute
hospital discharges, despite a known prevalence of clinically recognizable
psychopathology in the general population of 15%. The Information and
Statistics Division (ISD) has suggested that mental health problems in
general hospitals are under-diagnosed and under-recorded (2002).
In fact, despite the amount of data issued from neurobiological and
pharmacological research on the influence of Central nervous system (CNS)
dopaminergic neuronal functioning, serotoninergic cells’ 5-HTA re-uptake
dynamics, and more recently, glutamate metabolism in depressive disorder
(Segal, Pearson, and Thase, 2003; Paul & Skolnick, 2003), there is not a
single biologic marker that can be used in simple laboratory monitoring
tests with animal or human models.
Furthermore, symptomatic subclinical depression often presents with
somatic and functional complaints that can mimic medical illness. This is
particularly true with depression in the elderly because of the psychophysi-
ologic changes, often of a declining or degenerative nature, that occur in
this phase of the life cycle. Psychosomatic involution, deteriorating physical
conditions, grieving, and dementia are common events that can mask
depression.
In fact, depression can be clinically expressed among the elderly in a
way that can sometimes be difficult to recognize, given its more atypical
clinical presentation as compared with other age groups. Instead of looking
sad, elders with depression will frequently present with physical ailments,
headaches and stomachaches (with no medical cause) as well as with
fatigue and irritability or both (Alexopoulos et al., 1999; Caine & Conwell,
2001; Conwell & Pearson, 2002; Karel et al., 2002; Williamson, Shaffer, and
Parmelee, 2002).
 Geriatric Depression Scale for Portuguese Elders 225

This complex clinical picture, plus the fact that psychiatric diagnoses
are based chiefly on the clinician’s subjective assessment of the patient’s
symptoms, explains the interest accorded to the development of an opera-
tive set of diagnostic measurable criteria based on internationally validated
structured clinical interviews (Evans & Mottram, 2000; Sinclair et al, 2001)
Furthermore, such a process can allow for the implementation of cross-
comparison and psychometric validation methods for data from interna-
tional studies. This is why Mcsweeney and Creer (1995) state that such
Downloaded by [b-on: Biblioteca do conhecimento online UEvora] at 18:48 16 December 2017

assessment tools must comply with methodological rules of validity, reliabil-

ity, and outlining of normative values in order to be scientifically useful as
research tools.
This is the rationale for the design of standardized systems that can
provide a more objective approach to the clinical and epidemiological
assessment of depression (Social Care Institute for Excellence [SCIE], 2004).
Much of the literature has been concerned with the development and
selection of the appropriate screening systems and diagnostic tools, e.g.,
Geriatric Depression Scale (GDS), Center for Epidemiological Depression
Scale (CES-D), psychogeriatric assessment scales (PAS), the Hamilton
Depression Rating Scale, Hospital Anxiety and Depression Scale (HADS),
and the Camberwell Assessment of Need for the Elderly (CANE) (Hamilton,
1967; Burns et al., 1999).
Accordingly, Burns et al. (2001) recommend that the use of standard-
ized screening instruments should be encouraged in: 1) primary care set-
tings, in order to identify people with psychiatric disorders who can then
be referred to 2) secondary care (specialized) services, where these screen-
ing tools can be extremely useful both for the assessment of the severity of
pathology and the evaluation of the best therapeutic management for each
individual.
In brief, it is important to emphasize that scales (or other evaluation
tools) presenting with reliable psychometric characteristics and based on
rigorous diagnostic criteria are the most useful in general population studies
for preventive and therapeutic purposes. This is precisely the case with the
geriatric depression scale (GDS), which is a self-report assessment tool for
depression in the elderly with demonstrated advantages over other instru-
ments of the same type, namely, it is easier to administer compared with
other self-reports; it is rather useful for detecting depressive disorder among
general population samples; it has excellent psychometric properties. It was
created and developed in 1982 by Brink, Yesavage, Lum, Heersma, Adey, et al.,
and various studies have confirmed its validity, reliability and stability qualities
to measure depressive disorders in general population samples (Montorio,
1996; Mui, 1996; Stiles & McGarrahan, 1998)
The GDS is designed as a 30-item inventory, with a yes/no format. It
takes 10 to 15 minutes to administer. Yesavage et al. (1983) reported a
Cronbach’s alpha internal consistency reliability for GDS of .94. These
 226 M. T. S. Pocinho et al.

psychometric properties have been well supported by data issued from a

number of studies conducted afterward (e.g., Koneig et al., 1988; Nitcher et al.,
1993; Mui, 1996).
The scale’s scores range from 0 to 30 representing the total number of
depressive symptoms According to Brink et al. (1982), those who report 10
or fewer symptoms are considered normal, those who report 11 to 20 symp-
toms are considered mildly depressed, and those who report 21 or more
symptoms are considered moderately to severely depressed.
Downloaded by [b-on: Biblioteca do conhecimento online UEvora] at 18:48 16 December 2017

The GDS has so far been translated and adapted into 26 languages,
including Brazilian (Zilenovski, 1991), Italian (Ferrario et al., 1990), and
Spanish (Perlado, 1987; González, 1988; González & Szurek, 1990; Montorio &
Izal, 1993).
In this paper, we will present data from the psychometric validity study
for the GDS Portuguese version. The field work for this study was under-
taken with a representative sample of Portuguese elders recruited from vari-
ous sites and residential settings. Additionally, this study evaluated the
suitability of the GDS in clinically assessing elders presenting with depressive
symptoms, in particular those suffering from moderate feelings of loneliness
linked to recurring suicidal ideation. This was the reason Alentejo was chosen
as the main study site. Alentejo has the highest suicide rate among the eld-
erly in Portugal, and has also one of the higher suicide rates in Europe.
Although there has been a previous validation study for a Brazilian version
of GDS (Zilenovski, 1991) and even if this version seemed to fit reasonably
well with a Portuguese language adaptation of the original scale, there are
subtle linguistic (semantic and pragmatic) differences between natives from
both countries that can be particularly significant for an elderly population
group (especially for those living outside greater urban centers and having a
lesser education level).
On the other hand, there is also an ecological validity issue, based on
socio-cultural differences (social traditions, personal meaning and commu-
nitarian judgment of certain emotions or somatic complaints in the context
of specific affective-behavioral, moral, and religious patterns) between elders
living in a South American and a European country, despite historical roots
they might share.
Moreover, there is a growing body of literature recommending a tran-
scultural approach for the creation of psychometric instruments with clinical
and epidemiologic characteristics allowing its usefulness in multicentric
studies. This is definitely the case for GDS.

METHODS

We used a number of sampling procedures in order to achieve five different

population samples. Participants in three of the five subsamples were
 Geriatric Depression Scale for Portuguese Elders 227

recruited from elders inhabiting the Center and North regions of Portugal
(whose prevalence rates for suicide are lower); another subsample was con-
stituted by elderly patients carrying a diagnosis of depression who were fol-
lowed in private psychiatric practice; finally, elderly people from Alentejo
(the region with the higher prevalence rate for suicide) were the subjects
recruited for the last subsample.

First Step: Translation and Back-Translation

Downloaded by [b-on: Biblioteca do conhecimento online UEvora] at 18:48 16 December 2017

The study began by the translation and back-translation of the scale’s items,
followed by initial pre-testing in a sample of 200 elderly people, age 65 to
92 years. Sixty-three percent (mean ± SD = 76.63 ± 6.44) of them (126 sub-
jects) were female, and 37% (n = 74) were male. Also, 62% of the subjects
enrolled in the pre-testing phase (n = 24 ) came from rural areas, whereas
38% (n = 76) lived in an urban environment.

Second Step: Concurrent Validity

After pre-test the next step was the evaluation of concurrent validity. To do
so, the Hamilton Rating Scale for Depression was used. This is a well
known and reliable instrument for measuring depression both in clinical
and general population samples. The sample was composed by 30 elders,
aged 65 to 84 years (mean ± SD = 71.83 ± 4.30). Twenty-five of them
(76.7%) were between the ages of 65 and 74 years. There were 14 women
(46.7%) and 16 ‘lived in an urban area.

Third Step: Factorial Validation

The following step was the check for reliability through a three-way path:

1. Factorial analysis and internal consistency, through principal compo-

nents methodology: Cronbach alpha and Kuder-Richardson.
In order to accomplish this objective, the scale was administered to
660 elders (58% came from the central of Portugal, 29% from the
northern region, 12% from Lisbon & Tagus valley and 1% from the
Azores & Madeira). In terms of gender distribution, 396 (60%) were
female and 264 (40%) were male. The age distribution showed a shift
toward the 65 to 74 age group (361 subjects, i.e., 54.7%) followed by
the 75 to 84 age group (250, accounting for 37.9% of the sample), and
by 7.4% (49) who were age 85 years of age or older. Among these
elders, 64.7% (427) came from rural areas and 35.3% (233) from
urban ones.
2. Temporal consistency (test-retest using an 8-day interval).
 228 M. T. S. Pocinho et al.

For this procedure, the scale was administered to 38 elderly people

(68% of them from the central Portugal, 24% from the north, and 8% from
Lisbon and Tagus valley). Almost 40% were women (n = 15), and 60.5%
(23) were men. For the most part, subjects were between 65 and 74
years old (n = 31; 81.6%), although a quarter were older (18.4% were 75 to
84 years old) or much older (7.4% were aged 85 years or more). For this
sample, 55.3% (21 elders) came from urban areas, whereas 44.7% (17
subjects) came from the country.
3. Inter-rater reliability was determined by administrating the instruments
Downloaded by [b-on: Biblioteca do conhecimento online UEvora] at 18:48 16 December 2017

separately to a sample of 15 subjects (presented below) who were eval-

uated three times within a 12-day interval by different interviewers).
This pre-test was applied to a sample of 15 elderly people (66.7%
between 65 and 74 years old, 33.3% from 75 to 84 years old). In terms of
gender, 80% (n = 12) were female and 20% were male. Almost two-
thirds lived in an urban area, and a third (n = 10) lived in a rural area.

Fourth Step: Normative Values

DISCRIMINANT ANALYSIS

Cut-off points were established based on the mean values obtained for the
general population sample (the 660 elders noted above, in the first proce-
dure of the third step). Notice that almost all of them (656 elders) fully
completed the questionnaire, and so that the age/gender distribution, as
well as the residential characteristics, were analogous to the whole popula-
tion sample (n = 660).

CLINICAL VALIDATION

This step served as the final assessment procedure for establishing norma-
tive values. The clinical validation of cut-off points derived from the statisti-
cal analysis was performed through diagnostic interviews conducted on a
subsample of 20 clinically depressed elderly patients (50% were men and
50% women; 30% were 65 to 74 years old; 50% were 75 to 84 years old, and
20% were age 85 years or more; almost equal groups were from urban and rural
areas) In this subsample, a depression diagnosis (according to ICD diagnostic
criteria) was confirmed for 14 of these patients, whereas 6 subjects had no
confirmed diagnosis.

RESULTS

From the results of the first pre-test (following item translation and retrover-
sion) conducted on the above mentioned sample of 200 elderly people, it
 Geriatric Depression Scale for Portuguese Elders 229

was possible to conclude that 3 of the 30 items of the original scale (items
27, 29, and 30) showed no internal consistency and that their positive corre-
lation with each one of the scale’s dimensions was always less than .3 (eval-
uated by means of Cronbach’s alpha and factorial analysis, respectively).

Concurrent Validity
After this first pre-test, the concurrent validity analysis of the GDS was then
Downloaded by [b-on: Biblioteca do conhecimento online UEvora] at 18:48 16 December 2017

performed using the Hamilton Rating Scale for Depression. The result of this
analysis revealed a correlation score of .83 among the 30 elderly people
assessed, a value that indicates a high Pearson correlation coefficient
(Bryman & Cramer, 1993; Pestana & Gajeiro, 2000).

Factorial Validity
Since current mathematical thinking about statistics does not consider factor
analysis using the principal component method to be the most appropriate
procedure for dealing with dichotomous scales of measure, the Bartlett
spherecity test and the Kaiser-Meyer-Olkin (KMO) test were performed in
this study. The Bartlett test revealed a .000 level of significance (p < .05) and
a KMO value of .932, indicating that factor analysis seemed to be highly
adjusted to GDS scores.
After defining the type of analytic procedure to use it was then impor-
tant to decide on the most suitable method to perform it. Based on calcula-
tion of the anti-image matrix (where sample adjustment values—MSA
[Measures of Sampling Adequacy] inscribed on the diagonal were rather
high whilst their anti-image was low) principal component analysis was
chosen (Table 1).

Factor Analysis—Principal Component

Factor analysis, after entering a varimax rotation, revealed three distinct factors:

• Factor 1 explaining 20.4% of the variance and reflecting well-being/ill-being,

• Factor 2 explaining 12.3% of the variance and reflecting sad mood,
• Factor 3 explaining 10.7% of total variance.reflecting mental & physical
problems

Reliability
TEMPORAL CONSISTENCY—TEST-RETEST

The instrument was administered twice to 38 elderly people, within an

8-day interval. The relationship between the two testing periods shows a
 230 M. T. S. Pocinho et al.

TABLE 1 GDs Principal Component Matrix

(Rotated by Varimax Method Using Kaiser
Normalization)

Factors

1 2 3

GDS 1 .766
GDS 7 .752
GDS 9 .737
Downloaded by [b-on: Biblioteca do conhecimento online UEvora] at 18:48 16 December 2017

GDS 19 .667
GDS 15 .643
GDS 10 .634
GDS 3 .613
GDS 5 .606
GDS 17 .551 .337
GDS 16 .551 .404
GDS 22 .435 .314
GDS 21 .422 .376
GDS 2 .391 .380
GDS 24 .649
GDS 11 .648
GDS 25 .621
GDS 6 .321 .604
GDS 13 .571
GDS 8 .492 .362
GDS 4 .476 .480
GDS 18 .477
GDS 20 .674
GDS 28 .572
GDS 14 .550
GDS 12 .545
GDS 26 .532
GDS 23 .363 .409
Note: Output criteria was selected in such a way
that correlations <.3 were not considered.

high temporal consistency for GDS, whether measured by Pearson correla-

tion coefficient (0.995) or by Intraclass Correlation Coefficient - (0.979)

INTERNAL CONSISTENCY—CRONBACH’S ALPHA AND KUDER-RICHARDSON

The global reliability study revealed rather high reliability (α =. 906; K-R = .907).

INTER-RATER RELIABILITY—TEST-RETEST

The inter-rater reliability was assessed through means of a three-part testing

procedure conducted in a sub-sample of 15 elderly people, within a 12-day
interval, by different interviewers. The results of repeated measures analysis
are listed in Table 2.
 Geriatric Depression Scale for Portuguese Elders 231

TABLE 2 GDS Inter-Judge Confidence—Repeated Measures

95% Confidence interval

Raters N M SD Lower bound Upper bound

1 15 8.800 .200 8.371 9.229

2 15 8.867 .133 8.581 9.153
3 15 8.867 .133 8.581 9.153
Total 45 8.844 .085 8.662 9.027
Downloaded by [b-on: Biblioteca do conhecimento online UEvora] at 18:48 16 December 2017

Note: Wilks’ Lambda = .994 p > .05.

One can see that inter-rater results are concordant (p > .05). Since a
Lambda value of .961 is a good indicator (values close to zero indicate dif-
ferences) GDS demonstrated good stability, allowing it to be administrated
by different raters without significant effect on the results.
Another concordance assessment test is Cohen’s kappa. According to
Pestana & Gajeiro (2000) the level of observed concordance is most accu-
rately measured by Cohen’s Kappa. After establishing the initial normative
data the results were divided by the cut-off point 10, and a Kappa calcula-
tion was then performed. Results are shown in Table 3.
GDS shows a quite high concordance value between the two assess-
ments (0.87), regardless of who administers the scale.
After performing all the statistical analysis discussed above, a 27-item
final version of the GDS was developed (items of GDS Portuguese language
version are listed in the Appendix).

Normative Values
The GDS mean score for our sample was 9.6. However, this score decreased to
4.9 when three conditions were observed: 1) good family relationship; 2)
age under 74 years; 3) absence of polypharmacy. Conversely mean depres-
sion scores increase considerably (up to 15.9) whenever family relationship

TABLE 3 GDS Inter-Judge Confidence—Cohens’ Kappa

Retest

Absence Presence
of depression of depression Total

Test
Absence of depression 23 2 25
Presence of depression 1 19 20
Total 24 21 45
Value Asymp. SE Approx. T
Agreement: Kappa 0.866 0.075 5.813***
N of Valid Cases 45
*** P < .0001.
 232 M. T. S. Pocinho et al.

was bad or of a poor affective quality, when age was greater than 74 years,
or when polypharmacy was present.

Discriminant Analysis
To establish cut-off points based on mean values obtained for the study
sample, a presentation of more than 10 symptoms (mean value, 9.6) was
used as the cut-off point between absence and presence of depression.
Downloaded by [b-on: Biblioteca do conhecimento online UEvora] at 18:48 16 December 2017

Using this cut-off point, 4 out of 10 of the elders in the study (40.2%) were
considered positive for depression.
To further verify the accuracy of our psychometric hypothesis, a
discriminant function analysis test was performed. In fact, assuming an
equal probability for absence or presence of depression among the elderly,
a theoretical cut-off point of 10 out of 27 symptoms does not statistically
mean depression is clinically present. Nevertheless, such a speculative
assumption was insufficient to validate our hypothesis. Further validation
was achieved through calculation of results obtained against group size
(meaning a .598 probability of not having depression and a .402 probability
of having depression). Results are displayed in Table 4.
Assuming that 59.8% of the subjects were not depressed, while 40.2%
were depressed, discriminant analysis revealed that 96.3% of the sample
was correctly assessed. The statistical cut-off point was then established by
calculating the contingency between predictive group values and global
GDS values. The results of the discriminant analysis indicate a predictive
cut-off point of 11 of 27 symptoms. However, in order to achieve the most
coherent decision possible, these results were compared with clinically vali-
dated cut-off points.

Clinical Validation
To accomplish such a task, GDS was administrated to elderly patients by a
psychiatrist in private practice who performed a diagnostic interview using

TABLE 4 Cut-off Point’s Validation*—Prior Probabilities Computed for Group Size

Predicted group membership

Absence Presence
Cohort of depression of depression Total

Validation n Absence of depression 392 0 392

Presence of depression 24 240 264
% Absence of depression 100 0 100
Presence of depression 9.1 90.9 100
Note: *96.3% of original grouped cases correctly classified.
 Geriatric Depression Scale for Portuguese Elders 233

TABLE 5 Normative Data & GDS Diagnostic Value

Clinical diagnostic (gold standard)

Presence Absence
of depression of depression Total

GDS
Presence of Depression 14 1 15
(>11 symptoms)
Absence of Depression 0 5 5
Downloaded by [b-on: Biblioteca do conhecimento online UEvora] at 18:48 16 December 2017

(≤ 11 symptoms)
Total 14 6 20

PE IC− IC+
S 100% 1 1
E 83% 0.5 1
VPP (PPV) 93% 0.8 1
VPN (PNV) 100% 1 1
FP 17% 0 0.3
FN 0% 0 0
A 95% 0.9 1

ICD diagnostic criteria (Ballone). This clinical diagnosis of depression using

ICD diagnostic criteria was used as the “gold standard” for GDS accuracy
analysis in our study.
The results were then calculated following two types of procedures:

• The first procedure was the one proposed by Fisher. Using this procedure,
scores obtained in the GDS were calculated using Fisher’s formula1 in order to
determine the scale’s cut-off point (P)2. A cut-off point of 11 was obtained.3
• The second procedure comes from epidemiologic studies and can be
applied whenever there is need for clinical confirmation of dimensional
diagnostic data generated by psychometric instruments.

The results from both procedures are shown in Table 5.

Based on the above analyses, it was possible to conclude that if 11 is
used as the cut-off point, the GDS was 95% accurate in detecting a depres-
sive disorder (compared with gold standard of clinical assessment by a clini-
cian) and 83% accurate in ruling out depression. These findings support the
choice of 11 as the cut-off point for establishing a depression diagnosis
when using the GDS as screening instrument.

CONCLUSION

The Portuguese language version of the GDS (with its 27-item scale struc-
ture) has high psychometric accuracy as a screening instrument to detect
 234 M. T. S. Pocinho et al.

depressive disorder among elders, both in the general and clinical popula-
tion settings. Based on the data obtained in this study regarding its factorial
structure, internal and temporal consistencies, and its inter-rater reliability,
the GDS appears to be a highly valid and reliable test for use in screening
Portuguese elders for depression.
The normative values identified in this study strongly suggest that the
identification of 11 or more symptoms out of 27 in this version of the GDS
constitute a robust criterion for the identification of depressive disorder
Downloaded by [b-on: Biblioteca do conhecimento online UEvora] at 18:48 16 December 2017

among the elderly attending primary care facilities in Portugal. These find-
ings indicate that the GDS can be a very useful screen, especially since
depressive disorders are considered to be a priority public health problem
by the World Health organization (WHO).

NOTES
(x1 + s 1 )+ (x2 − s2 )
1. P =
2
2. In the utilization of this formula we must bear in mind that c1 < c2 ; so in this case c1 = median
value for general population individuals and c 2 = median value for depressed individuals.

3. (9.6 + 6.8 ) + (12.7 − 6.2)

P = = 11.25.
2

REFERENCES

Alexopoulos, G., Bruce, M. L., Hull, J., Sirey, J., & Kakuma, T. (1999). Clinical deter-
minants of suicidal ideation and behavior in geriatric depression. Archives of
General Psychiatry, 56 (11): 1048–1053.
Ballone C.G., (2005). ICD10. Classificação internacional das doenças [On-line].
Available at http://www.psicologia.com.pt/instrumentos/dsm_cid/cid.php.
Brink, T.L., Yesavage, J.A., Lum, O., Heersema, P., Adey, M.B., & Rose, T.L. (1982).
Screening tests for geriatric depression. Clinical Gerontologist, 1, 37–44.
Bryman, A., & Cramer, D. (1993). Análise de dados em Ciências Sociais—
Introdução às Técnicas Utilizando o SPSS. Oeiras, Portugal: Celta.
Burns, A., Beevor, A., Lelliott, P., Wing, J., Blakey, A., Orrell, M. (1999). Health of
the Nation Outcome Scales for elderly people. British Journal of Psychiatry,
174, 435–438.
Burns, A., Dening, T., & Baldwin, R. (2001). Mental health problems. British
Medical Journal, 322, 789–791.
Caine, E.D., & Conwell, Y. (2001) Suicide in the elderly. International Clinical
Psychopharmacology, 16(Suppl 2), S25–S30.
Conwell, Y., & Pearson, J.L. (2002) Suicidal behaviors in older adults. American
Journal of Geriatric Psychiatry, 10(4), 359–361.
González Felipe, M. A. (1988). Características técnicas y modifications introducid en
dos scales de depression: Estudio pilot para la baremación y adaptación de la
 Geriatric Depression Scale for Portuguese Elders 235

GDS y el BDI a la población geriátrica española. Revista Iberoamericana de

Geriatría y Gerontología, 4, 235–240.
González Felipe, M. A., & Szurek Soler, S. (1990). La Geriatric Depression Scale
(GDS). In I. Montorio & M. Izal, The Geriatric Depression Rating Scale: a
review of its development and utility. International Psychoageriatrics, 8,
103–112.
Hamilton, M. (1967). Development of a rating scale for primary depressive illness.
British Journal of Social and Clinical Psychology, 6, 278–296.
ISD - Information and Statistics Division (2002). The co-occurrence of mental health
Downloaded by [b-on: Biblioteca do conhecimento online UEvora] at 18:48 16 December 2017

and substance use problems in Scotland: National data sources.

Karel, M.J., Ogland-Hand, S., Gatz, M., & Unuetzer, J. (2002). Assessing and treating
late-life depression. New York: Basic Books.
Koenig, H.G., George L.K. (1998). Depression and physical disability outcomes in
depressed medically ill hospitalized older adults, American Journal of Geriatric
Psychiatry, 6(3), 230–247.
Koneig, H.G., Meador, K.G., Cohen, H.J., & Blazer, D.G. (1988). Self-rated depres-
sion scales and screening for major depression in the older hospitalized patient
with medical illness. Journal of the American Geriatric Society, 36, 699–706.
Mavis Evans, M., & Pat Mottram, P. (2000). Diagnosis of depression in elderly
patients. Advances in Psychiatric Treatment, 6, 49–56.
Mcsweeny, A.J., Creer, T.L. (1995). Health-related quality of life assessment in
medical care. Disease-A-Month, 41, 5–71.
Montorio, I., Izal, M. (1996) The Geriatric Depression Rating Scale: a review of its
development and utility. International Psychoageriatrics, 8, 103–112.
Mui, A. (1996). Geriatric Depression Scale as a community screening instrument for
elderly Chinese immigrants, International Psychogeriatrics, 8(3), 445–458.
Nitcher, R.L., Burke, W.J., Roccaforte, W.H., Wengel, S.P. (1993). Collateral source
version of the Geriatric Depression Rating Scale. American Journal of Geriatric
Psychiatry, 1(2), 143–152.
Paul, I.A., Skolnick, P. (2003). Glutamate and depression: clinical and preclinical
studies. Annals of N.Y. Academy of Sciences, 1003, 250–272.
Pestana, M.H., Gageiro, J.N. (2000). Análise de Dados para Ciências Sociais—A
Complementaridade do SPSS, 2nd (Revista e aumentada). Lisbon: Sílabo.
Segal, Z.V., Pearson, J.L., & Thase, M.E. (2003) Challenges in preventing relapse in
major depression - Report of a National Institute of Mental Health Workshop
on state of the science of relapse prevention in major depression. Journal of
Affective Disorders, 77(2), 97–108.
Sinclair, Paula A., Lyness, Jeffrey M., King, Deborah A., Cox, Christopher, & Caine,
Eric D. (2001). Depression and self-reported functional status in older primary
care patients. American Journal of Psychiatry, 158, 416–419.
Stiles P.G., & McGarrahan J. F. (1991). The Geriatric Depression Scale: a compre-
hensive review, Journal of Clinical Gerontopsychology, 8(4), 89–110.
WHO. (1987). Mental Disorders: Clinical descriptions and diagnostic guidelines. In
Tenth revision of the international classification of diseases (pp. 179–191).
Geneva: World Health Organization.
World Health Organization. (2001). Suicide: International rates [online]. Available
at http://www.who.int/mental_health/topic_suicide/suicide1.html
 236 M. T. S. Pocinho et al.

Williamson, G.M., Shaffer, D.R., & Parmelee, P.A. (2002). Physical illness and
depression in older adults. New York: Kluwer Academic/Plenum Publishers.
Yesavage, J.A., Brink, T.L., Rose, T.L., Lum, O., Huang, V., Adey, M.B., & Leirer V.O.
(1983). Development and validation of a geriatric depression screening scale: A
preliminary report. Journal of Psychiatric Research, 17, 37–49.
Zilenovski, A. (1991) Geriatric Depression Scale: A Portuguese translation. Clinical
Gerontologist, 10, 84–86.
Downloaded by [b-on: Biblioteca do conhecimento online UEvora] at 18:48 16 December 2017

Appendix: GDS, Portuguese Version

GDS Sim Não

1 - Está satisfeito com a sua vida actual 0 1

2 - Abandonou muitas das suas actividades e interesses 1 0
3 - Sente que a sua vida está vazia 1 0
4 - Anda muitas vezes aborrecido 1 0
5 - Encara o futuro com esperança 0 1
6 - Tem pensamentos que o incomodam e não consegue afastar 1 0
7 - Sente-se animado e com boa disposição a maior parte do tempo 0 1
8 - Anda com medo que lhe vá acontecer alguma coisa má 1 0
9 - Sente-se feliz a maior parte do tempo 0 1
10 - Sente-se muitas vezes desamparado ou desprotegido 1 0
11 - Fica muitas vezes inquieto e nervoso 1 0
12 - Prefere ficar em casa, em vez de sair e fazer outras coisas 1 0
13 - Anda muitas vezes preocupado com o futuro 1 0
14 - Acha que tem mais problemas de memória do que as outras pessoas 1 0
15 - Actualmente, sente-se muito contente por estar vivo 0 1
16 - Sente-se muitas vezes desanimado e abatido 1 0
17 - Sente que, nas condições actuais, é um pouco inútil 1 0
18 - Preocupa-se muito com o passado 1 0
19 - Sente-se cheio de interesse pela vida 0 1
20 - Custa-lhe muito meter-se em novas actividades 1 0
21 - Sente-se cheio de energia 0 1
22 - Sente que para a sua situação não há qualquer esperança 1 0
23 - Julga que a maior parte das pessoas passa bem melhor do que o senhor 1 0
24 - Aflige-se muitas vezes por coisas sem grande importância 1 0
25 - Dá-lhe muitas vezes vontade de chorar 1 0
26 - Sente dificuldade em se concentrar 1 0
28 - Evita estar em locais onde estejam muitas pessoas (reuniões sociais) 1 0
Note: Excluded items: 27, 29, 30.