SciELO - Scientific Electronic Library Online

vol.30 issue2Is depression a risk factor for mortality in chronic hemodialysis patients?Gender differences in aggressiveness in children and adolescents at risk for schizophrenia author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand




Related links


Brazilian Journal of Psychiatry

Print version ISSN 1516-4446On-line version ISSN 1809-452X

Rev. Bras. Psiquiatr. vol.30 no.2 São Paulo June 2008  Epub Apr 28, 2008 



Factors associated with depressive symptoms measured by the 12-item General Health Questionnaire in Community-Dwelling Older Adults (The Bambuí Health Aging Study)


Fatores associados aos sintomas depressivos avaliados pelo General Health Questionnaire (12 itens) em idosos residentes na comunidade (Projeto Bambuí)



Érico Castro-CostaI,II,III; Maria Fernanda Lima-CostaI,IV; Sandra CarvalhaisI; Josélia O A FirmoI; Elizabeth UchoaI,V

IPublic Health and Ageing Research Group (PHARG), Fundação Oswaldo Cruz e Universidade Federal de Minas Gerais (UFMG), Belo Horizonte (MG), Brazil
IIHealth Research Service Department, Institute of Psychiatry, King's College London, London, United Kingdom
IIISchool of Human Health and Ecology (FASEH), Vespasiano (MG), Brazil
IVDepartment of Preventive and Social Medicine, Medical School, Universidade Federal de Minas Gerais (UFMG), Belo Horizonte (MG), Brazil
VDepartment of Mental Health, Medical School, Universidade Federal de Minas Gerais (UFMG), Belo Horizonte (MG)





OBJECTIVE: To investigate factors associated with depressive symptoms in community-dwelling older adults.
This study evaluated 1,510 participants of the Bambuí Health Aging Study baseline. The dependent variable was the presence of depressive symptoms assessed by the 12-item General Health Questionnaire and predictive variables were sociodemographic characteristics, social support network, lifestyle and health conditions.
RESULTS: The prevalence of depressive symptoms was 38.5% (12-item General Health Questionnaire ≥ 5). Depressive symptoms were positively and independently associated with female gender (PR = 1.15; 95%CI 1.01-1.33), aged 80 years or over (PR = 1.22; 95%CI: 1.02-1.54) compared with 60-69 years, single (PR = 1.25; 95%CI: 1.02-1.46) or separated (PR = 1.30; 95%CI: 1.03-1.65) marital status, less than 4 years of schooling (PR = 1.42; 95%CI: 1.04-2.00), self-reported global health (reasonable: PR = 1.84; 95%CI 1.45-2.34; bad or very bad: PR = 2.44; 95%CI 1.91-3.12), incapacity or great difficulty in performing daily activities (PR = 1.39; 95%CI: 1.22-1.57) and complaint of insomnia in the last month (PR = 1.77; 95%CI: 1.22-1.99).
The similarities between factors associated with depressive symptoms in this population and in others do not explain the high prevalence rates previously reported in Bambuí.  These findings may guide efforts to investigate others factors to elucidate the etiopathogenesis of depression in this population.

Descriptors: Depression; Aged; Precipitating factors; Comparative study; Epidemiologic mesurements


OBJETIVO: Investigar os fatores associados aos sintomas depressivos em idosos residentes na comunidade.
MÉTODO: Este estudo seccional foi desenvolvido em 1.510 idosos, que correspondem a 86% do total de residentes na cidade de Bambuí-MG com 60 ou mais anos de idade. A variável dependente deste estudo é a presença de sintomas depressivos, determinada por meio do General Health Questionnaire (12 itens). As variáveis independentes incluíram características sociodemográficas, indicadores da rede social de apoio, estilos de vida e indicadores das condições de saúde.
RESULTADOS: A prevalência de sintomas depressivos foi de 38,5% (escore no General Health Questionnaire ≥ 5). Associações positivas e independentes com esses sintomas foram observadas para sexo feminino (RP = 1,15; IC95%: 1,01-1,33), faixa etária de 80 anos ou mais (RP = 1,22; IC95%: 1,02-1,54) comparada com idades entre 60 e 69 anos, ser solteiro (RP = 1,25; IC95%: 1,02-1,46) e ser separado (RP = 1,30; IC95%: 1,03-1,65), anos de escolaridade inferior a quatro anos (RP = 1,42; IC95%: 1,04-2,00), percepção da saúde como razoável (RP = 1,84; IC95%: 1,45-2,34) e ruim ou muito ruim (RP = 2,44; IC95%: 1,91-3,12), incapacidade funcional (RP = 1,39; IC95%: 1,22-1,57) e insônia nos últimos 30 dias (RP = 1,77; IC95%: 1,22-1,57).
CONCLUSÃO: Os fatores associados aos sintomas depressivos são semelhantes aos descritos em outros estudos e não explicam a alta prevalência de depressão encontrada em Bambuí. Esses achados demonstram a necessidade de investigação de outros fatores na tentativa de elucidar a etiopatogenia da depressão nessa população.

Descritores: Depressão; Idosos; Fatores desencadeantes; Estudo comparativo; Medidas em epidemiologia




Evidence suggests that depression is the most prevalent1 and relevant2,3 psychiatric condition present in late-life. There are wide variations in reported prevalences (0.4 to 35%). However, methodological differences between studies, particularly regarding sampling, definition and assessment of outcome preclude firm conclusions about cross-cultural and geographical differences.4

In the general population, where the great majority will not show any signs of psychopathology, clinical diagnostic interviews are often too expensive and time-consuming to justify their use for the purposes of mental health assessment in epidemiological research. This is particularly the case in contexts where mental health is not the primary focus of research. In such situations, a lay administered screening scale may be more effective.5

The General Health Questionnaire (GHQ)6 is probably the most widely used screening scale for mental disorders. It was selected by the World Health Organization (WHO) as the screening scale in a large multicentre primary care study because of its high sensitivity and specificity in various settings and cultures.6 The shorter 12-item version has been shown to be equally effective as the longer 28-item version in screening for common mental disorders.7 A previous study that simultaneously used the GHQ-12 and the GDS-30,8 a screening scale which was specially designed to evaluate elderly people, demonstrated that the GDS-30 did not perform better than the GHQ-12 and also was biased in similar ways.9 Tentatively, it was suggested that the GHQ-12 might be preferred, particularly where assessment of psychological morbidity across a wide age range, from younger to older adults, is required.9

Several studies have demonstrated that depression among the elderly is associated with many factors. A recent meta-analysis showed that among elderly community subjects bereavement, sleep disturbance, disability, prior depression and female gender appear to be the most important risk factors for current depression.10 To the best of our knowledge, there is only one Brazilian study about factors associated with depression in older adults, which was conducted in a probability sample of 583 elderly residents of the city of Pelotas-RS.11 However, this study has the limitation of applying a non-standardized instrument created by the authors using eight depressive symptoms commonly present in questionnaires and depression scales administered to the elderly.11

The aim of the present study was 1) to determine the factors associated with depressive symptoms in the baseline of the Bambuí Health Aging Study (BHAS) using the GHQ-12, and 2) to compare the findings with studies that applied scales specifically developed for the assessment of depression in older adults.



This study was accomplished at the baseline of the BHAS, which is a population-based study on health and aging. Bambuí is a city of 15,000 inhabitants in the state of Minas Gerais,12 where agriculture, dairy production and commerce are the main economic activities. The city has a general hospital (62 beds), one physician for every thousand inhabitants and no psychiatrist. Human development rate is 0.70 and life expectancy at birth is 70.2 years. The baseline was composed of 1606 individuals, corresponding to 92.2% of all Bambuí residents aged 60 years or older, identified by a census. Characteristics of participants were similar to those of the total population in the same age group regarding gender, number of persons in the home, marital status, family income and schooling. For the present study, all participants from the baseline who answered the GHQ-12 were selected (n = 1,510). Informed consent was obtained and the Ethical Committee of the Oswaldo Cruz Foundation approved the project.



1. Dependent variable

 The dependent variable of this study was the presence of depressive symptoms measured by the GHQ-12. It detects non-psychotic mental disorders and has been well studied in many settings and cultures,6 including Brazil.13 Recently, it was also validated for detecting depressive symptoms in the older population.9,14 Cut-off point ≥ 5 presented the best trade-off between sensitivity and specificity for the detection of depressive symptoms in the elderly population of Bambuí.9

2. Independent variables

The independent variables of this study included socio-demographic characteristics, indicators of social support network, lifestyle, and health conditions. The socio-demographic variables were gender, age group, marital status and number of complete years of schooling. As indicators of social support network, living with offspring, frequency of visits from offspring, relatives and neighbours in the last month were considered.

Lifestyle considerations included frequency of 20 to 30 minutes of physical exercise in the last 90 days, present consumption of cigarettes among those who had already smoked at least 100 cigarettes during their lifetime and episodes of excessive consumption of alcohol (5 or more drinks on a single occasion) in the last 30 days. Further information can be found in previous publications.15,16

Health condition indicators used in this study were: self-evaluation (determined by the question: “In general, how would you say your health is?”), impairment or great difficulty to perform at least one among five daily activities (dressing, eating, walking around inside the house, bathing and using the toilet), complaint of insomnia in the last 30 days,17 diabetes mellitus (fasting glycemia ≥ 126 mg/dL or present use of a hypoglycemic agent),18 high blood pressure (average of two among three measurements of blood pressure ≥ 140 and/or ≥ 90 mmHg or current,19 body mass index (BMI = kg/m2) and waist circumference.20

The interview was administered by lay interviewers who were selected from the community. They had at least 11 years of schooling and were intensely trained by a psychiatrist (Uchôa E.) for the standardized application of the interviews.


Statistical analysis

In cross-sectional studies, two measures of effect are more commonly used: the prevalence odds ratio (POR) and the prevalence ratio (PR).21,22 The choice between them has been a source of an ongoing debate in the epidemiological literature over the past few years.21-26 There is no dispute that both will be similar when the event is a rare disease, but they may be very discrepant for common diseases, which are often the focus of cross-sectional studies.22,23

The use of adjusted odds ratio to estimate adjusted relative risk is appropriate in studies of rare outcomes. However it may be misleading when the outcome is common and in this case the PR should be preferred.21-23 PR prevents the overestimation of the magnitude of association with a better control of confounders than the odds ratio (OR).21,23 PR is difficult to estimate in multivariate settings. The simplest way is to transform the odds ratio obtained by logistic regression into PR.27 Alternatives explored in the epidemiological literature are Cox regression,28 log-binomial regression,27 Poisson regression with robust estimates26 and complementary log-log model.29 In the present study, statistical analysis was based on crude and adjusted PR estimated by Robust Poisson Regression, which provides correct estimates of standard errors.26,30 The mathematical computations for PR are identical to the relative risk in cross-sectional studies.22

Gender and age group were considered a priori confounding variables and were included in all multivariate models. The initial multivariate model also included all variables associated with depressive symptoms in the univariate analysis, at a level of 0.20 or less. In the final model, only variables that maintained an association with these symptoms at a level of 0.05 or less were retained. A maximum-likelihood approach was used to establish the most parsimonious predictive model from the logistic regression. In this approach, each reduced hypothesis was tested against the general model (gender, age, marital status, schooling, physical exercise, alcohol consumption, self-reported health, impairment in daily activities, insomnia, waist circumference) using a likelihood ratio test. The most parsimonious model represents the combination of all non-rejected hypotheses and is desirable because it uses fewer parameters.  The statistical analyses were performed using STATA software, version 9.1. 



Among the 1,742 elderly residents of the study area, 86.7% answered the GHQ-12 (94 refused or had an incomplete GHQ-12).  Women (61.1%), younger elderly (59.8% were less than 70 years old) and those with low schooling levels (64.1% had never gone to school or had 4 or less years of schooling) comprised the largest groups. 38.5% of the participants presented depressive symptoms (GHQ-12 score ≥ 5).

In Table 1, the univariate and the multivariate analyses, adjusted by gender and age, for the association between depressive symptoms and sociodemographic characteristics are presented. Female gender, 80 or more years of age, separated marital status and less than 4 years of schooling were associated with depressive symptoms in the adjusted analysis.

Regarding the association of depressive symptoms with social support network, health conditions and lifestyle, only the absence of physical exercise during leisure time, perception of global health as reasonable or bad/very bad, impairment or difficulty in performing activities of daily life and insomnia were still associated with depressive symptoms after adjusting for gender and age (Table 2 and Table 3).

In Table 4, we display the results of the multivariate analysis simultaneously adjusted for all factors associated with depressive symptoms in the previous models. In this final model, female gender,  age 80 years or over, single or separated marital status, less than 4 years of schooling, reasonable or bad/very bad self- perception of global health, impairment in performing daily activities and insomnia remained independently associated with depressive symptoms measured by GHQ-12.




The results of this study show that the depressive symptoms among the elderly measured by GHQ-12 was associated with female gender, the oldest age group, the lowest levels of schooling, worse health conditions, impairment in daily activities and insomnia. These results are consistent with the literature10,31 and confirm previous observations in this population, where depressive symptoms were associated with worse health conditions among the older adult participants of Bambuí.32,33

A recent meta-analysis reported that 25 risk factors for depression were considered in studies using only univariate analysis while 15 factors were considered in studies using multivariate analysis.10 According to authors, all studies that used screening scales and multivariate analysis found that disability (4 studies), new medical illness (2) and sleep disturbances (2), poor self-perceived health (2 studies) and poor social support (1) were associated with depression. Poor health status and female gender were associated with depression in half of these studies whereas no association was reported with unmarried marital status and lower education.10

POR and PR have both been used as estimators of effect in cross-sectional studies.21 However, there are many evidences that the PR is more appropriate. It is more interpretable and more consistent for estimating the true effect. On the other hand, the POR is difficult to interpret as an effect measure (beyond specific settings) and can sometimes under- or overestimate the effect. Finally, the use of the POR will not necessarily lead to the same conclusions as that of the PR about effect modification or confounding.

As expected, comparing the PR used in the current study with the OR estimated in previous studies demonstrated an overestimation obtained from the OR.21,23 In studies which reported an association of depression with female gender the OR varied from 1.3 to 3.4,31 slightly  higher than the value found in the final model of the current study (PR = 1.15)  Association between aging and depression has been inconsistent10,31 and  has tended to disappear after adjustment for functional disability, co-morbid medical disorders and social deprivation.34 In our study, despite controlling for functional disability, co-morbid medical disorders and schooling (a variable related to social deprivation), depression was still more prevalent in older participants. This may have occurred because a screening test was used to define the presence of depressive symptoms. In a previous study in Bambuí, which used the Schedules for Clinical Assessment in Neuropsychiatry (SCAN) for diagnostic definition, this association was not observed.33         

The association of marital status with depressive symptoms is also inconsistent among studies.10,31 In the meta-analysis mentioned before, this association was not observed, maybe due to methodological differences. In the present study marital status was measured as a categorical variable (married/living together, single, separated, widowed) instead of a dichotomic one (married/unmarried).  Studies that demonstrated this association showed OR values from 1.1 to 1.9 for married, separated or divorced subjects,31 similar to the ones we obtained after adjustment for all other variables (PR = 1.25 for single and PR = 1.30 for separated marital status).

Associations between lower schooling level and depression have been frequently described, with OR from 1.5 to 1.8,31 similar to our results (PR = 1.42). However, some studies did not show this relationship because of different definitions of schooling level or because they were carried out in developed countries, where the level is much higher than in this study population.35,36

Regarding poor health status, many studies have shown this association using both screening37 and clinical diagnostic assessments.36,38 Findings of the current study are also consistent with this, showing a magnitude of association very similar to those presented in the literature.31 OR for depression with good, fair and poor health compared with excellent health were 1.8, 3.1 and 5.6, respectively,39 while in our study, PR were 1.84 and 2.44, for reasonable and bad/very bad health, in that order. Health-related functional impairment, especially in personal and instrumental Activities of Daily Living (ADL) was a risk factor for depressive disorders with an OR varying from 1.540 to 6.3,41 and a PR of 1.39 in our study.

Finally, sleep disturbances, particularly insomnia, are also frequently reported as independent predictors of depressive symptoms with magnitudes very similar to those of the current study.10,31 The other Brazilian study11 also found an association with female gender, oldest age and lower education. Other comparisons could not be made because the remaining factors investigated were different in each study.

Although our results are very similar to the literature, some methodological issues should be appraised. First, the response rate of this study was over 85%. Therefore the study sample has very high chances of being representative of this population. Second, the magnitude of associations was measured by the PR, which is more conservative than the OR, reducing chances of a type I error. Third, a validated screening scale was used to detect depressive symptoms, unlike in the previous Brazilian study.11

However, in the validation of the GHQ-12 and GDS-30 in this population,9 both scales performed insufficiently when compared with the SCAN/ICD-10 diagnosis and could not be recommended as efficient screening measures (the sum of sensitivity and specificity was less than 1.6),42 since their use might lead to high rates of misclassification. Also, the cross-sectional nature of the present study allows to establish associations in a specific time frame but causal inferences are precluded. Consequently, the results of the current study must be cautiously interpreted.

In conclusion, this is the first study that used a standardized instrument to investigate factors associated with depressive symptoms in elderly people in Brazil. Although, the GHQ-12 was not designed to assess this population, it did not perform worse than specific scales for this age band.9 The similarity between factors associated with depressive symptoms in this population and in others does not explain the high prevalence rates previously reported.32,33 Despite methodological limitations, our findings may guide efforts to investigate other factors to elucidate the etiopathogenesis of depression in this population.



This work was supported by grants of the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and Financiadora de Estudos e Projetos (FINEP).  MF Lima-Costa and E Uchoa are fellows of the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq). E Castro-Costa is supported by the Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG).




1. Blazer DG. Depression in late life: review and commentary. J Gerontol A Biol Sci Med Sci. 2003;58(3):249-65.         [ Links ]

2. Beekman AT, Penninx BW, Deeg DJ, de Beurs E, Geerling SW, van Tilburg W. The impact of depression on the well-being, disability, and use of services in older adults: a longitudinal perspective. Acta Psychiatr Scand. 2002;105(1):20-7.         [ Links ]

3. Penninx BW, Geerlings SW, Deeg DJ, van Eijk JT, van Tilburg W, Beekman AT. Minor and major depression and the risk of death in older persons. Arch Gen Psychiatry. 1999;56 (10):889-95.         [ Links ]

4. Beekman AT, Copeland JR, Prince MJ. Review of community prevalence of depression in later life. Br J Psychiatry. 1999;174:307-11.         [ Links ]

5.  Dunn G, Pickles A, Tansella M, Vázquez-Barquero JL. Two-phase epidemiological surveys in psychiatric research. Brit J Psychiatry. 1999;174:95-100.         [ Links ]

6. Goldberg D, Williams P.  A User's Guide to the General Health Questionnaire. Windsor: NFER - Nelson; 1988.         [ Links ]

7. Goldberg DP, Gater R, Sartorius N, Ustun TB, Piccinelli M, Gureje O, Rutter C. The validity of two versions of the GHQ in the WHO study of mental illness in general health care. Psychol Med. 1997;27(1):191-7.         [ Links ]

8. Yesavage JA, Brink TL, Rose TL, Lum O, Huang V, Adey M, Leirer VO. Development and validation of a geriatric depression screening scale: a preliminary report. J Psychiatr Res. 1982-1983;17(1):37-49.         [ Links ]

9. Costa E, Barreto SM, Uchoa E, Firmo JO, Lima-Costa MF, Prince M. Is the GDS-30 better than the GHQ-12 for screening depression in the elderly people in the community? The Bambui Health Aging Study (BHAS). Int Psychogeriatr. 2006;18(3):493-503.         [ Links ]

10. Cole MG, Dendukuri N. Risk factors for depression among elderly community subjects: a systematic review and meta-analysis. Am J Psychiatry. 2003;160(6):1147-56.         [ Links ]

11. Gazalle FK, Lima MS, Tavares BF, Hallal PC.Sintomas depressivos e fatores associados em população idosa no Sul do Brasil. Rev Saude Publica. 2004;38:365-71.         [ Links ]

12.  Costa MF, Uchoa E, Guerra HL, Firmo JO, Vidigal PG, Barreto SM. The Bambui health and ageing study (BHAS): methodological approach and preliminary results of a population-based cohort study of the elderly in Brazil. Rev Saude Publica. 2000;34(2):126-35.         [ Links ]

13. Mari JJ, Williams P. A comparison of the validity of two psychiatric screening questionnaires (GHQ-12 and SRQ-20) in Brazil using Relative Operating Characteristics (ROC) analysis. Psychol Med. 1985;15(3):651-9.         [ Links ]

14. Papassotiropoulos A, Heun R. Screening for depression in the elderly: a study on misclassification by screening instruments and improvement of scale performance. Progress Neuropsychopharmacol Biol Psychiatry. 1999;23 (3):431-46.         [ Links ]

15. Lima-Costa MF, Barreto SM, Uchôa E, Firmo JO, Vidigal PG, Guerra HL. The Bambui Health and Aging Study (BHAS): prevalence of risk factors and use of preventive health care services. Rev Panam Salud Publica. 2001;9(4):219-27.         [ Links ]

16. Prais HA, Loyola Filho AI, Firmo JO, Lima-Costa MF, Uchoa E. Estudo de base populacional sobre consumo excessivo de álcool entre homens idosos:evidências dos inquéritos de saúde de Belo Horizonte e Bambuí. Rev Bras Psiquiatr. 2008 (in press).         [ Links ]

17. Rocha FL, Guerra HL, Lima-Costa MF. Prevalence of insomnia and associated socio-demographic factors in a Brazilian community: The Bambui Study. Sleep Med. 2002;3(2):121-6.         [ Links ]

18. Passos VM, Barreto SM, Diniz LM, Lima-Costa. Type 2 diabetes: prevalence and associated factors in a Brazilian community - The Bambui health and aging study. Sao Paulo Med J. 2005;123(2):66-71.         [ Links ]

19. Barreto SM, Passos V, Firmo JO, Guerra HL, Vidigal PG, Lima-Costa MF. Hypertension and clustering of cardiovascular risk factors in a community in Southeast Brazil - The Bambui Health and Ageing Study. Arq Bras Cardiol. 2001;77(6):576-81.         [ Links ]

20. Leite ML, Nicolosi A, Firmo JO, Lima-Costa MF. Features of metabolic syndrome in non-diabetic Italian and Brazilians: a discriminant analysis. Int J Clin Pract. 2007;61(1):32-8.         [ Links ]

21. Thompson ML, Myers JE, Kriebel D. Prevalence odds ratio or prevalence ratio in the analysis of cross sectional data: what is to be done? Occup Environ Med. 1998;55(4):272-7.         [ Links ]

22. McNutt LA, Wu C, Xue X, Hafner JP. Estimating the relative risk in cohort studies and clinical trials of common outcomes. Am J Epidemiol. 2003;157(10):940-3.         [ Links ]

23.  Barros AJ, Hirakata VN. Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models tha directly estimate the prevalence ratio. BMC Med Res Methodol. 2003;3:21.         [ Links ]

24. Lee J, Chia KS. Use of the prevalence ratio v the prevalence odds ratio as a measure of risk in cross sectional studies. Occup Environ Med. 1994;51(12):841.         [ Links ]

25. Lee J. Odds ratio or relative risk for cross-sectional data? Int J Epidemiol. 1994;23(1):201-3.         [ Links ]

26. Behrens T, Taeger D, Wellmann J, Keil U. Different methods to calculate effect estimates in cross-sectional studies. A comparison between prevalence odds ratio and prevalence ratio. Methods Inf Med. 2004;43(5):505-9.         [ Links ]

27. Zocchetti C, Consonni D, Bertazzi PA. Estimation of prevalence rate ratios from cross-sectional data. Int J Epidemiol. 1995;24(5):1064-7.         [ Links ]

28. Lee J, Chia KS. Estimation of prevalence rate ratios for cross sectional data: an example in occupational epidemiology. Brit J Ind Med. 1993;50(9):861-2.         [ Links ]

29. Martuzzi M, Elliott P. Estimating the incidence rate ratio in cross-sectional studies using a simple alternative to logistic regression. Ann Epidemiol. 1998;8(1):52-5.         [ Links ]

30. Zou G. A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol. 2004;159(7):702-6.         [ Links ]

31. Djernes JK. Prevalence and predictors of depression in populations of elderly: a review. Acta Psychiatr Scand. 2006;113(5):372-87.         [ Links ]

32. Vorcaro CM, Lima-Costa MF, Barreto SM, Uchoa E. Unexpected high prevalence of 1-month depression in a small Brazilian community: the Bambuí Study. Acta Psychiatr Scand. 2001;104(4):257-63.         [ Links ]

33. Costa E, Barreto SM, Uchoa E,  Firmo JO, Lima-Costa MF, Prince M. Prevalence of International Classification of Diseases, 10th Revision Common Mental Disorders in the Elderly in a Brazilian Community: The Bambui Health Ageing Study. Am J Psychiatry. 2007;15:17-27.         [ Links ]

34. McDougall FA, Kvaal K, Matthews FE, Paykel E, Jones PB, Dewey ME, Brayne C; Medical Research Council Cognitive Function and Ageing Study. Prevalence of depression in older people in England and Wales: the MRC CFA Study. Psychol Med.  2007;37(12):1787-95.         [ Links ]

35. Livingston G, Watkin V, Milne B, Manela MV, Katona C. Who becomes depressed? The Islington study of older people. J Affect Disord. 2000;58(2):125-33.         [ Links ]

36. Schoevers RA, Beekman AT, Deeg DJ, Geerlings MI, Jonker C, van Tilburg W. Risk factors for depression in later life: results of a prospective community based study (AMSTEL). J Affect Disord. 2000;59(2):127-37.         [ Links ]

37. Harlow SD, Goldberg EL, Comstock GW. A longitudinal study of risk factors for depressive symptomatology in elderly widowed and married. Am J Epidemiol. 1991;134(5):526-38.         [ Links ]

38. Prince MJ, Harwood RH, Thomas A, Mann AH. A prospective population-based cohort study of the effects of disablement and social milieu on the onset and maintenance of lat-life depression. The Gospel Oak Project VII. Psychol Med. 1998;28(2):337-50.         [ Links ]

39. Hybels CF, Blazer DG, Pieper CF. Toward a threshold for subthreshold depression: an analysis of correlates of depression by severity of symptoms using data from an elderly community sample. Gerontologist. 2001;41(3):357-65.         [ Links ]

40. Minicuci N, Maggi S, Pavan M, Enzi G, Crepaldi G. Prevalence rate and correlates of depressive symptoms in older individual: the Veneto Study. J Gerontol A Biol Sci Med Sci. 2002;57(3):M155-61.         [ Links ]

41. Roberts RE, Kaplan GA, Shema SJ, Strawbridge WJ. Prevalence and correlates of depression in an aging cohort: the Alameda Country study. J Gerontol B Psychol Sci Soc Sci. 1997;52(5):S252-8.         [ Links ]

42. McNamee R. Efficiency of two-phase designs for prevalence estimation. Int J Epidemiol. 2003;32(6):1072-8.         [ Links ]



Erico de Castro e Costa
Av. Augusto de Lima, 1715 – Barro Preto
Belo Horizonte, MG, Brasil

Submitted: February 14, 2007
Accepted: December 12, 2007

Financial support: Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Financiadora de Estudos e Projetos (FINEP) and Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG)
Conflict of interests: None

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License