Professional Documents
Culture Documents
2002 Yesavage
2002 Yesavage
2002 Yesavage
SUMMARY
Objective To estimate the predictive value of the 30-question Geriatric Depression Scale (GDS) in Spanish and calculate
the most adequate cut-point for its use in Primary Health Care consultations.
Method 218 patients over the age of 64 treated at three health centers of Area 10 in Madrid were selected. In the ®rst
phase, the subjects completed the GDS, the Mini-Mental State Examination (MMSE) and a questionnaire on health and
socio-demographic variables. They were later interviewed using the Geriatric Mental Schedule (GMS), used as the gold
standard by doctors who were unaware of the results of the GDS. Two categories were contemplated according to the results
of the GMS: cases of depression (diagnosis of psychotic or neurotic depression) and non-psychiatric cases (no psychiatric
diagnosis, although isolated symptoms could be present).
Results 192 aged subjects were interviewed using the GDS and the GMS. Of these, 103 were considered `non-cases of
depression' and 60 others made up the `cases of psychotic/neurotic depression' group. For the most effective cut-point
(9/10), sensitivity was 86.7% and speci®city 63.1%. Considering a prevalence of depression of 30%, the predictive value
for positives was 50.2% and for negatives 91.7%. The Cronbach alpha coef®cient was 0.82, and the area below the ROC
curve obtained was 0.85. Those patients with cognitive deterioration had a mean GDS score similar to those that did not
present deterioration (11.16 vs 10.52; p > 0.05).
Conclusions The Geriatric Depression Scale is valid as a screening test in Primary Care consultations due to its high sen-
sitivity and negative predictive value. The most effective Spanish GDS cut-point (9/10) is lower than that obtained in the
original English version (10/11). Copyright # 2002 John Wiley & Sons, Ltd.
INTRODUCTION
*Correspondence to: M. I. FernaÂndez-San MartõÂn, Area de
Medicina Preventiva y Salud PuÂblica, Universidad de AlcalaÂ, Depression is the most frequent psychiatric problem in
Campus Universitario-Facultad de Medicina, Ctra. Madrid-Barce- the aged. Its high prevalence has been revealed by stu-
lona, Km. 33,600. E-28871, Alcala de HenaresÐMadrid, Spain.
Tel: 34 1 8854569 34 1 8854532. Fax: 34 1 8854874. dies carried out in the general population (Dewey et al.,
E-mail: maribel.fernandez@uah.es 1993; Devanand, 1994; FernaÂndez et al., 1995; CerdaÂ
Contract/grant sponsor: Health Ministry (Fondo de InvestigacioÂn
et al., 1997) as well as in patients treated in primary
Sanitaria). care (Evans and Katona, 1993; Marwijk et al., 1994;
Contract/grant number: 98/0726. Gottfries et al., 1997; Barry et al., 1998).
Received 2 July 2001
Copyright # 2002 John Wiley & Sons, Ltd. Accepted 30 October 2001
280 m. i. fernaÂndez-san martõÂn et al.
The detection of depression is dif®cult for primary calculating the most adequate cut-off point for its
care doctors for various reasons, including the insuf- use in primary health care consultations.
®cient training of the physician in mental health and
the scarce amount of time available to each patient in
the consultation (Goldberg and Huxley, 1980; METHOD
Van Hemert et al., 1993). Also, depression in the
elderly is revealed by symptoms that can be confused The present study was performed in Area 10, public
with problems of the aging process itself (slowness, health and administrative zone of the Spanish
insomnia), progressive cognitive deterioration (loss National Health System. Area 10 is located in the
of memory and attention span, disorientation) or phy- south of the Community of Madrid and covers some
sical diseases for which the aging patient consults his/ 235,000 inhabitants. The study population are persons
her physician (pain, constipation), symptoms which over the age of 64 assigned to three health centers of
would more probably be attributed to depression in Area 10 in Madrid (5,000 aged). Those elderly
an adult (Yesavage and Brink, 1983). This compli- patients treated outside the centers, such as those
cates the diagnosis of mental diseases in general in who are handicapped or in rest homes, were excluded.
seniors (Kanowski, 1994). We also excluded `displaced' patients, being those
Screening tests can be useful for the early detection that live less than four consecutive months at a resi-
of depression in Primary Care. However, due to the dence corresponding to one of the health centers.
aforementioned particularities of this disease in the We systematically selected 218 persons over the
aged, the tests used in patients under the age of 65 age of 64 who were treated at the health center under
are not valid for those over this age, thus speci®c tests the `on demand' method of consultation (consulta-
have been designed for the aging population. There are tions solicited by the patients themselves, thus
various tests that have been validated in an older popu- appointments ordered by the doctor were also
lation, such as the Zung Self-Rating Scale for Depres- excluded) during the months of February, March
sion or SDS (Zung, 1965) and the Geriatric and April of 1998.
Depression Scale or GDS (Brink et al., 1982; Yesavage The following variables were gathered:
and Brink, 1983). The latter has been translated into a
number of languages and used at various levels of Descriptive variables
health assistance (primary care, geriatric clinics, hos-
pitals) and in community studies (Montorio and Izal, Date of birth, marital status, sex, academic studies
1996). The authors speci®cally designed the scale and social class according to occupation and position
for the elderly, and as such the scale uses a Yes/No for- held in said occupation: ordinal variable ranging
mat to simplify its completion. They also excluded between classes l and V (Alvarez-Dardet et al., 1995).
from the symptoms those which could be confused
with somatic diseases or pseudo-dementia. Variables related with state of health
Different studies of the adaptation of the GDS scale
to Spanish have been written. Its predictive value has (1) Psychopharmaceuticals prescribed in the last six
been assessed in patients treated in a psychiatric months. Information obtained from the patient's clini-
department (Ramos et al., 1991), in a small group cal records; (2) health as perceived by the patient; (3)
of elderly residents in a home (Lobo et al., 1990), presence of mental disease at any time during the sub-
and its psychometric characteristics have been studied ject's life, according to information given by the
in community elderly subjects (Salamero and Marcos, patient; (4) sensory alterations: vision perceived (with
1992). In addition, other studies evaluate the psycho- glasses if used); hearing perceived (with implant if
metric behavior of the scale in aged rest home resi- used) and (5) Cognitive deterioration evaluated by
dents, without any modi®cations to its structure means of the MMS (Folstein et al., 1975) and adapted
(Izal and Montorio, 1993) and adding two more for Spanish by Lobo et al. (1979). The cut-point used
answer options (PeÂrez et al., 1990). was 23/24 (sensitivity: 85% and speci®city: 82%).
Given that none of the previously mentioned ver-
sions have been validated in a primary care setting
Measurement of depression
and that GDS is frequently used at this level, the pre-
sent study was proposed, with the objective of esti- The screening test used was the GDS and its results
mating the predictive value of the 30-question were compared with those obtained with the GMS,
Geriatric Depression Scale (GDS) in Spanish and which was used as a gold standard.
Copyright # 2002 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2002; 17: 279±287.
spanish version of the gds 281
Geriatric Depression Scale GDS Brink et al. (1982) interviewed the patient and completed the GDS and a
and Yesavage and Brink (1983). This scale is questionnaire that included the variables described.
comprised of 30 questions with a dichotomous After a few days (2 weeks maximum) the patient
YES/NO response which is evaluated as either 1 or was given another appointment and, in this second
0. A response with a value of 1 indicates the presence phase, doctors trained in psychiatry performed
of a depressive symptom. There are 10 inverted the GMS interview. These doctors were trained in
questions in which a negative response is scored as 1. the use of GMS by a psychiatrist; they were also
The validation of the 30-question version obtained a unaware of the result obtained by the patient in the
sensitivity of 80% and a speci®city of 100% (cut- GDS.
point 10/11). The original English version was The values for Sensitivity (S) and Speci®city (Sp)
translated into Spanish in this study by two bilingual were calculated for various cut-points. The Positive
health professionals. Predictive Values (PPVs) and Negative Predictive
Values (NPVs) were obtained by means of the Bayes
Geriatric Mental Schedule (GMS). Standardized Theorem, considering various values of prevalence.
clinical interview that con®rms the diagnosis of a Likewise, the ROC curve was drawn, calculating the
psychiatric case. Created by Copeland et al. (1976) area below the curve, and the internal consistency of
and Gurland et al. (1976) and adapted for Spanish the GDS was evaluated with the Cronbach alpha coef-
(Lobo et al., 1988; Saz, 1991). ®cient. Con®dence intervals (CI) were calculated for
These tests were complemented with the AGECAT the measurements up to 95%.
computerized diagnostic program (Copeland et al., A multiple logistic regression analysis was carried
1986; Dewey and Copeland, 1986). The interview out in order to adjust the classi®cation of the GDS
consists of 286 questions referring to the month prior with the MMS cognitive deterioration test, meaning
to the interview, exploring the presence of 152 symp- that we checked whether the S and Sp values of the
toms. The diagnostic decisions which the programs GDS were modi®ed with the results of the MMS test.
contemplates are the following: no diagnosis, organic The dependent variable is the classi®cation as a case
psychosis, functional psychosis, hypomania, psycho- of depression/non-psychiatric case in accordance with
tic depression, others psychoses, obsessive neurosis, the GMS interview. Two independent variables were
phobic neurosis, hypochondria, neurotic depression, included: the classi®cation from the GDS (with the
state of anxiety. Psychotic and neurotic depression most effective cut-point) and the classi®cation from
joined in the same group follow the diagnostic criteria the MMS (with the 23/24 cut-point). The statistical
for depression of the DSM-III classi®cation. In addi- signi®cance of the coef®cients was evaluated by
tion to the diagnoses, a Case Probability Index (CPI) means of the Wald test.
is also generated, which is scored from 1 to 5 (no
symptoms, mild symptoms, threshold case, moderate
RESULTS
case and serious case). The psychiatric case is de®ned
starting from a score of 3, which is considered the The percentage of patients who refused the GMS psy-
threshold level. In the study carried out by Copeland chiatric interview after having answered the GDS was
et al. (1986) and Dewey and Copeland (1986), the 11.9% (26/218). The socio-demographic and clinical
authors demonstrated that the inter-observer reliabil- characteristics of the aged subjects, including those
ity between the AGECAT diagnoses and those given who refused the GMS as well as those who did not,
by the psychiatrists that evaluated the same subjects are presented in Table 1. No statistically signi®cant
was 0.80 for depression, 0.88 for organic deteriora- differences were found between the two groups.
tion and 0.74 for the diagnosis of lack of psychiatric The seniors interviewed (192) were mostly women
pathology. (63.5%), 39.5% were 75 years of age or older, 65.6%
In this study, the cases were divided into two cate- were married, an important percentage had received
gories according to the results of the GMS: cases of no education (64.2%) and the majority had been
depression (aged patient with diagnosis of psychotic unskilled or skilled manual workers (87.7%).
or neurotic depression, level 3 or higher on the CPI) 47.1% of the subjects described their state of health
and non-psychiatric cases (aged patient without any as being average, more than half reported having suf-
psychiatric diagnosis, although isolated symptoms fered insomnia, and around 40% reported having suf-
may be present; level 2 or lower on the CPI). fered depression or anxiety at some time in their lives.
The interviews took place in the health centers. In According to the corresponding clinical records,
an initial phase, medical residents of Family Medicine 27.4% had taken anxiolytics in the previous six
Copyright # 2002 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2002; 17: 279±287.
282 m. i. fernaÂndez-san martõÂn et al.
Sex
Male 70 36.5 6 23.1
Female 122 63.5 20 26.9
Age (in years)
65 to 69 56 30.3 3 12.5
70 to 74 56 30.3 8 33.3
75 to 79 41 22.2 8 33.3
over 80 32 17.3 5 20.8
Marital status
Single 4 2.1 1 3.8
Married 126 65.6 15 57.7
Separated/divorced 4 2.1 Ð Ð
Widowed 58 30.2 10 38.5
Academic studies
Illiterate 26 13.7 4 15.4
Can read and write, no studies 96 50.5 13 50.0
Primary studies 63 33.2 9 34.6
Secondary studies 5 2.7 Ð Ð
Social class
I and II (Businessmen, managers) 3 1.6 Ð Ð
III (Skilled worker-non-manual) 20 10.7 2 8.0
IV (Skilled worker-manual) 79 42.2 9 36.0
V (Unskilled workers) 85 45.5 14 56.0
Self-perception of state of health:
Very good 9 4.7 Ð Ð
Good 65 34.0 14 53.8
Average 90 47.1 9 34.6
Bad 22 11.5 2 7.7
Very bad 5 2.6 1 3.8
Report having suffered the
following at some time in their
lives**
Insomnia 107 57.2 11 42.3
Depression 73 38.8 8 32.0
Anxiety 64 40.0 10 38.5
Psychopharmaceuticals in the
previous six months**
Anxiolytics 46 27.4 6 27.3
Antidepressants 22 13.1 3 13.6
Sensory alterations**
Visual 55 28.6 5 20.0
Auditory 27 14.1 4 16.0
Cognitive deterioration (MMS) 34 17.8 7 26.9
Mean GDS score (CI 95%) 10.67 (9.94 ±11.42) 10.87 (8.70±13.05)
months. According to the MMS, 17.8% of the patients and 22 (`Do you feel that your situation is hopeless?').
presented cognitive deterioration. The question-total correlation coef®cients of the scale
The mean score obtained on the GDS by the 192 were statistically signi®cant ( p < 0.01), except the
aged patients was 10.67 (CI: 9.94±11.42). The GDS coef®cients representing two questions: numbers 28
questions that were least answered (Table 2) were (`Do you prefer to avoid social gatherings?') and 29
numbers 5 (`Are you hopeful about the future?'), 20 (`Is it easy for you to make decisions?'). The mean cor-
(`Is it hard for you to get started on new projects?') relation for these items was 0.40 (range 0.04 ± 0.66).
Copyright # 2002 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2002; 17: 279±287.
spanish version of the gds 283
Table 2. Patients who did not answer each question of the GDS scale
GDS SCALE No.* Item-whole correlation**
*n 192.
**All the correlation coef®cients are statistically signi®cant ( p < 0.01), except the coef®cients corresponding to items 28 and 29.
So as not to exclude any patients from the analysis, the analysis of the validity of the GDS test, we there-
the score obtained by those who did not complete the fore compared the 103 non-psychiatric cases (75 plus
scale was adjusted to 30, taking into account the num- 28 with some symptoms but without diagnosis) with
ber of questions answered. This was carried out once the 60 cases of psychotic and neurotic depression
we checked that there were no statistically signi®cant (54 plus 6).
differences between the group of patients that Table 3 shows the GDS values for the aged patients
answered all the questions and the group that left diagnosed as cases of depression and the group
some questions incomplete, regarding the following without psychiatric pathology. The cases of depres-
variables: sex ( p 0.87), mean age ( p 0.39), aver- sion had a mean of 14.85 positive answers, while
age GDS score ( p 0.53) and cognitive deterioration the non-depression cases had a mean of 8.27
according to the MMS ( p 0.52). ( p < 0.001). 78% of the cases were in the threshold
Later, the psychiatric diagnoses obtained by means category of the case probability index and 62% of
of the GMS-AGECAT interview in these 192 elderly the non-cases had symptoms that did not constitute
subjects were the following: 75 did not receive any a psychiatric case. Women represented 86.7% of the
diagnosis, 27 were classi®ed with organic psychosis, cases and 51.5% of the non-cases ( p < 0.001).
two were assigned to the group of other psychoses, 54 The area below the ROC curve was 0.85 (CI: 0.79±
received a diagnosis of psychotic depression and six 0.91; p < 0.001 under the assumption of null hypoth-
of neurotic depression. The remaining 28 persons pre- esis that the true area is 0.5). The most effective cut-
sented symptoms for some diagnoses, but without points, taking into account the S and Sp values, were
reaching the threshold to be considered a case. For points 9/10 and 10/11 (Table 4). For the former, S was
Copyright # 2002 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2002; 17: 279±287.
284 m. i. fernaÂndez-san martõÂn et al.
Table 3. Comparison between cases and non-cases of depression according to the GMS interview
Total GMS depression case GMS non-depression case
n 163 n 60 n 103
Table 4. Characteristics of the GDS scale for the selected cut- the GDS and are nonetheless cases of depression, and
points true negatives were: 87.5% of women vs 43.1% of
Cut-point 9/10 Cut-point 10/11 women among the true negatives, and a mean age
of 82.10 years vs a mean age of 72.25 years among
Validation the true negatives. The mean score on the GDS was
Sensibility 86.7% (78.1±95.3%) 81.7% (71.9±91.5%)
Speci®city 63.1% (53.8±72.4%) 68.0% (59.0±77.0%)
also clearly higher, the score range in this group of
Positive predictive false negatives being between 7 and 9 points.
value for The false positives, meaning those which obtained
different prevalences: a score higher than 9 on the GDS and were not cases
15% 29.3% (19.9±38.7%) 31.1% (21.0±41.1%) of depression, differed from the true positives as fol-
20% 37.0% (27.0±47.0%) 39.0% (28.4±49.5%)
25% 43.9% (33.7±54.2%) 46.0% (35.2±56.8%) lows: 65.8% were women vs 86.5%, respectively. The
30% 50.2% (39.8±60.5%) 52.2% (41.4±63.1%) mean score obtained on the GDS was quite lower in
35% 55.9% (45.6±66.1%) 57.9% (47.2±68.6%) the group of false positives (12.67) than in the group
40% 61.0% (51.0±71.1%) 63.0% (52.5±73.4%) of true positives (15.89). The GDS score range for the
Negative predictive
value for
false positives was 10 to 17. 76.3% had psychiatric
different prevalences: symptoms on the GMS without reaching threshold
15% 96.4% (92.1±100%) 95.5% (90.9±100%) case level.
20% 95.0% (90.0±100%) 93.7% (88.4±99.0%) The internal consistency of the GDS, measured by
25% 93.4% (87.8±99.1%) 91.8% (85.8±97.8%) means of the Cronbach alpha coef®cient, was 0.82.
30% 91.7% (85.4±98.0%) 89.7% (83.0±96.3%)
35% 89.8% (82.9±96.7%) 87.3% (80.1±94.6%) An adjustment was made to the GDS (cut-point 9/
40% 87.7% (80.1±95.2%) 84.8% (77.0±92.6%) 10) with the MMS cognitive deterioration test, by
means of multiple logistical regression. The depen-
dent variable was the classi®cation as case/non-case
86.7% and Sp was 63.1%. For the latter cut-point, S according to the GMS interview. The result of the
decreases somewhat (81.7%) and Sp increases Wald test to evaluate the null hypothesis if the coef-
slightly (68.0%). If the test were applied in a popula- ®cient of the MMS variable is zero was not statisti-
tion with a prevalence of depressive disorders of 30%, cally signi®cant ( p 0.57). Therefore, there was no
PPV would be 50.2% and 52.2%, while NPV would relationship between the classi®cation made using
be 91.7% and 89.7%, for cut-off points 9/10 and the MMS and the result of the GMS, nor were the S
10/11, respectively. and Sp values of the GDS modi®ed according to the
Described below are the characteristics of the false results obtained on the MMS. The mean score of the
positives and negatives resulting from a cut-point GDS for the cases of cognitive deterioration was
value of 9/10 (Table 5). The differences between false 11.16 (CI: 9.58±12.74) and 10.52 for the non-cases
negatives, meaning those which scored less than 10 on (CI: 9.68±11.35).
Copyright # 2002 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2002; 17: 279±287.
spanish version of the gds 285
Table 5. Comparison between subjects correctly and incorrectly classi®ed by GDS (cut-point 9/10)
Negative Positive
TN FN TP FP
n 65 n8 n 52 n 38
Sex
Male 56.9% 12.5% 13.5% 34.2%
Female 43.1% 87.5% 86.5% 65.8%
p 0.03* p 0.02
Mean age (CI) 72.25 82.10 73.18 74.26
(70.86±73.64) (75.78±88.43) (71.41±74.94) (72.26±76.27)
p 0.001** NS
Mean GDS score (CI) 5.70 8.08 15.89 12.67
(5.13±6.28) (7.33±8.83) (14.73±17.06) (12.09±13.26)
p 0.004** p < 0.001
GDS score
0±1 4.5% Ð
2±3 15.4% Ð
4±5 26.1% Ð
6±7 30.8% 25.0%
8±9 23.1% 75.0%
10±14 42.3% 81.6%
15±19 32.7% 18.4%
20±24 23.1% Ð
25±30 1.9% Ð
Case probability index
No symptoms 46.2% Ð Ð 23.7%
Mild symptoms 53.8% Ð Ð 76.3%
Threshold case Ð 75% 78.8% Ð
Moderate case Ð 25% 21.2% Ð
Serious case Ð Ð Ð Ð
Copyright # 2002 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2002; 17: 279±287.
286 m. i. fernaÂndez-san martõÂn et al.
50%. However, given that the GDS would be used by Contemplating the results obtained, the current
the general practitioner as a screening tool, it is proposal of the GDS scale with a cut-point of 9/10
important that the instrument has high S and NPV should be used when applied in Primary Care consul-
in order to detect the majority of cases, albeit at the tations, given the difference in results in the various
cost of labeling as ill those patients that are not in rea- Spanish versions carried out in different environ-
lity. Those aged patients with positive results who ments.
therefore have a higher probability for depression
should be evaluated in depth by the general physician
ACKNOWLEDGEMENT
before initiating any treatment. The authors of the
GDS designed the scale so that it would be able to This work is part of a Project ®nanced by Health Min-
detect depression independently of the cognitive istry (Fondo de InvestigacioÂn Sanitaria) investigation
deterioration of the aged subject, thus those symp- grant 98/0726.
toms that could be indicative of both pathologies were
not included. In our study, the mean score of the GDS
was similar in patients with and without cognitive REFERENCES
deterioration; the same happened in the Burke et al. AlvaÂrez-Dardet C, Alonso J, Domingo A, Regidor E. 1995. La
study (1992); likewise, it has been reported that that MedicioÂn de la Clase Social en Ciencias de la Salud. SG Edi-
GDS is valid for patients with mild or moderate tores: Barcelona.
Andrade C, FernaÂndez San MartõÂn MI, Rivero M, Asiel A, Redondo
dementia (Feher et al., 1992). However, it must AM, GarcõÂa R. 1999. Deterioro Cognitivo y Nivel Educativo en
be pointed out that patients classi®ed as having pacientes ancianos de AtencioÂn Primaria. IX Congreso de la
cognitive deterioration can have high scores on the Sociedad MadrilenÄa Medicina Familiar y Comunitaria, 21±22
MMS due at least in part to a low educational level. mayo. Madrid, Spain.
Barry KL, Fleming MF, Manwell LB, Copeland LA, Appel S. 1998.
This fact was manifested in a study created by the Prevalence of and factors associated with current and lifetime
authors with these same patients (Andrade et al., depression in older adult primary care patients. Fam Med 30:
1999), as well as in the revalidation of the MMS scale 366±371.
by the author of the ®rst translation into Spanish Brink TL, Yesavage JA, Lum O, Heersema P, Adey M, Rose TL.
(Lobo et al., 1999). 1982. Screening tests for geriatric depression. Clin Gerontologist
1: 37±44.
The GDS was devised by its authors to be com- Burke WJ, Nitcher RL, Roccaforte WH, Wengel SP. 1992. A pro-
pleted by patients. However, it can be dif®cult for spective evaluation of the Geriatric Depression Scale in an out-
aged Spaniards to do so by themselves, as 14% of patient geriatric assessment center. JAGS 40: 1227±1230.
the aged subjects in our study were illiterate and Cerda R, LoÂpez-Torres J, FernaÂndez C, LoÂpez-Verdejo MA, Otero
A. 1997. DepresioÂn en personas ancianas. Factores asociados.
50% had received no education whatsoever. For this Aten Primaria 19: 12±17.
reason, it was the doctor who questioned the patient Copeland JRM, Kelleher MJ, Kellet JM, et al. 1976. A semi-
and completed the scale with the answers given. We structural clinical interview for the assessment of diagnosis
believe that the comprehension of the questions was and mental state in the elderly. The Geriatric Mental State Sche-
very high. Only those questions related to future pro- dule. I. Development and reliability. Psychol Med 6: 439±449.
Copeland JRM, Dewey ME, Grif®ths-Jones HM. 1986. Psychiatric
jects in the lives of the aged patients (numbers 5 and case nomenclature and a computerized diagnostic system
20) or number 22, which may re¯ect a more abstract for elderly subjects: GMS and AGECAT. Psychol Med 16:
idea, were more likely to be left unanswered despite 89±99.
the interviewers' insistence, most probably due to not D'Ath P, Katona P, Mullan E, Evans S, Katona C. 1994. Screening,
detection and management of depression in elderly primary care
having been understood. attenders. I: the acceptability and performance of the 15 item
What remains to be contemplated is a second ana- geriatric depression scale (GDS15) and the development of short
lysis of this scale to try to improve its validity starting versions. Fam Pract 11: 260±266.
with a reduction in the number of questions, taking Devanand DP. 1994. Is dysthymia a different disorder in the
into account the psychometric properties, as other elderly? Am J Psychiatry 151: 1592±1599.
Dewey ME, Copeland JRM. 1986. Computerized Psychiatric diag-
authors have already done (D'Ath et al., 1994). This nosis in the elderly: AGECAT. J Microcomput Appl 9: 135±140.
would lead to greater use of the GDS in Primary Care Dewey ME, de la CaÂmara C, Copeland JRM, Lobo A, Saz P. 1993.
consultations because less time would be needed for it Cross-cultural comparison of depression and depressive symp-
to be completed. Scales with fewer questions are toms in older people. Acta Psychiatr Scand 87: 369±373.
Evans S, Katona C. 1993. Epidemiology of depressive symptoms in
already being used in a multitude of languages (some elderly primary care attenders. Dementia 4: 327±333.
of the translations are available on Internet at Feher EP, Larrabee GJ, Crook TH. 1992. Factors attenuating the
Yesavage, 2001) and as a screening test in Primary validity of the Geriatric Depression Scale in a dementia popula-
Care (Gottfries et al., 1997). tion. JAGS 40: 906±909.
Copyright # 2002 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2002; 17: 279±287.
spanish version of the gds 287
FernaÂndez L, Villaverde ML, Gracia R, Morales CR, Morera A, RevalidacioÂn y estandarizacioÂn del Mini-Examen-Cognoscitivo
Fuente J. 1995. Estudio comunitario de prevalencia de trastornos (primera versioÂn espanÄola del Mini-Mental-Status-Examination)
depresivos en poblacioÂn geriaÂtrica. An Psiquiatr 11: 99±102. en poblacioÂn geriaÂtrica general. Med Clin (Barc) 112: 767±774.
Folstein MF, Folstein SE, McHugh PR. 1975. `Mini Mental State': Marwijk HV, Hoeksema HL, Hermans J, Kaptein AA, Mulder JD.
a practical method for grading the cognitive state of patients for 1994. Prevalence of depressive symptoms and depressive disor-
the clinician. J Psychiatr Res 12: 189±198. der in primary care patients over 65 years of age. Fam Pract 11:
Goldberg D, Huxley P. 1980. Mental Illness in the Community. 80±84.
Tavistock: London. Montorio I, Izal M. 1996. The Geriatric Depression Scale: a review
Gottfries GG, Noltorp S, Norgaard N. 1997. Experience with a of its development and utility. Int Psychogeriatr 8: 103±112.
Swedish version of the Geriatric Depression Scale in primary PeÂrez E, Gonzalez MA, Moraleda P, Szurek S, Gonzalez JA. 1990.
care centres. Int J Geriatr Psychiatry 12: 1029±1034. La Geriatric Depression Scale (GDS) como instrumento para la
Gurland BJ, Fleiss J, Goldberg K, et al. 1976. A semi-structural evaluacioÂn de la depresioÂn: bases de la misma. Modi®caciones
clinical interview for the assessment of diagnosis and mental introducidas y adaptacioÂn de la prueba a la realidad psicogeriaÂ-
state in the elderly. The Geriatric Mental State Schedule. II. A trica espanÄola. Rev Esp Geriatr Gerontol 25: 173±180.
factor analysis. Psychol Med 6: 451±459. Ramos JA, Montejo ML, Lafuente R, Ponce de LeoÂn C, Moreno A.
Izal M, Montorio I. 1993. Adaptation of the geriatric depression 1991. ValidacioÂn de la escalaÐcriba geriaÂtrica para la depre-
scale in Spain: a preliminary study. Clin Gerontologist 13: 83± sioÂn. Actas LusoÐEsp Neurol Psiquiatr 19: 174±177.
90. Salamero M, Marcos T. 1992. Factor study of the Geriatric Depres-
Kanowski S. 1994. Depression in the elderly: clinical considera- sion Scale. Acta Psychiatr Scand 86: 283±286.
tions and therapeutic approaches. J Clin Psychiatry 55: 166±173. Saz P. 1991. GMS-AGECAT: validacioÂn y estudio de su utilidad en
Lobo A, Escolar V, Esquerra J, Seva A. 1979. El Mini Examen Cog- la comunidad geriaÂtrica. [Tesis doctoral]. Universidad de Zara-
noscitivo: un test sencillo y praÂctico para detectar alteraciones goza: Zaragoza.
intelectivas en pacientes meÂdicos. Actas Luso-Esp Neurol Psi- Van Hemert AM, Hengeveld MW, Bolk JH, Rooijmans HG,
quiatr 3: 149±153. Vandenbroucke JPSO. 1993. Psychiatric disorders in relation
Lobo A, Saz P, Dia JL, GonzaÂlez JL. 1988. El Geriatric Mental to medical illness among patients of a general medical outpatient
State en poblaciones espanÄolas: validacioÂn de paraÂmetros orgaÂ- clinic. Psychol Med 23: 167±173.
nicos y afectivos. Actas de la XIII ReunioÂn de la Sociedad Espa- Yesavage JA, Brink TL. 1983. Development and validation of a ger-
nÄola de PsiquiatrõÂa BioloÂgica 333±339. iatric depression screening scale: a preliminary report. J Psichiat
Lobo A, Ventura T, Marcos C. 1990. Psychiatric morbidity among Res 17: 37±49.
residents in a home for theÐelderly in Spain: prevalence of dis- Yesavage JA. 2001. Geriatric Depression Scale. Available in:
order and validity of screening. Int J Geriat Psychiatry 5: 83±91. www:stanford:edu=~yesavage=GDS:html:
Lobo A, Saz P, Marcos G, Dia JL, de la Camara C, Ventura T, Mor- Zung WWK. A self rating depression scale. 1965. Arch Gen Psy-
ales Asin F, Fernando Pascual L, Montanes JA, Aznar S. 1999. chiat 12: 63±70.
Copyright # 2002 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2002; 17: 279±287.