Full text of DOI:10.1002/mpr.91

174 International Journal of Methods in Psychiatric Research, Volume 9, Number 4
The Galicia Study of Mental Health of the
Elderly II: the use of the Galician DIS
R. MATEOS, Department of Psychiatry, School of Medicine, University of Santiago de Compostela (USC) and
Psychogeriatric Unit, Conxo Psychiatric Hospital, University Hospital Complex of Santiago de Compostela
(CHUS), Spain
A. DROUX, Ministry of Health, Division for Treatment and Prevention of Drugdependency, Northern Regional
Directorate (Portugal)
M. PÁRAMO, CHUS, Santiago de Compostela, Spain
F. GONZALEZ, Obradoiro de Socioloxía, SL, Santiago de Compostela, Spain
I. CARRERA, Galician Plan on Addictive Drugs, Division for Assistance, Education and Research, Spain
J. MAZAIRA, CHUS, Santiago de Compostela, Spain
L. GUNTIN, P. GONZALEZ, Obradoiro de Socioloxía, SL, Santiago de Compostela, Spain
M.C. GARCÍA, A. RODRIGUEZ-LÓPEZ, CHUS and Department of Psychiatry, USC, Santiago de Compostela,
Spain
ABSTRACT Reports of epidemiological surveys do not always provide adequate careful descriptions of the methodology used
and the sociocultural context involved. Galicia, a natural region in south-west Europe, possesses certain characteristics that
differ from those in other communities within the Spanish state. Galician is a Romance language, which is, in fact, closer to
Portuguese than it is to Spanish. The population (2.7 million inhabitants) is widely dispersed, with two-thirds of them living in
the countryside. More than 18% are older than 65, and the educational level of this elderly population is low. A total of 681
persons over 60 years of age were interviewed in their homes using the DIS-III as the main diagnostic instrument during the
second phase of the Galicia Study of Mental Health of the Elderly. This paper analyses some methodological implications of
this epidemiological study, focusing on the translation of the DIS into Galician and its performance and acceptability in this
community. The possible cultural bias that can affect the rates of prevalence in some specific disorders, the limitations of life-
time prevalence and the advantages of using six-month prevalence rates is discussed. Besides some interesting anecdotes, which
are described in the article, the main result is the very high acceptance of DIS-III in this elderly population.
Key words: epidemiologic methods, mental disorders, health surveys, elderly, Diagnostic Interview Schedule
(DIS), translations, prevalence, Spain
Introduction
peculiar to Galicia, the methodology used in the field-
work and the estimate of prevalence rates. Here we
need only remember that the main objective was to
study the range of psychiatric pathology in the aged
segment of the population, covering all the social lev-
els of the natural community of Galicia in south-
western Europe, with a population of 2.7 million. The
instrument chosen to produce the DSM-III diagnoses
in the second phase of the study was the DIS-III
(Robins et al., 1981).
In this paper we describe the process of translating
the DIS into Galician, to what extent the DIS was
accepted by this community, and we will provide anec-
dotes of anthropological interest. We discuss the
reliability of lifetime prevalence rates and the possible
cultural bias concerning the operational definition of
The scientific status of psychiatric epidemiology caus-
es it to be viewed in an ‘objective’ way. This very
‘objectivity’ can make us forget the anthropological
side of man’s existence (Rodríguez-López, 1994). From
this comes the need to explain fully the influence that
sociocultural contexts can have on the source of epi-
demiological information. Unfortunately, with some
notable exceptions (for example, Eaton and Kessler,
1985) ‘published reports of survey results rarely pause
to describe the many problems and unforeseen events
that occurred during data collection’ (Kohout, 1992).
This paper continues our reflection on the method-
ology used in the Galicia Community Study of Mental
Health of the Elderly. Elsewhere (Mateos et al., 2000)
we describe the sociocultural characteristics that are

Galicia Study of Mental Health of the Elderly II 175
some diagnoses. As far as possible, we have backed up
our discussion by three kinds of empirical data: the
prevalence rates produced by the survey, the observa-
tions made by the interviewers at the end of each
interview and discussion groups dealing with difficul-
ties, and anecdotes.
them being a native Portuguese speaker (Carrera).
Afterwards, the Galician version was supervised by
another Portuguese (Droux), who speaks fluent English
and Spanish. Finally, this almost definitive version was
given a stylistic revision to adapt certain words to the
official normalized Galician, a job that was done by a
member of the survey research company (Guntín).
We consider that, to a great extent, the resulting
interview remained faithful to the spirit of the English
version. Nevertheless, in using instruments of psy-
chopathological research created in another language,
there is a risk of certain biases upon which it is necessary
to comment, especially considering the sociological
differences between the US and Galicia. The following
goes on to objectify and quantify the said possible bias.
From the beginning our idea was to translate the
DIS into Galician rather than into Spanish. The rea-
sons for this decision were:
Difficulties in translating the DIS into Galician,
with special references to the Portuguese and
several Spanish versions of the instrument
If all translation is, ultimately, an impossible task
(Straight, 1981; cited in Karno et al., 1985, p. 175 ), it
becomes even more difficult when we are dealing with
the translation of an interview that must enable people
to speak about something as personal as their
psychopathological experiences. The (imperfect)
bilingualism of Galician society constitutes an added
difficulty (Mateos et al., 2000), as has been found in
other Hispanic countries (Marcos and Trujillo, 1981,
cited in Karno et al., 1985). All these difficulties are
increased by the fact that, until a few years ago,
Galician, the mother tongue of the majority of the
older population, was considered an uncultured and
socially marginalized language.
In preparing the DIS-III we used all the versions in
Spanish that we could find as well the Portuguese one.
The Spanish versions that we consulted were not all
completely alike. Even though it is not the objective of
this paper, it seems opportune to point out that we
have found some limitations in the various Spanish
versions that we have examined (for example, an
excessively literal translation of the English text).
These make it difficult to understand the questions in
the Spanish that is spoken in our country. The
Portuguese version, created in 1984 by Droux (Droux
and Meireles, 1990) under the supervision of J. Helzer,
proved to be particularly useful because of the close
relation between Portuguese and Galician. This ver-
sion faithfully followed all the questions of the English
one, introducing only minor adaptations to preserve
the spirit of the original questions.
• both Galician and Spanish are spoken in Galicia,
but the elderly often use the former as a result of
their inferior educational level and because a
majority of them live in a rural environment in
which Galician predominates (Mateos et al.,
2000);
• in the case of the interviewee not understanding
the meaning of a specific term in Galician, it was
be easier for the interviewer to explain its meaning
in Spanish.
Finally, the laborious translation of the computer algo-
rithms of the DIS into the language of the statistics
package SPSS was carried out by a psychiatrist who is
an expert in computing (Páramo).
The performance and acceptability of the survey in
this community
General acceptance was high, which must be considered
within the context of the success of epidemiological
studies that had taken place in Galicia previously
(Rodríguez-López, 1993), and therefore it was relative-
ly easy to follow on from the ECA project (Munson et
al., 1985). In doing the interviews, some comparisons
were made between the first and second stage. The
GHQ-60 (Goldberg, 1970) proved much more difficult
for elderly people, even though it is shorter than the
DIS. The DIS was seen by the elderly as a kind of
medical interview in which the patient had the
opportunity to talk about his or her problems.
The translation into Galician took place in various
stages. First, a group of four researchers from the
Department of Psychiatry of the University of Santiago
de Compostela (Mateos, Carrera, Mazaira, Páramo)
compared each of the Spanish versions of the DIS with
the English and Portuguese ones. Three members of
this group had advanced fluency in English, and all four
of them understood written Portuguese, with one of

176 Mateos et al. (II)
Anecdotes
act of witchcraft. One asked us directly: ‘Are you
trying to put a spell on me?’ These examples
illustrate the profound roots that the belief in
witchcraft and magic have here in Galicia. They
are frequently used toward paramedical ends.
• The Venezuelan soap opera Cristal, called a ‘cule-
brón’ in our language, which literally means a
snake eating itself, became our inseparable compe-
tition in the after-lunch hours between 16h00 and
17h00 in the afternoon. The interviewers always
had to wait until the show was over as the old peo-
ple watched it with immense interest.
Now we would like to relate a series of anecdotes, with
the intention of sketching certain anthropological fea-
tures of the Galician people. These should not be
taken as general rules, but they give an anthropological
profile that provides the reader with some under-
standing of our culture, with reference to the so-called
‘deep Galicia’. All of these events took place in the
rural areas, the traditional reserve of the oldest
Galicians:
• One country woman agreed to be interviewed, but
on condition that it would be done in the early
evening when her husband would be taking the
cows out to pasture. The interview was going well
until it began to rain, causing the early return of
the husband, who, on seeing the interviewer
became very angry and shouted ‘What are you
doing with my wife?’ He then began to chase the
interviewer, trying to strike him with one of his
farm tools. This interview was eventually placed on
the refusal list. However, it was the only violent
incident that we encountered.
• In one very humble household, the interviewer, in
his first visit, came upon an old man, shabbily
clothed, in the company of his Mongoloid son. The
interview was conducted in a kitchen that was in a
lamentable state. The second time, the interviewer
was received in the same room, but which was now
completely unrecognizable as it had been cleaned
from top to bottom. Accompanying the old man
were all of his family members, dressed in their
Sunday finest, including his Mongoloid son (wear-
ing a bow tie) along with two village neighbours.
There is no doubting the extreme importance this
visit had for this rural family. It is also a good exam-
ple of Galician idiosyncrasy. In this case, the old
man had not accepted the first interview with
much enthusiasm, but once the initial resistance
had been overcome, he and everyone else in his
family did all they could to collaborate.
• In certain cases the subject invited the interviewer
to stay for dinner, and more than one even asked,
‘How much do I owe you?’
The interviewers perceived a difference from the first
phase, which could be classified as more ‘socially based’
and less revealing on a personal level. In the course of
the DIS the subjects would sometimes cry, and
the interviewers would be emotionally affected by the
interview all day. Many subjects confessed ‘I’ve never
told anybody as many things about myself as I have just
told you’. The climate created was generally one of
confidence and the interviewers felt that they had
been put in the role of being doctors.
Difficulties
Table 1 shows a summary analysis of the difficulties
reported by interviewers. These difficulties, of course,
differed in their degree of importance. In the majority
of cases we are referring to specific difficulties that
more than anything affect how the MMSE is carried
out. All in all, we believe that where some doubts have
emerged concerning the validity of the interviews it is
only with those subjects where severe deterioration
was evident (2.2%) and perhaps to some extent in
those individuals who were very suspicious in their
attitudes (3.7%). Of the remaining difficulties that
could be considered relatively serious, illiteracy was
the most important factor. It was reported to affect
16% of the total number of interviews (but the level of
functional illiteracy is much higher).
The fact that the interview was conducted in
Galician was seen by all the interviewers as extremely
positive. Only in very few cases, especially in an urban
environment, were major explanations in Spanish
necessary. With reference to the language used, the
interviewers also valued the fact that they themselves
were expert Galician speakers with the ability to
• Another subject continually rejected a second
interview – altogether on five different occasions –
and on one of them refused to even respond to his
name when we asked for him at a bar. The last time
the interviewer tried to talk to him, he jumped out
of his house from a window and hid in a nearby
cornfield.
• More than one person thought that asking them to
fold a piece of paper in half to do the MMSE was an

Galicia Study of Mental Health of the Elderly II 177
Table 1: Difficulties found by the interviewers while carrying
out the DIS
the subject would feel as if he or she were being cross
examined.
The ‘personality disorders’ section was more diffi-
cult to understand. This was the case with questions of
the following kind: ‘have you ever spanked or hit any
child, hard enough so that he or she had bruises or had
to stay in bed or see a doctor?’ or: ‘did you ever skip
school or play hooky at least twice in one year?’ The
first usually caused as much surprise as it did suspicion.
So far as the second was concerned, for elderly
Galicians living in a rural environment the normal
education they would have received was work in the
family’s fields instead of attending classes. In rural
Galicia 50 years ago, children’s education was a fre-
quent source of conflict; one of these being linguistic,
as Galician was suppressed at that time. As a result of
these conflicts, questions concerning education have a
tendency not to be very revealing.
1
a) General difficulties
Evident severe cognitive decline
Hearing difficulties
%
2.2
0.9
3.7
Distrust/suspicion
b) Specific difficulties (affecting the MMSE)
Lack of formal education
Sight problems
16.0
1.6
2.9
Problems with motor functions
TOTAL
One or more difficulties
More than one
22.0
5.0
1
Raw percentages of the 681 subjects interviewed using the
DIS.
maintain the ‘idiomatic rhythm’, an extremely impor-
tant factor when dealing with the less-educated indi-
viduals in the rural environment. In Spain an ironic
sentence emphasizes the proverbial ambiguity of the
Galician, describing him as ‘someone who if you find
him in the middle of a stairway, you can’t find out if
he’s going up or coming down’. And so the abilities of
a good interviewer, knowing how to get the subject to
respond in a precise way to a specific question, have
been a decisive factor in the execution of the work and
have contributed to reducing difficulties to a mini-
mum.
All in all, the interviewers were surprised at how
well the DIS functioned. During the training course,
comments like ‘these kinds of interviews don’t work
very well with old people here’ were frequent.
Nevertheless, in looking over the questionnaire and,
moreover, after conducting various interviews, they
admitted that the interview was very well thought out,
and that ‘once you get the hang of it you can have a
very friendly dialogue with the subject which is not
overly rigid’.
The interviewers found that the greatest difficulty
was the fact that they were sometimes not left alone
with the subjects. When this was the case, the family
members present tended to comment on the questions
and answers, drawing out the time of the interviewer
and causing difficulties. The interviewers placed great
value on the fact that the interview began by asking
the subjects about their medical problems, and equally,
that the MMSE was left for the end. This was the most
difficult part of the interview, as it was inevitable that
As has already been said, the most difficult part of
the interview was the MMSE. All the subjects felt as if
they were being given the ‘third degree’, and some of
them responded with feelings of shame, others with
nervousness, and many wanted to terminate the inter-
view right there and then. It was often necessary in
these situations to motivate the subject to respond.
This is not the time to make an exhaustive criticism of
the most dubious items on the MMSE, but in general
the subjects found it difficult to understand the
directions to: ‘write a sentence’, and calculation
(subtraction) was a task most found very difficult
indeed.
In the psychotic disorders section there were doubts
concerning both delusions and hallucinations. When
certain subjects were asked if they believed that others
were trying to hurt them, they made allusion to a very
typical problem in Galicia: the neighbour who
purposefully moves a stone marker that limits the
boundaries of the subject’s tiny piece of land. Equally,
the affirmation of some subjects as to having seen the
Virgin Mary or other religious beings is firmly support-
ed within Galician culture. There were cases in which
a moment of great tension occurred when this topic
came up in an interview, as the subject perceived that
his or her religious belief was being taken as a symptom
of an unbalanced mind. It was this type of psychotic
symptomatology that caused the greatest feeling of
insecurity in the memories of the interviewers.
Concerning the ‘consumption of alcohol’, doubts also
arose: the questions about drinking seemed too direct,
for the older generations of Galicians are unaccustomed

178 Mateos et al. (II)
to counting up the numbers of drinks they have had
(generally associated with mealtimes here). On the
other hand, those that had alcohol problems would
more than likely underestimate the quantity that they
consume when asked in such a direct way. In a previ-
ous survey (Mateos et al., 1995, Mateos et al., 1999) we
seemed to receive positive results by camouflaging
questions concerning alcohol consumption under the
general heading of ‘nutrition’ (in which we asked a
wide variety of questions concerning the subject’s gen-
eral intake of food products). Perhaps it would be use-
ful to present a sort of preliminary section about how
to begin an interview in such a way that the subject is
most likely to accept being questioned in this particu-
lar topic.
Asking the elderly about their work is not nearly as
pertinent as asking someone who is actively employed;
in order to achieve meaningful responses it will proba-
bly be necessary to reform questions concerning work.
They did not find it difficult to respond to questions
concerning sexuality even though the interviewers had
expected this to be a problem. The interviewers were
pleasantly surprised with just how well the questions
were received and responded to in this section of the
interview.
symptoms suffered in the past. However, different fac-
tors could contribute towards a lapse of memory:
• The mere fact that the subject is of an advanced age
implies that he has to remember longer periods of
the life cycle.
• The subject, influenced by his culture of reference,
does not identify the symptom as such – that is to say,
as the expression of a potentially serious illness –
and therefore does not consider it to be deserving of
some kind of professional attention. We should
remember that some traditional cultures do not
even possess terms to express certain emotional
states and symptoms (Katon et al., 1982; Somerwell
et al., 1992). In Galicia there is evidence to suggest
(Rodríguez-López, 1994; Benavente et al., submitted)
that the rural–urban polarity induces a pathoplastic
effect that reproduces the transcultural research
paradigm of depression: a predominance of somatic
symptoms and hypochondria in non-Western tradi-
tional cultures whereas guilt feelings and suicidal
thoughts predominate in Western ones.
• The subject has not received medical or para-
medical attention for the symptom. This depends
not only on the cultural-attitudinal aspect previ-
ously mentioned, but also on the unavailability of
health services in the past. Galician society has
gone through a late but enormous process of mod-
ernization in the last two or three decades: this
goes from the availability of health services, to
profound cultural changes, including the cultural-
ly homogenizing role that television has played.
The availability of the present public health sys-
tem, universal and free of charge, for all ills, is
totally different from the situation of a few
decades ago, when medical attention was essen-
tially of a private nature, and therefore dependent
upon the financial situation of each individual.
• The symptom did not have a major impact on
the life of the subject (because it did not alter the
social role that he or she played).
A methodological interpretation of some prevalence
rates
This section will discuss two aspects of prevalence
rates:
• the reliability of the lifetime prevalence rates
obtained through the DIS interview in this elderly
population;
• the sociocultural bias that can affect the rates of
certain diagnoses when they are made by the com-
puterized algorithm of the DIS.
Discussion of the lifetime prevalence rates
In order to explain the psychopathology of this section
of the population we have chosen the six-month
prevalence from the ranges provided by the DIS algo-
rithms (lifetime prevalence, one year, six months, one
month, two weeks). Two types of factors, theoretical
and empirical, have brought about this decision.
From a theoretical point of view, the preparation of
diagnoses that cover broad periods of the life cycle start
by assuming that the subject is capable of remembering
All of these factors impede the memory of those
elderly people who were interviewed. We chose to
leave to the end the most obvious limiting factor on
the memory of the elderly, which is cognitive impair-
ment. The problem here would not so much be severe
impairment, which is easy to identify, and therefore
to diagnose, but slight impairment, which would

Galicia Study of Mental Health of the Elderly II 179
affect a large segment of the population. To what
extent slight impairment can be separated from
‘normal aging decline’ is, to say the least, open to
discussion (Mateos and Rodríguez-López, 1989b;
Colsher, 1992).
This is even more striking when one considers that
a diagnosis of dysthymia requres symptoms of
depression to be present for at least two years. This
surprising phenomenon reproduces the findings of
the ECA project and confirms our suspicions about
the possible bias that the problems of age-related
deficits in recall would introduce in the lifetime
prevalence estimates (Robins et al., 1984). We
agree with Tweed et al. (1992) that: ‘The lifetime
prevalence rates were found to parallel the one-
month prevalence rates (Regier et al., 1988), with
the highest rates being found in the younger age
groups and the lowest rates being found among the
elderly. Unless there was an overwhelming cohort
effect, it made little sense that the elderly (who had
the greatest cumulative exposure to risk) had low-
est rates.’
Discussion based on the empirical evidence of our results
The 12 sections of the DIS-III were administered, with
the exception of the additional questions on sexual
problems, in section 10 (questions S1-S24). For the 16
principal diagnoses of the DIS-III, we have analysed
the different prevalences provided by the five DIS
algorithms: lifetime, one-year, six-month, one-month,
and two-week prevalences. It should be pointed out
that we are using here a broader-than-usual case-
operative definition due to the fact that we also
include mild cognitive impairment, whereas the ECA
project reports usually consider severe cognitive
impairment.
There is no lack of significant warnings in the litera-
ture about the problem (Andreasen et al., 1981;
Aneshensel et al., 1987; Bromet et al., 1986; Kendell,
1988). Probably the studies that aim to measure life-
time prevalence, especially in the elderly, should take
into account the difficulties mentioned and should
provide tools that make remembering easier.
For the time being we will not enter into the details
of these figures, which will be discussed in future pub-
lications; for now we will concentrate on the relation-
ship that exists between each pair.
Table 2 clearly shows two facts:
• We observe a discreet but progressive difference
between the two-week prevalence and the one-year
prevalence (45.9% and 49.9% respectively). The
percentage of subjects with a single diagnosis is
hardly modified, but there is an increase in those
subjects with two or more disorders (from 5.5% to
9.3%).
• On the other hand, the limited difference between
the one-year prevalence and the lifetime preva-
lence (49.9 % and 63.8%, respectively) is striking.
Initially we preferred to publish the six-month
prevalence in detail because it allows a higher number
of diagnoses than the two-week or one-month
prevalences (an essential aspect of the low prevalence
diagnoses) and it would not be as subject to memory
biases as the one-year prevalence or the lifetime
prevalence. We now go on to analyse some of the rates
of certain specific DIS diagnoses. We have selected
some disorders that on first sight might seem to be
contaminated by some methodological bias.
Table 2: Row prevalence rates for any of the main 16 DIS Diagnoses according to temporal criteria
1
No. of diagnoses
Period/prevalence
Any
One
40.4%
41.5%
42.4%
40.6%
40.6%
≥ 2
5.5%
5.8%
6.6%
Two weeks
One month
Six months
One year
45.9%
47.3%
49.0%
49.9%
63.8%
9.3%
23.3%
2
Lifetime
1
In contrast to the ECA project reports, here mild cognitive impairment is included.
Diagnosis of dysthymia included.
2

180 Mateos et al. (II)
The severity of cognitive impairment
debate on what is the standard for ‘normal sexuality’ in
the elderly and if it can be defined just on the basis of
the sexual performance of younger adult people
(Oppenheimer, 1997).
In the first place, it seems obvious that a bias is intro-
duced into a screening tool like the Mini Mental State
by the low educational level of the subjects inter-
viewed. We have pointed out that the rate of subjects
who were definitely free of cognitive impairment
(category ‘absent’ at the ‘orgbrain’ variable of the DIS
algorithms) was extremely correlated with the level of
education: 3.0% (3/99), 34.1% (119/349), 60.1% (125/
208) and 90.5% (19/21), for those who were illiterate,
reading and writing, completed compulsory elementary
education, and secondary school or higher, respectively.
If we prefer to analyse the phenomenon by observing
the other tail of the ‘orgbrain’ variable, the proportion of
people with ‘severe’ cognitive impairment, according to
the level of education is: illiterate 17.2% (17/99),
reading and writing 8.6% (30/349), and com-
pulsory elementary education or higher 3.9% (9/229),
respectively.
Even excluding illiterates from the analysis, the
association between performance on the MMSE and
educational levels is maintained, with breaks in each
level of education. It is not possible at this time to tell
whether this association is simply a cultural bias intro-
duced by the instrument or if we see the expected
effect of cultural development preventing the deterio-
ration of mental functions (Leibovici et al., 1996;
Christensen et al., 1997). But the magnitude of this
association in our judgement would lead one to believe
that the MMSE should be complemented with other
items freer from cultural biases (Escobar et al., 1986;
Mateos et al., 1989; Kraemer et al., 1988).
Alcohol dependency
The six-month prevalence rates for alcohol dependen-
cy and alcohol abuse were 0.7% each. Not having
detected any women with alcohol problems, the rate
for men must be doubled.
In our previous community survey (Mateos and
Rodríguez-López, 1989a; Mateos et al., 1995), among
those surveyed with an age of over 60 years, a rate of
7% to 9% on over-drinking was revealed, based on a
limit of 75 cc or 50 cc of pure alcohol per day, respec-
tively (in other words, about 30 or 40 drinks a week).
Lacking a more detailed study that would take into
account the differences in methodology of the two
studies, it is difficult to draw conclusions, but perhaps
the rates given by the DIS underestimate the southern
European style of drinking, so different from the
Anglo-Saxon one, particularly in the elderly.
Schizophrenia
The six-month prevalence for schizophrenia or schizo-
phreniform disorder was zero, whereas the lifetime
prevalence was 0.07%. These figures are even lower than
those found in the ECA project for people of 65 years
and over (0.1% and 0.3%, respectively) (Keith et al.,
1991).
Our study did not include an institutionalized popu-
lation and it is known that these subjects are usually
concentrated in old psychiatric hospitals as well as in
nursing homes, but it does not seem plausible that omis-
sion of the institutionalized could explain such a low
prevalence. In our study the phenomenon can probably
be better explained by the enormous difficulty involved
in attempting to collect psychotic symptoms through lay
personnel. Problems were foreseen right from the begin-
ning of the DIS training course and the later comments
by the interviewers confirmed this. It seems clear that
this is the Achilles’ heel of instruments such as the DIS.
The aspects that are discussed above can make com-
parison of epidemiological rates particularly difficult. We
have analysed the possible bias that the inclusion of
diagnoses of dubious validity (sexual dysfunction), which
are difficult to operationalize (cognitive impairment), or
those habitually disregarded in epidemiological studies
(tobacco dependence) can have on the calculation of
the overall prevalence. Table 3 shows nine operative
Sexual dysfunction
From the four general questions on sexuality (in our
study we did not ask the sexual questions for somatization
disorder) a surprisingly high rate of sexual disorders was
shown. However, these rates could be deceptive in
terms of the significance of the disorders that were
detected. The six-month prevalence (13%) of the
sexual dysfunctions, which was much greater among
women, was probably overestimated. Attitudes and
beliefs, which are very different in old people in
Galicia compared with the young, have affected these
results. We must be very cautious when considering
these elevated rates of ‘sexual disorders’, and suspect an
important cultural bias in the origins of their operative
definitions. Alternatively, it could be argued that
results may be accurate as a decline in sexual activity is
a worldwide phenomenon, but this would move the

Galicia Study of Mental Health of the Elderly II 181
Table 3: Six-month prevalence for nine possible operative definitions of the condition ‘DIS case’, depending upon the
1
severity of cognitive impairment and the total number of DIS diagnoses covered
2
1) All levels CI
included
2) Only severe CI
included
3) CI
included
Diagnoses taken into account
(A) = All the diagnoses provided by the DIS
(B) = Tobacco dependence excluded
(C) = Tobacco dependence and sexual dysfunction
excluded.
56.6
54.2
47.6
37.3
34.7
24.0
29.9
27.4
16.6
3
1
Weighted rates.
CI = Cognitive impairment.
2
3
Reasons for choosing this operative definition of DIS case: it is the most reliable, the easiest to compare with the ECA
reports and the one with the highest clinical meaning.
cultural medium. The one exception here would be the
definitions of the condition of presenting ‘at least one
important area of alcohol problems and the difficulties
DIS diagnosis’. It also clarifies the important contribu-
these pose in Galicia. In order to bring a solution to
tion of conditions such as sexual dysfunction, tobacco
the said limitations, certain minor changes need to be
dependence and cognitive impairment. In order to
made when formulating some parts of the DIS-III, the
facilitate meaningful comparisons between our results
majority of which, on the other hand, have already
and those in the literature, we have chosen the option
been taken into account in later editions of the instru-
ment, DIS IIIR and DIS IV.
that includes ‘severe cognitive impairment’ and
excludes the diagnoses of ‘sexual dysfunction’
and ‘tobacco dependence’. This definition has two
Being pragmatic, we sincerely believe that limita-
tions like those expressed above, when they have been
advantages:
analysed and are handled prudently, are acceptable in
community surveys whose intention is to find out scien-
tifically the wide range of psychiatric pathology. For this
reason we subscribe to the two conclusions in Tweed et
al’s (1992) review: ‘We feel that much can be done with
existing instruments, such as the DIS . . . but more vali-
dation studies should be performed at the symptom
assessment and diagnostic levels, with a special regard
for how these processes can interact with characteristics
of the elderly to give biased information.’
We consider that the comments made in this work
are the best way we have of giving a certain
anthropological sense to the objective data in our
epidemiological survey. Furthermore, as far as our lim-
ited experience will allow, they are a way of helping to
improve the DIS.
• it discards two diagnoses of limited validity and/or
reliability: mild cognitive impairment and sexual
dysfunction;
• it allows comparison with the ECA study, where
publications usually omit both of these diagnoses as
well as the analysis of tobacco dependence.
Conclusions
After carefully reviewing all the limitations, difficulties
or doubts that the use of the DIS has raised in our com-
munity survey, we have had no option but to ask
ourselves whether this is perhaps the inevitable price
one has to pay in order to obtain an instrument that is
easy to administer to large samples of the population.
The perfect instrument, which is easy to administer and
economical, is still not available, and although we hope
to be proved wrong, we fear that it never will be. We
have to admit that the criteria used to define a case, as
well as how it is put into practice in a diagnostic instru-
ment, will vary depending upon the objectives and
resources of the researchers (Wing et al., 1981).
All things considered, these observations present
minor difficulties that do not put in question the
validity of the nucleus of the DIS interviews within our
Although certain reservations have been expressed
regarding the use of the DIS in our community survey,
we would like to point out that, even though there is
room for improvement, we continue to consider it a
highly valuable tool for use in research of this kind.
Acknowledgements
This research has been financed by the Galician
Government (Xunta de Galicia, Consellería de Sanidade e
Servicios Sociais, Servicio de Saúde Mental).

182 Mateos et al. (II)
The following people supplied different Spanish versions
of the DIS: Professor J Guimon (‘DIS Spanish version of the
DIS’, with permission of the work team from Los Angeles);
Professor G Canino, Dr M Bravo and Dr G Bernal (‘DIS-DS
1987 Questionnaire’); Professor RE Roberts (‘Spanish DIS’);
Professor Marvin Karno (‘English-Spanish multilingual ver-
sion of the DIS’).
I Azevedo, A Cunhal, C Gimaraes and T Pereira collabo-
rated with A Droux in teaching our interviewers how to use
the DIS interview. The authors would like to thank the fol-
lowing people for the enthusiasm and professionalism they
demonstrated in administering the DIS interview: MT
Sánchez, C Calvo, M Borrás, M Durán, JL Gil, E Tacon, CM
Gayoso, G Casado, V Pumar, MJ González, MJ Sánchez, MJ
González, MJ Sánchez, I Murados and E Calvo. They would
also like to thank to Dr C González-Borrás and the nursing
staff at the Psychiatric Hospitalization Unit of the Hospital
General de Galicia for the facilities provided for interviewing
patients during training with the DIS.
Christensen H, Korten AE, Jorm AF. Education and decline
in cognitive performance. Int J Geriatr Psychiatry 1997;
12(3): 323–30.
Droux A, Meireles H. The use of the DIS III in a study of
prevalence of mental disorders in Gondomar Portugal. In
Book of Abstracts of the V Congress of the International
Federation of Psychiatric Epidemiology. Montreal, 1990,
p. 103.
Eaton WW, Kessler LG. Epidemiologic Field Methods in
Psychiatry. The NIMH Epidemiologic Catchment Area
Program. Orlando: Academic Press, 1985.
Escobar JI, Burnam A, Karno M, Forsythe A, Landsverk J,
Golding JM. Use of the Mini-Mental State Examination
(MMSE) in a community population of mixed ethnicity.
Cultural and linguistic artifacts. J Nerv Ment Dis 1986;
174: 607–14.
Goldberg DP. The detection of psychiatry illness by ques-
tionnaire. Maudsley Monographs 21. London: Oxford
University Press, 1972.
A travel grant from the University of Santiago de
Compostela enabled Dr Mateos to take part in a course on
the latest DIS/CIDI interviews at Washington University (St
Louis). The excellence of the course, directed by Professor
Lee Robins, and the contact with researchers from different
cultures provided us with new insights for our methodologi-
cal reflection.
A preliminary version of this paper was presented at the
VI Congress of the International Federation of Psychiatric
Epidemiology (Lisbon, 1993).
Karno M, Burnam A, Escobar JI, Hough RL, Eaton WW. The
Spanish Language Version of the Diagnostic Interview
Schedule. In WW Eaton, LG Kessler (eds) Epidemiologic
Field Methods in Psychiatry. The NIMH Epidemiologic
Catchment Area Program. Orlando: Academic Press,
1985, pp. 171–90.
Katon W, Kleinman A, Rosen G. Depression and somatiza-
tion: a review. Part I. Am J Med 1982; 72: 127–35.
Keith SJ, Regier D, Rae DS. Schizophrenic disorders. In L
Robins, D Regier (eds) Psychiatric Disorders in America.
the Epidemiologic Catchment Area Study. New York:
The Free Press, 1991, pp. 33–52.
Kendell R. What is a case? Food for thought for epidemiolo-
gists. Archives of General Psychiatry 1988; 45: 374–6.
Kohout FJ. The pragmatics of survey field work among the
elderly. In RB Wallace, RF Woolson (eds) The
Epidemiologic Study of the Elderly. New York: Oxford
University Press, 1992, pp. 91–119.
Kraemer HC, Moritz DJ, Yesavage J. Adjusting Mini-Mental
State Examination scores for age and educational level to
screen for dementia: correcting bias or reducing validity?
Int Psychogeriatr 1998; 10: 43–51.
Leibovici D, Ritchie K, Ledésert B, Touchon J. Does educa-
tion level determine the course of cognitive decline? Age
Ageing 1996; 25(5): 392–7.
Marcos LR, Trujillo MR. Culture, Language and
Communicative Behavior: The psychiatric examination
of the Spanish-Americans. In RP Duran (ed.) Latino
Language and Communicative Behavior. Newark: Alex
Publishing Co, 1981.
Mateos R, Vázquez A, Cimadevila C, Rodríguez-López A.
Socio-cultural bias of the Mini-Mental State (Spanish
Version). A methodological contribution from the
Galician Community Study of Psychiatric Epidemiology.
In Book of Abstracts of the XVI International Congress of
Gerontology, Acapulco, 1989, p. 330.
The authors are grateful to Lee Robins, Brian Cooper and
Javier I Escobar for their comments on a previous version of
this paper.
References
Andreasen N, Growe W, Shaphiro R, Keller M, Hirsfeld R,
McDonald-Scott P. Reliability of lifetime diagnosis: A
multicenter collaborative perspective. Arch Gen
Psychiatry 1981; 38: 400–5.
Aneshensel C, Estrada A, Hansell M, Clark V. Social
Psychological aspects of reporting behavior: Lifetime
depressive episode reports. J Health Soc Behav 1987; 28:
232–46.
Benavente JL, Mateos R, Cerdá T, Pérez-Vilariño J,
Carbonell C. Depressive phenomenology and socio-
cultural factors. A study on urban and rural populations of
Galicia (Spain). Submitted.
Bromet E, Dunn L, Connell M, Dew M, Schulberg H. Long-
term reliability of diagnosing lifetime major depression in
a community sample. Arch Gen Psychiatry 1986; 43:
435–40.
Colsher PL. Epidemiological Studies of Cognitive Function
in the Elderly: Rationale, Methods and Findings. In RB
Wallace, RF Woolson (eds) The Epidemiologic Study of
the Elderly. New York: Oxford University Press, 1992, pp.
130–56.

Galicia Study of Mental Health of the Elderly II 183
Mateos R, Rodríguez-López A. Estudio de Epidemiología
Psiquiátrica en la Comunidad Gallega. Santiago de
Compostela: Xunta de Galicia, 1989(a).
Mateos R, Rodríguez-López A. Epidemiología de las
Demencias: Perspectivas metodológicas y datos de preva-
lencia. Jano 1989(b); 3(4): 21–8.
Rodríguez-López A. Social Psychiatry and Psychiatric
Epidemiology, a classical research line in the Psychiatry
Chair of the University of Santiago de Compostela. In
Book of Abstracts of the VI Congress of the International
Federation of Psychiatric Epidemiology. Lisbon: 1993,
p. 202.
Mateos R, Páramo M, García MC, Carrera I, Rodríguez-López
A. Estudio epidemiológico comunitario del consumo de
alcohol y alcoholismo. In F Pulido and A Sierra eds. La
investigación Epidemiológica de las Drogodependencias.
Canarias: ICEPSS SL, 1995, pp. 101–11.
Mateos R, González F, Páramo M, García MC, Carollo MC,
Rodríguez-López A. The Galicia Study of Mental Health
of the Elderly I: General Description of Methodology.
International Journal of Methods in Psychiatric Research
2000; 9(4): 165–73.
Munson ML, Orvaschel H, Skinner EA, Goldring E,
Pennybacker M, Timbers DM. Interviewers: characteris-
tics, training and field work. In WW Eaton, LG Kessler
(eds) Epidemiologic Field Methods in Psychiatry. The
NIMH Epidemiologic Catchment Area Program.
Orlando: Academic Press, Inc, 1985, pp. 69–83.
Rodríguez-López A. De la Epidemiología a la Antropología
Psiquiátrica. In Proceedings of II Reunión Interregional
de Psiquiatría. Llanes, 1994, pp. 205–14.
Rodríguez-López A. Galicia: a Medicina Popular ante a
Depresión. In E González (ed.) Cultura e Saúde Mental.
Aproximacións Antropolóxicas á clínica psiquiátrica en
Galicia. Santiago de Compostela: Xunta de Galicia, 1994,
pp. 29–38.
Somervell PD, Beals J, Kinzie JD, Boehnlein J, Leung P,
Manson SM. Use of the CES-D in an American Indian
village. Cult Med Psychiatry 1992; 16: 503–17.
Straight HS. Knowledge, purpose and intuition: three
dimensions in the evaluation of translation. In MG Rose
(ed.) Translation Spectrum: Essays on Theory and
Practice. Albany: State University of New York Press,
1981
Oppenheimer C. Sexuality in old age. In R Jacoby, C
Oppenheimer (eds) Psychiatry in the Elderly. Oxford:
Oxford University Press, 1997, pp. 689–708.
Regier D, Boyd J, Burke J, Rae D, Myers J, Kramer M, Robins
L, George L, Karno M, Locke B. One-month prevalence
of mental disorders in the United States: based on five
epidemiologic catchment area sites. Arch Gen Psychiatry
1988; 45: 977–86.
Tweed DL, Blazer DG, Ciarlo JA. Psychiatric Epidemiology
in Elderly Populations. In RB Wallace, RF Woolson (eds)
The Epidemiologic Study of the Elderly. New York:
Oxford University Press, 1992, pp. 213–33.
Wing JK, Bebbington P, Robins LN. What is a Case? The
problem of definition in psychiatric community surveys.
London: Grant Mc Intyre, 1981.
Robins LN, Helzer JE, Croughan J, Ratcliff KS. National
Institute of Mental Health Diagnostic Interview
Schedule: its history, characteristics and validity. Arch
Gen Psychiatry 1981; 38: 381–9.
Robins L, Helzer J, Weissman M, Orvaschel H, Gruenberg E,
Burke J, Regier D. Lifetime prevalence of specific disor-
ders in three sites. Arch Gen Psychiatry 1984; 41: 949–58.
Correspondence:
Professor
Raimundo
Mateos,
Departamento de Psiquiatría, Facultad de Medicina,
Santiago de Compostela, E-15705 Spain. Fax
+34.981.95 19 01. E-mail mrmateos@usc.es