Psychological Medicine, University of Wales College of Medicine, Cardiff
Clinica Psiquiatrica, Facultad de Medicina, Universidad de Chile, Santiago, Chile
Department of Psychological Medicine, University of Wales College of Medicine, Cardiff
Correspondence: Dr R. Araya, Senior Lecturer, Psychological Medicine, University of Wales College of Medicine, Monmouth House, Heath Hospital, Cardiff CF14 4XN. E-mail: arayari{at}cf.ac.uk
Declaration of interest This study was funded by the European Community.
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Aims To determine the prevalence of common mental disorders and socio-demographic correlates among adults from Santiago, Chile.
Method Cross-sectional survey of private households with a probabilistic sampling design was used. Common mental disorders were measured using the Clinical Interview ScheduleRevised (CISR).
Results Three thousand eight hundred and seventy adults were interviewed. Twenty-five per cent were CISR cases and 13% met criteria for an ICD-10 diagnosis. Low education, female gender, unemployment, separation, low social status and lone parenthood were associated with a higher prevalence.
Conclusions Prevalence rates were higher than those found in urban areas of Great Britain, both for ICD-10 diagnoses and non-specific neurotic disorders. Similar socio-demographic factors were associated with an increased prevalence of common mental disorders in Chile as in the UK. There is a need to unify methodologies to be able to compare results internationally.
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Research from developed and developing countries has shown that mental disorders are more common among the more socially disadvantaged, people who have previously been married and women (Canino et al, 1987; Cheng, 1988; Chen et al, 1993; Kessler et al, 1994; Almeida-Filho et al, 1997; Jenkins et al, 1997).
This study used similar methodology to the National Psychiatric Morbidity survey of Great Britain (Jenkins et al, 1997) and aimed:
![]() |
METHOD |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Interviewers were instructed to make at least four visits before declaring an address as lost. The only criterion for exclusion, apart from age, was that the person selected was unable to answer questions because of physical impediments or other personal reasons such as being away for a long period. Household size was defined as the number of people aged 16-64 years residing in that property who were eligible for interview. The selected households were not visited before initiation of the fieldwork, so it was not possible to ascertain a priori if any of the residents did not meet the criteria for inclusion in the study. (Further details of the sampling design or procedures are available from R.A. upon request.)
Several pilot studies were carried out to study the validity, reliability and feasibility of the questionnaires and the procedures to be used (Lewis et al, 1992; Araya et al, 1994). One hundred and ninety-seven interviewers took part in the fieldwork. Each interviewer had to complete successfully a 3-day training course that included a reliability exercise marking three mock videotape interviews. Lay interviewers who did not achieve satisfactory agreement with most of the ratings on the video interviews were excluded from the fieldwork. Interviewers were closely supervised, with systematic revisiting of randomly chosen addresses to check the reliability of information. Data collection took place between October 1996 and April 1998.
Assessment
A detailed structured questionnaire covering socio-demographic factors,
general health, use of health services, social support, psychiatric morbidity,
and tobacco, alcohol and drug use was administered to each interviewee.
Common mental disorders
Common mental disorders (CMDs) were assessed using the Revised Clinical
Interview Schedule (CISR) (Lewis
et al, 1992). This is a structured interview that has
been fully standardised so that it can be administered by social survey
interviewers. The English and Spanish versions of the CISR have been
used extensively in primary care, occupational and community studies with good
reliability and validity (Lewis et
al, 1992). Although the CISR enquires about 14 common
neurotic symptoms present during the preceding week, it includes questions to
identify the onset and duration of each episode, so allowing categorisation
according to ICD10 (World Health
Organization, 1992) criteria for the most common disorders. Each
one of the 14 symptoms is rated with an individual score. The total sum of
these 14 scores can be used as a good indicator of the severity of a CMD.
People scoring 12 or above on the CISR were regarded as suffering from
a CMD (Lewis et al,
1992). The following ICD10 diagnoses were included:
depressive episodes (F32.00, 32.01, 32.10, 32.11, 32.2); phobias (F40.00,
40.01, 40.01, F40.2); panic disorder (F41.1); generalised anxiety disorder
(F41.1); and obsessivecompulsive disorder (F42). The diagnostic group
of mixed anxiety and depressive disorder (F42.1) was not included because the
ICD10 does not provide an operational definition for research purposes.
All those subjects who were above the threshold on the CISR, but failed
to meet explicit ICD10 criteria for a psychiatric diagnosis were
grouped under a category denominated non-specific neurotic disorder, our
equivalent of the mixed anxiety and depressive disorder. The prevalence rates
of specific diagnoses cannot be added up because no attempt was made to
establish a hierarchy of specific diagnoses. ICD10 diagnoses were
reached using computer-assisted algorithms developed by Meltzer et al
(1995).
The CISR was chosen for several reasons. First, because it provides a measurement of the severity of symptoms including those with a sub-threshold intensity. Second, because it allows establishing ICD10 diagnoses of the most common disorders. Third, because the period of time enquired about is reasonably short, diminishing the possibility of memory distortions about conditions experienced many years ago. Fourth, because the time taken to administer this interview is brief in comparison with other similar instruments. The administration of the CISR took an average of 30 minutes in this study. This is important because the length of an interview influences the reliability of the answers, the refusal rates, and the costs of the fieldwork. Lastly, because the validity and reliability of the CISR are comparable to the other commonly used structured interview, Composite International Diagnostic Interview (CIDI) (Lewis et al, 1992; Andrews & Peters, 1998; Brugha et al, 2000).
Most of the questions used to measure other variables included in this study were derived from the National Psychiatric Morbidity Survey of Great Britain (NPMSGB; Meltzer et al, 1995). The Spanish questionnaires were translated and back-translated using a standard procedure recommended by the World Health Organization. These questionnaires are available upon request from R.A.
Family type
This variable is divided in the same five categories as in the NPMSGB.
Couples were divided into two groups: those with or without children. The
other three groups included were lone parents, one-person
families, and respondents living with parents. Lone parents and couples
with children could include children older than 16 years provided they were
not married or they did not have children of their own in the same household.
The one-person family does not necessarily mean a person living
alone because this person may be sharing premises with another family
unit.
Social class
This variable was based on the household's main earner's occupation. We
used the Chilean National Institute of Statistics scale to classify
occupations according to: "prestige, power, economic income, and
stability of occupation". Four categories were included: (a) low-status
and unstable occupation involving manual non-specialised working
freelance; (b) low-status but stable occupation involving manual
nonspecialised employees; (c) middle-status occupation involving
non-manual workers, with no professional qualifications; (d) high-status
occupation involving nonmanual professional or business people with
prestigious posts. For those households where no one was currently employed,
coding was based on the last occupation of the main earner.
Employment status
This was divided in four groups. The employed, including
those people doing unpaid work for a family business. The
unemployed category, including those who were looking for a job
and those individuals who were temporarily unemployed because of sickness or
injury. The economically inactive group, including housewives,
students, the retired and those permanently unable to work because of illness
or disability.
Analysis
This was approached in two different ways. Prevalence estimates for the
adult population of Greater Santiago, Chile, were calculated adjusting for
differential sampling and household size by using weights in the analysis. The
interviewed sample was compared by age and gender with population projections
based on the 1992 National Census results for Greater Santiago, and
differences were used to further modify the weights. In view of the
multi-stage random sampling design, 95% confidence intervals (95% CIs) were
calculated from standard errors estimated using the survey commands of the
computer program STATA (STATA,
1999), which takes into account the effect of the sampling
strategy (stratification and clustering) and sampling weights.
Associations between CMD and socio-demographic factors of the respondents were examined using odd ratios. These ratios and their 95% CIs were calculated using logistic regression, both before and after adjustment for various socio-demographic factors. The analysis was performed using STATA survey commands adjusting for sampling design effects as well as differential sampling weights.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
Prevalence of common neurotic symptoms
The most common symptom was worries, with a prevalence of
43%. Other non-specific symptoms were also common, such as fatigue,
irritability and sleep problems. Specific psychiatric symptoms such as
depression, anxiety or phobias were less common. All symptoms were more common
in women than in men. The prevalence by gender of the 14 neurotic symptoms
included in the CIS-R is presented in Fig.
1.
|
Prevalence of common mental disorders
The mean CIS-R total score was 7.78 (95% CI 7.45-8.30). The distribution of
total CIS-R scores for each gender for the surveyed household sample can be
seen in Fig. 1. Women were
overrepresented compared with men in the high-score bands of the CIS-R.
The overall prevalence of CMD in Greater Santiago was 26.7% (95% CI 24.5-29.0), including both CIS-R cases and ICD-10 diagnoses. Approximately half of these individuals met ICD-10 criteria for at least one of the five diagnoses studied and the remaining half were cases of non-specific neurosis (being a CIS-R case but not meeting criteria for any of the ICD-10 diagnoses included) (see Table 1). The 1-week prevalence of individuals scoring more than 12 points in the CIS-R was 25% (95% CI 22.8-27.3). The 1-week prevalence of individual ICD-10 psychiatric diagnoses is shown in Table 1. Depressive and generalised anxiety disorders were the most common ICD-10 diagnoses found in this study. Taken as a whole 9.5% (95% CI 8.1-10.9) of the people met criteria for either a diagnosis of ICD-10 depression or anxiety in the preceding week. Women had higher prevalence rates than men in all diagnostic categories with the exception of the least prevalent diagnosis of obsessive-compulsive disorder. Women had nearly twice the prevalence of non-specific neurotic disorders and overall ICD-10 diagnoses.
|
Common mental disorders and socio-demographic correlates
Many socio-demographic factors were associated with a higher prevalence of
CMD (see Tables 2 and
3). CMDs were significantly
more prevalent among women than men and the strength of this association
changed little after adjusting for other variables. Age was associated with
the prevalence of CMD: the highest prevalence was found among those
individuals aged 25-39 years and the lowest in the oldest group. These
differences remained significant in the multivariate analysis (adjusted Wald
test F=7.1, P=0.0001), after adjusting for other confounding
variables.
|
|
Respondents who had been separated or widowed were more likely to be a case of CMD than single individuals. However, only those individuals who were separated showed a statistically significant (P < 0.05) association with a higher prevalence of CMD, after adjustment for other variables. Lone parents with children had increased rates of psychiatric morbidity, an association that persisted after adjusting for other potential confounding variables.
Both education level (2=14.07, d.f.=3,
P<0.0001) and social class according to occupation
(
2=27.68, d.f.=3, P<0.0001) had linear
dose-response associations with the prevalence of CMD. Those in the most
disadvantaged groups with less education and in lower social class had a
higher prevalence of CMD. Although these gradients were reduced after
adjustment, they remained statistically significant. Although all employment
status groups had higher prevalence rates than the full-time employed, these
associations were not statistically significant after adjustment, with the
exception of being unemployed.
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
A quarter of the adult residents of Greater Santiago could be considered to be suffering from a common neurotic disorder, half of whom met criteria for an ICD-10 psychiatric diagnosis. Of the neurotic disorders considered, the two most common ICD-10 diagnoses were depressive and generalised anxiety disorders. Women and the most socially disadvantaged groups had a higher prevalence of CMD.
Strengths and weaknesses of this study
This study used a structured psychiatric interview (CIS-R) administered to
a representative sample of the adult population residing in private households
in Santiago, capital of Chile. Experienced lay interviewers underwent an
intensive training, including several tests of reliability, and were closely
supervised during the fieldwork. Studies of this design and scale are
difficult to conduct outside Western market economies. Despite these
strengths, this was a cross-sectional study involving only one city, thus it
would be unwise to generalise our results to the rest of the country or to
reach conclusions about aetiological relationships. However, it is still
possible to identify groups with poor health and identify strong associations
that might inform future research on causal relationships.
The validity of information collected using clinical interviews administered by lay people has been questioned (Brugha et al, 1999), but we agree with other researchers (Wittchen et al, 1999) in thinking that on balance this method represents the most cost-efficient way of undertaking community surveys of this size. From a practical perspective, it is difficult to conceive of a study over a large geographical area involving several thousand interviews that could be administered by psychiatrists in settings with a shortage of these specialists. We are not aware of any survey of the house-hold population that has achieved more than 1000 structured interviews administered by psychiatrists.
In the absence of gold standards one can only assess the reliability of measurement. Structured interviews administered by lay people are as reliable, if not more reliable, than interviews administered by psychiatrists (Lewis et al, 1992; Andrews & Peters, 1998; Wittchen et al, 1999). By using this methodology we can be confident that everyone in the study was asked the same questions and this aids comparison within and between studies.
Prevalence of common mental symptoms
Non-specific symptoms such as worries, irritability, fatigue and sleep
problems were by far the most frequent both in the Chilean and the British
urban samples. Specific psychological symptoms such as depressed or anxious
mood, essentials for reaching ICD-10 diagnoses, were much less common in both
countries. In the Chilean survey the most common reported symptom was
worries (Chile, 38%; GB urban, 21%) whereas
fatigue was the most common symptom in Great Britain (GB urban,
28%; Chile, 30%). The high prevalence of these non-specific symptoms might
help to explain the large proportion of people with high scores in the CIS-R
who did not meet criteria for an ICD-10 diagnosis.
All symptoms were more frequent in the Chilean than in the British sample. In both countries all symptoms were more prevalent in women than men. In addition, this study contributed to the growing body of research showing that people from many non-Western countries do report psychological symptoms when asked specifically about them. Notwithstanding this, somatic complaints and worries about physical health were more common in Chile than in urban areas of Great Britain: 19% v. 8% and 11% v. 5%, respectively.
Prevalence of CMDs
Prevalence rates of common neurotic disorders among adult residents from
the Greater Santiago area (27%) were higher than those found in urban areas of
Great Britain (18%). However, approximately the same proportion of ICD-10
diagnosable conditions and non-specific neurotic disorders were present in
both samples. It can be argued that non-specific neurotic disorders do not
represent clinically meaningful morbidity. However, previous
studies have demonstrated the public health importance of sub-threshold and
mild psychiatric disorders (Broadhead
et al, 1990). As far as diagnoses are concerned, higher
prevalence rates were found for the diagnoses of depressive disorders and
phobias in the Chilean compared with the UK sample: 5.5% v. 2.6% and
4.3% v. 2.1%, respectively.
After a literature search covering the period from 1985 up to the present, only two Latin American household surveys of psychiatric morbidity that had used a structured interview and a probabilistic sampling design were found in English peer-reviewed journals (Canino et al, 1987; Almeida-Filho et al, 1997). Other household surveys have been carried out but do not meet the criteria previously outlined.
Comparing our results with these two surveys is difficult because of methodological differences. The Puerto Rican study (Canino et al, 1987) utilised the Diagnostic Interview Schedule (DIS) with an island-wide probability sample and calculated 6-month prevalence estimates of DSM-III (American Psychiatric Association, 1980) diagnoses. The overall prevalence found in Puerto Rico is not comparable with our study because it includes psychotic disorders and substance misuse. However, rates of DSM-III major depression (3%) and anxiety disorders (7.5%) were similar to our study. However, it needs to be borne in mind that our study involved a much shorter prevalence period.
The Brazilian study (Almeida-Filho et al, 1997) was a two-stage design of population samples from three major cities. A subsample of individuals had diagnostic interviews using the Brazilian version of the DSM-III Symptom Check-List administered by psychiatrists. Comparisons with this study are even more cumbersome because their 1-year prevalence estimates were adjusted according to the potential need-for-treatment judged by the interviewing psychiatrist. The most common diagnoses were anxiety and phobic disorders, with 1-year prevalence estimates varying between 7% to 12%.
CMD and socio-demographic correlates
Most of the published research from the developing, as well as developed,
countries has found that women, people who had been previously married and
people belonging to the most socially disadvantaged groups (education, income,
and social class) have a much higher prevalence of CMD. This study replicated
these findings, showing a similar profile of associations between
socio-economic variables and CMD as in Great Britain
(Jenkins et al, 1997)
and Puerto Rico (Canino et al,
1987). Women, divorced or separated individuals, lone parents,
those people with less education or belonging to lower social class groups and
the unemployed all showed higher prevalence rates of common neurotic disorders
in both countries.
The mild but significant association between young adulthood (age 25-39 years) and the presence of a CMD found in the Chilean study persisted after adjusting for other variables. This association was not found in the British study, which showed a trend for the prevalence of mental disorders to decrease with greater age - this is similar to the trends found in the Chilean study for other age bands. In keeping with our results, some studies have found an increased prevalence of CMD in younger cohorts, with a decrease with age in other groups (Kessler et al, 1994). However, other studies have found a rise in the prevalence of CMD with increasing age (Canino et al, 1987). The higher prevalence of CMD among young adults in Chile and other countries deserves further investigation because this age group plays a vital role in any productive economy and these individuals could be experiencing more distress as a result of all the pressures of living in emerging economies.
On the basis of our study we estimate that about 390 134 people met criteria for an ICD-10 psychiatric diagnosis in Santiago, Chile. If only half of this group needed treatment at an average annual cost of US$150, this would total US$74 million per annum. Bearing in mind that the entire Chilean health budget was about US$1500 million in 1997 and that approximately 10% of this sum went into mental health, half of the mental health budget would have to be spent on the treatment of these common neurotic disorders. Based on this example it is clear that establishing priorities becomes an essential first step in dealing with this important disease burden. Unfortunately, local data on the cost and benefits of treating common psychiatric disorders are scarce. Information derived from community surveys should help decision makers know how best to allocate scarce resources.
![]() |
Clinical Implications and Limitations |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
LIMITATIONS
![]() |
ACKNOWLEDGMENTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
American Psychiatric Association (1980) Diagnostic and Statistical Manual of Mental Disorders (3rd edn) (DMS-III). Washington, DC: APA.
Andrews, G. & Peters, L. (1998) The psychometric properties of the Composite International Diagnostic Interview. Social Psychiatry and Psychiatric Epidemiology, 33, 80-88.[CrossRef][Medline]
Araya, R. (2000) Common Mental Disorders and Detection by Primary Care Physicians. Santiago, Chile. PhD Thesis. University of London: London.
Araya, R., Wynn, R., Leonard, R., et al (1994) Psychiatric morbidity in primary health care in Santiago, Chile. Preliminary findings. British Journal of Psychiatry, 165, 530 -533.[Abstract]
Broadhead, W. E., Blazer, D., George, L., et al (1990) Depression, disability days and days lost from work. Journal of the American Medical Association, 264, 425 -428.
Brugha, T. S., Bebbington, P. E. & Jenkins, R. (1999) A difference that matters: comparisons of structured and semi-structured psychiatric diagnostic interviews in the general population. Psychological Medicine, 29, 1013 -1020.[CrossRef][Medline]
Brugha, T. S., Bebbington, P. E., Jenkins, R., et al (2000) Cross-validation of a general population survey diagnostic interview: a comparison of CIS-R with SCAN ICD-10 diagnostic categories. Psychological Medicine, 29, 1029 -1042.[CrossRef]
Canino, G. J., Bird, H. R. & Shrout, P. E. (1987) The prevalence of specific psychiatric disorders in Puerto Rico. Archives of General Psychiatry, 44, 727 -735.[Abstract]
Chen, C., Wong, J., Lee, N., et al (1993) The Shatin Community Mental Health Survey in Hong Kong. Archives of General Psychiatry, 50, 125 -133.[Abstract]
Cheng, T. A. (1988) A community study of minor psychiatric morbidity in Taiwan. Psychological Medicine, 18, 953 -968.[Medline]
Jenkins, R., Lewis, G., Bebbington, P., et al (1997) The National Psychiatric Morbidity Surveys of Great Britain: initial findings from the household survey. Psychological Medicine, 27, 775 -790.[CrossRef][Medline]
Kessler, R. C., McGonagle, K. A., Zhao, S., et al (1994) Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States. Archives of General Psychiatry, 51, 8 -19.[Abstract]
Kish, L. (1965) Survey Sampling. Chichester: John Wiley & Sons.
Lewis, G., Pelosi, A. J., Araya, R., et al (1992) Measuring psychiatric disorder in the community: a standardised assessment for use by lay interviewers. Psychological Medicine, 22, 465 -486.[Medline]
Meltzer, H., Gill, B., Petticrew, M., et al (1995) OPCS Surveys of Psychiatric Morbidity. Report 1. The Prevalence of Psychiatric Morbidity among Adults Aged 16-64 Living in Private Households in Great Britain. London: Stationery Office.
STATA (1999) STATA. Version 6. Santa Monica, CA: STATA.
Ustun, B. (1999) The global burden of mental disorders. American Journal of Public Health, 89, 1315 -1318.[Abstract]
Ustun, B. & Sartorius, N. (1995) Mental Illness in General Health Care: An International Study. London: John Wiley & Sons.
Wittchen, H. U., Ustun, B. & Kessler, R. (1999) Diagnosing mental disorders in the community. A difference that matters? Psychological Medicine, 29, 1021 -1027.[CrossRef][Medline]
World Bank (1993) Development Report: Investing in Health. New York: Oxford University Press.
World Health Organization (1992) International Classification of Diseases and Related Disorders (ICD-10). Geneva: World Health Organization.
Received for publication March 8, 2000. Revision received July 21, 2000. Accepted for publication July 21, 2000.