Journal of Affective Disorders 132 (2011) 223–230 Contents lists available at ScienceDirect Journal of Affective Disorders j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j a d Research report Prevalence and correlates of generalized anxiety disorder among older adults in the Australian National Survey of Mental Health and Well-Being Daniela C. Gonçalves a,⁎, Nancy A. Pachana b, Gerard J. Byrne a,c a b c School of Medicine, University of Queensland, Brisbane, Q 4029, Australia School of Psychology, University of Queensland, Brisbane, Q 4067, Australia Older Persons Mental Health Service, Royal Brisbane & Women's Hospital, Brisbane, Q 4029, Australia a r t i c l e i n f o Article history: Received 24 December 2010 Received in revised form 23 February 2011 Accepted 23 February 2011 Available online 22 March 2011 Keywords: Aging Generalized anxiety disorder Older adult CIDI a b s t r a c t Background: Generalized anxiety disorder (GAD) occurs commonly, with widespread consequences including decreased functioning and wellbeing, and increased consumption of health resources. Notwithstanding its prevalence and impact, knowledge about GAD in older adults is still scarce. Accordingly, the main goals of this study were to estimate the prevalence and analyze the correlates of 12-month DSM-IV GAD in older community-residing adults. Methods: The sample was drawn from the 2007 Australian National Survey of Mental Health and Well-Being and consisted of 3035 participants aged between 55 and 85 years, assessed by lay interviewers with the fully-structured Composite International Diagnostic Interview. Results: Eighty-four participants were diagnosed with GAD, equivalent to a weighted 12-month population prevalence of 2.8% (95% CI: 2.0, 3.7). In a multivariate logistic regression model older age (OR = 0.24, p = 0.006), functional limitations (OR = 1.07, p = 0.001), lifetime depression comorbidity (OR = 5.31, p b 0.001), concerns about having a serious illness despite doctor's reassurance (OR = 2.29, p = 0.021), and family history of anxiety or depression (OR = 2.41, p = 0.007) were the most significant predictors of 12-month GAD in older adults. Limitations: This was a cross sectional study, limiting causal inferences. Conclusions: In community-residing older adults GAD is highly prevalent and strongly associated with functional limitations, psychiatric comorbidity and increased medication intake. These findings suggest the need for greater clinical awareness of GAD among older adults. © 2011 Elsevier B.V. All rights reserved. 1. Introduction Despite a recent increase in activity, research on anxiety disorders in older adults is still relatively underdeveloped (Byrne and Pachana, 2010). This is reflected in the limited number of publications and scarceness of empirical data on this topic, which partially explain why many obsolete beliefs are still uncritically accepted. These include the belief that anxiety disorders have a lower prevalence in older adults or ⁎ Corresponding author at: Academic Discipline of Psychiatry, School of Medicine, University of Queensland, K Floor, Mental Health Centre, Royal Brisbane & Women's Hospital, Herston, Q 4029, Australia. Tel.: + 61 7 3365 5572; fax: + 61 7 3365 5488. E-mail address: d.goncalves@uqconnect.edu.au (D.C. Gonçalves). 0165-0327/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jad.2011.02.023 that anxiety disorders occur mainly within a depressive context (Wetherell et al., 2005). Although prevalence studies indicated that anxiety disorders in general (e.g., Regier et al., 1988), and generalized anxiety disorder (GAD) in particular (e.g., Hunt et al., 2002), decline in older, community-residing cohorts, relative prevalence is still high in later life (e.g., Byers et al., 2010; for a recent review about the prevalence of anxiety disorders in older adults, see Bryant et al., 2008). The impact of anxiety in older adults has been found to be pervasive, being associated with memory impairment (Mantella et al., 2007), burden of disability (Brenes et al., 2005), and the use of both mental health services and general health services (Porensky et al., 2009). GAD had also been reported to worsen the prognosis of major depressive disorder (MDD) (Steffens and McQuoid, 2005) and increase the risk of adverse cardiac events 224 D.C. Gonçalves et al. / Journal of Affective Disorders 132 (2011) 223–230 (Frasure-Smith and Lesperance, 2008), independently of MDD (Barger and Sydeman, 2005). It is therefore important to identify the factors that increase or decrease the likelihood of GAD diagnosis in order to better understand the condition and design of clinical intervention strategies with a higher likelihood of success (Smit et al., 2007). Factors associated with anxiety disorders in older cohorts have been investigated in several clinical and epidemiological studies, often with contradictory results. A recent study reviewed risk factors for anxiety disorders and symptoms in subjects aged 50 years and older (Vink et al., 2008). These authors considered that a set of biological (chronic health conditions, poor self perceived health, and functional limitations), psychological (external locus of control, neuroticism, lack of self-efficacy, and past psychiatric history), and social (social network, stressful events, and demographic characteristics) factors were evidence based for anxiety disorders in community residing older adults. Cognitive impairment, high blood pressure, sensory loss, limited coping strategies, poor self-image, and low income were judged as risk factors for anxiety symptoms (Vink et al., 2008). The same study also analyzed factors associated with depressive symptoms and disorders in older adults. These include exercise, use of psychotropic medicine, alcohol use and smoking (Vink et al., 2008). Against this background, the principal goals of the present study were to estimate the 12-month prevalence of DSM-IV GAD and analyze the factors associated with this diagnosis in a representative sample of community-dwelling adults aged 55 years and older. We defined a model for GAD in older adults with putative factors derived from social, psychological, and biological dimensions. Socio-demographic factors included age, gender, education, marital status, country of birth, living arrangements, urbanization, financial problems, stressful life events, social network, and caregiver status. Psychological factors were past or current history of major depressive disorder and worry about illness despite doctor's reassurance. Biological factors included chronic diseases, functional limitations, self-assessed physical health, intake of anxiety related medication, lifestyle characteristics (exercise, alcohol intake, and smoking), and family history of anxiety or depression. These factors were established based on previous reported results concerning GAD in older adults, and additional putative factors were also included, selected for their relevance for mental health in older cohorts. 2. Methods 2.1. Sample and measures The sample was drawn from the Confidentialised Unit Record File (CURF) of the National Survey of Mental Health and Wellbeing (NSMHWB), conducted by the Australian Bureau of Statistics (ABS) in 2007, through a multi-stage sampling of people aged 16–85 years old living in private dwellings (ABS, 2007). A detailed description of the survey has been provided elsewhere (Slade et al., 2009). A total of 14,805 private households were randomly selected, with a national response rate of 60%, including 8841 full responses. For the present analyses, participants aged 55 years or older were considered (N = 3178; 53% female). Data were collected by trained lay interviewers. The Composite International Diagnostic Interview, version 3.0 (CIDI 3.0), (Kessler and Ustun, 2004), was employed to collect self-report data regarding anxiety, affective, and substance abuse disorders, generating ICD-10 and DSM-IV diagnoses. Additional information collected included socio-demographic characteristics, chronic medical conditions, disability, lifetime and current use of health services, traumatic life experiences, social network, and family history of mental illness. 2.2. Putative associated factors Social factors were assessed through demographic information (age, gender, education, marital status, country of birth, number of persons in the household, urbanization level, and financial problems) and social characteristics (stressful life events, social network, and caregiver status). For the present analysis, educational level (completion of 8 years or less of formal schooling or below vs. completion of 9 years or more), marital status [not married (single, divorced, widowed) vs. married], country of birth (Australia vs. other), number of persons in the household (living alone vs. living with others), urbanization level (major urban center vs. other), and financial problems (none vs. one or more) were dichotomized. Financial problems were ascertained by a list of ten questions concerning money shortage over the last 12 months (e.g., inability to pay bills). Lifetime significant life events were queried through a list of 27 potentially stressful experiences, including traumatic personal experiences (e.g., combat experience), events involving personal violence (e.g., beaten up as a child), and events affecting others (e.g., child's serious illness). Social network composition was ascertained by asking about the frequency of contact with family members (never; monthly or less; weekly; or daily). Caregiver status was defined as spending time providing instrumental or emotional care to close relatives because of their health problems. Psychological factors consisted of past or current history of major depressive disorder (MDD) and worry about illness despite receiving reassurance from a doctor. Past or current history of MDD was derived from the lifetime diagnosis of MDD, based on DSM-IV criteria. Illness concerns were determined by a dichotomous question. Biological factors were ascertained through lifetime history of chronic diseases, functional limitations, frequency of weekly exercise, selfassessed physical health, intake of anxiety related medication, smoking and alcohol intake frequency, and family history of anxiety or depression. Chronic diseases were based on the Australian National Health Priority Areas. Participants were asked whether they were ever told by a doctor or a nurse that they had asthma, cancer, stroke, heart or circulatory condition, rheumatism, or diabetes. Obesity, also being a national priority area, was included and defined as having a body mass index (BMI) of 30 or more, calculated from self-reported height and weight information. Functional limitations were measured on the 12-item WHO Disability Assessment Schedule, using the simple sum method to obtain the total score (Andrews et al., 2009). Frequency of vigorous weekly exercise was inquired through an open question and answers were converted into ordinal categories (never; 1–3 times per week; more than 3 times per week). Self-assessed physical health was assessed on a 3-point Likert scale (poor/ D.C. Gonçalves et al. / Journal of Affective Disorders 132 (2011) 223–230 fair, good, very good/excellent). Intake of tablets for sleep and anxiety or nerves was used to dichotomously determine medication intake, whereas alcohol and smoking habits were determined by their frequency (never; weekly or less; daily). Family history of anxiety or depression was obtained by inquiring about the presence of these disorders in close relatives. 2.3. Data analyses Data analyses were conducted in Stata 11.1 (StataCorp, 2009) using survey data routines. Individualized person weights were used to allow population estimates to be calculated. Standard errors of prevalence estimates and confidence intervals around odds ratios were calculated on the basis of 60 delete-one jackknife replications using replicate weights provided by the Australian Bureau of Statistics. Reported results for the prevalence estimates take into account the probability of being sampled and have been standardized to the projected 2007 age and sex distribution of the Australian population based on the 2006 national population census (ABS, 2007). Data were analyzed through logistic regression models, using the 12-month prevalence of GAD as the outcome variable and the previously described factors as the putative associated variables. The hierarchical rule, which states that GAD cannot occur exclusively during a mood disorder (APA, 1994), was suspended. Previous studies indicated that there were no significant differences between outpatients diagnosed with GAD with or without this hierarchical rule applied, but they were quite distinct from 225 outpatients diagnosed with MDD, namely reporting more impaired social functioning (Zimmerman and Chelminski, 2003). In order to decrease confounding, participants who did not present with 12-month GAD symptoms but were diagnosed with lifetime GAD were excluded from subsequent analysis (n = 143), rendering a sample of 3035 participants aged 55 years and older. Bivariate logistic regression analyses were used to assess the statistical significance of each of the variables against the presence of GAD, after which all the significant variables were included in a multivariate logistic regression model. An inclusive alpha level for statistical significance was established at α b 0.25 (Hosmer and Lemeshow, 2000) on the bivariate analysis. For the multivariate model statistical significance was considered at α b 0.05. 3. Results 3.1. Sample characteristics The sample consisted of 3035 older adults (M = 68, SD = 8, range 55–85 years, 52% female), the majority of whom were born in Australia (70%) and lived in a major urban center (59%). Approximately one fifth of the participants had completed 8 years or less of formal schooling. Almost 56% of the participants were married, and approximately 40% lived alone, although almost all participants (90%) were in contact with family members at least weekly. A minority of the participants (15%) reported one or more financial problems. One fifth of the participants reported no stressful life events, whereas more than half reported two or Table 1 Socio-demographic characteristics of older adults with and without GAD and bivariate logistic regression analyses. Variable Age 55–64 65–74 75–85 Gender Female Education 9 years or more Marital status Married Country of birth Outside Australia Living arrangements Living with someone Urbanicity Outside major urban center Financial problems One or more Frequency family contact Never Monthly Weekly Daily Caregiver status Does not provide care Life events No 12-month GAD (n = 2951) 12-month GAD (n = 84) N (%) a N (%) a OR (95% CI) t p −3.1 −3.6 1132 (47.6%) 1044 (30.9%) 775 (21.5%) 58 (76.4%) 16 (16.2%) 10 (7.4%) 1 0.33 (0.16, 0.67) 0.22 (0.09, 0.50) 1523 (50.5%) 55 (66.2%) 1.91 (1.01, 3.61) 2.05 0.045 2282 (78.3%) 68 (80.8%) 1.17 (0.56, 2.43) 0.42 0.673 1652 (69.5%) 40 (66.3%) 0.86 (0.49, 1.51) − 0.54 0.590 878 (32.3%) 26 (38.9%) 1.33 (0.68, 2.61) 0.85 0.396 1810 (77.2%) 42 (70.2%) 0.70 (0.42, 1.16) − 1.41 0.164 1206 (37.8%) 41 (37.8%) 1.00 (0.57, 1.77) − 0.00 0.999 155 (4.58%) 20 (24.2%) 6.65 (3.17, 13.96) 5.12 b 0.001 4 (3.0%) 9 (6.2%) 23 (25.0%) 48 (65.8%) 1 0.24 (0.05, 1.26) 0.28 (0.06, 1.24) 0.34 (0.08, 1.38) −1.72 −1.71 −1.54 0.091 0.092 0.128 20 (31.4%) M (SD) 3.32 (2.63) 2.48 (1.32, 4.68) 2.87 0.006 1.21 (1.12, 1.31) 5.06 b 0.001 37 280 945 1689 (1.0%) (8.1%) (29.0%) (62.0%) 418 (15.6) M (SD) 2.11 (2.04) GAD: generalized anxiety disorder; OR: odds ratio; CI: confidence interval. a Values correspond to raw numbers and population weighted prevalence. 0.003 0.001 226 D.C. Gonçalves et al. / Journal of Affective Disorders 132 (2011) 223–230 more. Almost 10% of the participants were diagnosed with lifetime MDD, which included both past and current episodes. Almost 80% of the participants reported one or more chronic diseases and mean disability score on the WHODAS was 4 (SD = 6, range 0–48). Less than one fifth of the participants assessed their health as “fair” or “poor”, whereas half reported “very good” or “excellent” physical health. Family history of either anxiety or depression was reported by 23% of the participants. 3.2. Prevalence and correlates of GAD Eighty-four participants were diagnosed with 12-month GAD, according to DSM-IV criteria with the hierarchical rule suspended, which is equivalent to a raw prevalence of 2.8% (95% CI: 2.2, 3.4) and a weighted population prevalence of 2.8% (95% CI: 2.0, 3.7). The majority of those diagnosed with GAD were women (n = 55, corresponding to two thirds of the participants diagnosed with GAD). Forty-five participants, which is equivalent to 54% of the GAD sub-sample, had both 12-month GAD and MDD, the majority of whom were also women (n = 27, 26% of the GAD subsample). There were significant differences between survey participants with and without GAD on biological, psychological and social factors. Women had greater odds of being diagnosed with GAD than men (OR = 1.91, 95% CI: 1.01, 3.61). Older participants were less likely than younger participants to have GAD (OR = 0.33, 95% CI: 0.16, 0.67, for the category 65–74 and OR= 0.22, 95% CI: 0.09, 0.50, for the category 75–85, when compared with 55–64). The odds of GAD increased as the number of stressful life events experienced increased (OR = 1.21, 95% CI: 1.12, 1.31), as the number of reported chronic diseases increased (OR = 1.36, 95% CI: 1.05, 1.75), and as functional limitations increased (OR = 1.10, 95% CI: 1.06, 1.13). Financial problems (OR = 6.65, 95% CI: 3.17, 13.96), providing instrumental and emotional care to close relatives (OR = 2.48, 95% CI: 1.32, 4.68), lifetime diagnosis of MDD (OR = 12.45, 95% CI: 6.70, 23.12), and worrying about illness despite reassurance from a doctor (OR = 6.22, 95% CI: 3.47, 11.14) were all significantly related with increased odds of 12month GAD. Conversely, living with someone (OR = 0.70, 95% CI: 0.42, 1.16), higher frequency of family contact (OR = 0.24, 95% CI: 0.05, 1.26, for the category “monthly” when compared with “never”), more exercise (OR = 0.19, 95% CI: 0.02, 1.59, for the category “more than 3 times per week” when compared with “never”), and better self-assessed health (OR = 0.53, 95% CI: 0.28, 1.00, for the category “good”, and OR = 0.19, 95% CI: 0.08, 0.41, for the category “very good/excellent”, when compared with “poor/fair”) were all associated with decreased odds of 12-month GAD. Among putative biological factors, smoking (OR = 2.24, 95% CI: 0.93, 5.44, for the category “daily” when compared with “never”), family history of anxiety or depression (OR = 4.30, 95% CI: 2.59, 7.12) and self-reported anxiety medication intake (OR = 4.94, 95% CI: 2.59, 9.44) were both associated with an increased odds of GAD. However, country of birth, educational level, marital status, urban location, and alcohol intake were not significantly associated with GAD. There were also no significant results for some of Table 2 Clinical characteristics of older adults with and without GAD and bivariate logistic regression analyses. Variable Lifetime MDD diagnosis Yes Worry about illness Yes Exercise frequency (per week) Never 1–3 times week More than 3 times Self assessed health Poor/fair Good Very good/excellent Anxiety medication intake Yes Alcohol intake Never Weekly or less Daily Smoking Never Weekly or less Daily Family history anxiety/depression Yes Chronic diseases Functional limitations No 12-month GAD (n = 2951) 12-month GAD (n = 84) N (%) a N (%) a OR (95% CI) t p 241 (7.4%) 45 (49.9%) 12.45 (6.70, 23.12) 8.15 b 0.001 242 (8.9%) 33 (37.7%) 6.22 (3.47, 11.14) 6.27 b 0.001 2422 (80.3%) 358 (13.7%) 171 (6.0%) 68 (84.2%) 14 (14.6%) 2 (1.1%) 1 1.02 (0.52, 1.98) 0.19 (0.02, 1.59) 0.05 −1.56 0.960 0.123 608 (20.3%) 1007 (34.3%) 1336 (45.4%) 36 (43.3%) 33 (38.7%) 15 (18.0%) 1 0.53 (0.28, 1.00) 0.19 (0.08, 0.41) −1.99 −4.26 0.051 b 0.001 289 (8.8%) 28 (32.3%) 4.94 (2.59, 9.44) 4.94 b 0.001 967 (31.3%) 1198 (41.0%) 786 (27.6%) 32 (38.1%) 33 (37.7%) 19 (24.3%) 1 0.76 (0.40, 1.44) 0.72 (0.32, 1.66) −0.86 −0.78 0.391 0.439 2576 (87.4%) 18 (1.7%) 337 (10.9%) 66 (76.9%) 2 (1.6%) 16 (21.6%) 1 1.07 (0.15, 7.68) 2.24 (0.93, 5.44) 0.07 1.83 0.942 0.072 4.30 (2.59, 7.12) 5.78 b 0.001 1.36 (1.05, 1.75) 1.10 (1.06, 1.13) 2.39 5.84 0.020 b 0.001 636 M 1.68 3.68 (22.1%) (SD) (1.26) (5.74) GAD: generalized anxiety disorder; OR: odds ratio; CI: confidence interval. a Values correspond to raw numbers and population weighted prevalence. 44 M 2.22 10.17 (55.0%) (SD) (1.60) (9.88) D.C. Gonçalves et al. / Journal of Affective Disorders 132 (2011) 223–230 the categories of ordinal variables, namely daily family contact and smoking weekly or less. Tables 1 and 2 summarize the frequencies for both groups as well as the results of the bivariate analyses. The factors that attained statistical significance in the bivariate analyses were then included in a multivariate logistic regression model (Table 3). The Hosmer–Lemeshow statistic indicated satisfactory goodness-of fit for the model [χ2(2818) = 2580.47, p = 0.99], correctly classifying approximately 97% of the sample and presenting an area under the receiver operating characteristic (ROC) curve of 0.89. The model was free of multicollinearity, with a variance inflation factor (VIF) of 1.14 and a tolerance value of 0.88. In the final model, age (OR = 0.24, 95% CI: 0.09, 0.65, p = 0.006, for the category 75–85 when compared with 55–65), lifetime MDD (OR = 5.31, 95% CI: 2.39, 11.78, p b 0.001), worry about illness (OR = 2.29, 95% CI: 1.14, 4.61, p = 0.021), functional limitations (OR = 1.07, 95% CI: 1.03, 1.12, p = 0.001), and family history of anxiety or depression (OR = 2.41, 95% CI: 1.28, 5.55, p = 0.007) were significantly associated with 12month GAD. Table 3 Multivariate logistic regression analysis of generalized anxiety disorder. Variable Age 55–64 65–74 75–85 Gender Female Living arrangements Living with someone Financial problems One or more Frequency family contact Never Monthly Weekly Daily Caregiver status Yes Stressful life events Lifetime diagnosis of MDD Yes Worry about illness Yes Exercise frequency (per week) Never 1–3 times week More than 3 times Chronic illnesses Functional limitations Self assessed health Poor/fair Good Very good/excellent Anxiety medication intake Yes Smoking Never Weekly or less Daily Family history of anxiety or depression Yes OR (95% CI) t 1 0.35 (0.14, 0.86) 0.24 (0.09, 0.65) −2.33 −2.88 0.023 0.006 1.48 (0.59, 3.71) 0.86 0.394 0.76 (0.42, 1.42) −0.84 0.403 1.10 (0.74, 1.66) 0.49 0.624 1 0.33 (0.07, 1.54) 0.45 (0.12, 1.72) 0.69 (0.19, 2.59) −1.44 −1.19 −0.55 0.156 0.240 0.581 0.98 (0.52, 1.85) 1.04 (0.92, 1.17) −0.07 0.56 0.946 0.575 5.31 (2.39, 11.78) 4.19 b0.001 2.29 (1.14, 4.61) 2.37 0.021 2.65) 2.39) 1.30) 1.12) 0.44 −1.28 0.23 3.41 0.662 0.205 0.817 0.001 1 1.57 (0.73, 3.38) 0.77 (0.32, 1.82) 1.18 −0.61 0.244 0.541 2.31 (0.99, 5.34) 1.98 0.053 1 1.10 (0.54, 2.23) 1.35 (0.40, 4.52) 0.26 0.50 0.792 0.622 2.41 (1.28, 5.55) 2.78 0.007 1 1.19 0.21 1.03 1.07 (0.54, (0.02, (0.81, (1.03, p 227 4. Discussion In a population-based sample of community-dwelling older adults the weighted 12-month prevalence of non-hierarchical DSM-IV GAD was 2.8%, similar to values previously reported (Byers et al., 2010). The present prevalence finding is considerably lower than the one reported by the Longitudinal Aging Study Amsterdam (LASA), which found a 6-month prevalence for DSM-III GAD in community living adults aged between 55 and 85 to be 7.3% (Beekman et al., 1998). This significant difference is partially explained by changes in the DSM criteria between DSM-III and DSM-IV, particularly the introduction of the excessiveness criterion and the increase in the minimum required duration of symptoms from one month to six months, thus being more restrictive. A study that manipulated DSM-IV diagnostic criteria for GAD found that when the excessiveness criterion was suspended and the duration of anxiety was changed to one month or more the 12-month prevalence of GAD in the entire sample of the NCS-R rose from 3.0 to 6.2% (Ruscio et al., 2007). In the current analysis, age, history of lifetime comorbid MDD, functional limitations, and family history of anxiety or depression were the factors most significantly associated with GAD in older adults. Worry about illness despite a doctor's reassurance was also associated with GAD, albeit with relatively wide confidence intervals. These findings are consistent with previously published work. Cross-sectional studies usually provide lower prevalence estimates for GAD in older adults in comparison with younger adults (Hunt et al., 2002), and longitudinal findings indicate that increasing age is associated with decreasing severity of GAD symptoms in middle-aged adults (Ramsawh et al., 2009). Several methodological explanations have been put forward to explain the apparent decline of anxiety disorders with age, including the systematic exclusion of impaired, frail or institutionalized older adults (Jeste et al., 2005) and a subtle shift in phenomenology that is not detected by the traditionally employed diagnostic criteria (Flint, 2005). Even though the present study did not survey institutionalized older adults, many participants reported functional impairment and chronic disorders, thus frailty and impairment were represented to some extent. The phenomenological characteristics of GAD in older adults, which might be overlooked by the current assessment techniques, warrant further investigation to properly distinguish age effects from cohort effects. Alternatively, decreasing prevalence might be attributable to substantive rather than methodological factors. According to the socioemotional selectivity theory emotional goals become more salient with age, with older adults exhibiting age-differentiated strategies of emotional control and tending to engage in strategies to optimize social exchanges and relationships (Carstensen et al., 1999). Additionally, it has been shown that older adults display a memory bias that favors positive information (Mather and Carstensen, 2003), experiencing the same number of positive emotions and fewer negative emotions than their younger counterparts (Carstensen et al., 2000). Functional limitations had a direct and significant association with GAD. Because functional limitations increase with age they might create a specific vulnerability factor for older cohorts. Functional limitations and their associated disability 228 D.C. Gonçalves et al. / Journal of Affective Disorders 132 (2011) 223–230 have been frequently reported in association with anxiety disorders in older cohorts, in both cross-sectional and longitudinal studies (Vink et al., 2008). The mechanisms that underlie the interplay between functional limitations and anxiety are not yet established but a dynamic and reciprocal interaction between mental and physical health has been postulated. Anxiety symptoms are themselves disabling, as well as impacting on different physiological systems, including the cardiovascular and respiratory systems, thus increasing the likelihood of disability (Culpepper, 2009). In a recent study that followed more than 1000 older women for 3 years, anxiety symptoms were found to be an independent predictor of disability, both in activities of daily living and light housework chores (Brenes et al., 2005). On the other hand, functional limitations at baseline and low performance on basic or instrumental activities of daily living were significant risk factors for GAD and mixed anxiety and depression (Schoevers et al., 2005). We also found that a comorbid lifetime diagnosis of MDD was highly related with GAD in older adults. In addition, almost half of those diagnosed with GAD were also diagnosed with MDD, which is consistent with the findings of other studies of older cohorts (Manela et al., 1996; Schoevers et al., 2005). Several hypotheses exist to explain the high comorbidity between disorders, including shared genetic risk factors (Cerda et al., 2010; Kendler et al., 1992) and social and functional vulnerability created by GAD that subsequently triggers depression (Simon, 2009). However, biological data corroborate that MDD and GAD are distinct disorders (Goldberg, 2008). Furthermore, recent longitudinal findings indicate that although the association between GAD and MDD is significant, it is similar to that observed between GAD and other anxiety disorders, and even MDD and other anxiety disorders (Beesdo et al., 2010). The impact of family history of anxiety or depression upon the development of GAD can be explained by genetic and behavioral mechanisms, and the interplay between the two. In a study conducted with almost 4000 pairs of older twins, approximately one fourth of the variation in GAD was explained by genetic factors (Mackintosh et al., 2006). Although less evidence exists for the role played by family environment (Hettema et al., 2002), it has been suggested that attachment and vicarious learning during childhood increase the likelihood of having GAD as an adult (McLaughlin et al., 2008). Another relevant aspect to take into account is that participants diagnosed with GAD might have a selective memory for family history of mental disorder. The relevance of family history of anxiety or depression for current anxiety disorders was found to decrease with age (Beekman et al., 1998). This was also partially replicated in our sample, as the stratified analysis for age groups indicated that family history was more significant for those aged between 55 and 65 years. These results might indicate that as age increases environmental factors are more relevant or reflect memory bias in older cohorts (Beekman et al., 1998). It might also be the case that older adults are less likely to identify and report anxiety symptoms in themselves (Palmer et al., 1997). Our finding concerning worry about illness despite a doctor's reassurance should be interpreted with caution, as it was associated with a wide confidence interval. As excessive worry is the cornerstone of GAD it is natural that concerns about health and illness are frequently reported. However, being worried about a serious illness even when there is no medical evidence for it is also the main diagnostic criterion for hypochondriasis, a disorder found in frequent comorbid relation to GAD (Barsky et al., 1992). A similar trend was found in our sample, where almost 40% of the participants with GAD was worried about having a serious illness despite doctor's reassurance, whereas less than one tenth of those without the disorder expressed the same concern. There is further evidence for the tight association between GAD and illness worry. A study about the natural course of GAD indicated that at 40-year follow-up somatoform disorders were the most frequent outcome (Rubio and Lopez-Ibor, 2007). Because ours was a cross-sectional study and diagnosis was not established for hypochondriasis, caution is required when interpreting the results. However, it seems likely that worrying about a serious illness creates an additional vulnerability factor for those with GAD or that having GAD increases concern about having a serious illness in the absence of sufficient evidence for such. The strengths of our study include its use of a large, population-based, sample of older adults and a fully structured psychiatric interview. There are however some limitations. Although ours was a representative epidemiologic sample, it was not specifically designed to assess risk factors for GAD in older adults. Some factors previously identified as relevant to the etiology of GAD, such as personality traits, were not included. Biological variables were scarce in the survey dataset and the inclusion of relevant information for the understanding of anxiety disorders, namely imaging data, would have provided a richer set of results. However, the present dataset was comprehensive and comprised factors from multiple domains, allowing for a comprehensive analysis of GAD. The sampling strategy excluded people aged 86 and older, as well as institutionalized persons, thus excluding frailer or more impaired older adults. Nevertheless frailty and impairment were represented in the sample population, as indicated by the functional limitations and cognitive screening results. Both the DSM-IV diagnostic criteria themselves and the CIDI instrument used to assess them were initially developed to assess younger populations, and it remains uncertain whether either is entirely suitable for use with older adults (O'Connor and Parslow, 2010). On the other hand the CIDI is a detailed structured interview especially developed for epidemiological research (Kessler and Ustun, 2004), which has been used with older adults before, and no consensus has yet emerged about how younger and older adults might differ in their anxiety symptom presentation (Wolitzky-Taylor et al., 2010). The final limitation concerns the cross sectional nature of the design, which limits causal inferences. This study analyzed the factors associated with GAD in older adults, using a population-based representative sample of Australians. Along with depression, anxiety disorders are the leading cause of incident non-fatal disease burden in Australia (Mathers et al., 1999), representing over 14% of the years lost due to disability. Unfortunately, the majority of the identified factors are not modifiable. Nevertheless, by identifying risk factors for GAD it might be possible to design effective preventive or early intervention strategies (Smit et al., 2007), ultimately decreasing medical costs and D.C. Gonçalves et al. / Journal of Affective Disorders 132 (2011) 223–230 increasing quality of life and disability-free years of life for older people. Data support the efficacy of both behavioral and pharmacological interventions for anxiety disorders in older people (Pinquart and Duberstein, 2007). From a clinical perspective, the model we have described allows the identification of community-residing older adults with a higher likelihood of having GAD, thus making them potentially identifiable and suitable candidates for interventions. Despite the existence of previously published research, there are still several unanswered questions about this topic, especially when compared with the available data for other disorders on this population, namely depressive disorders (Vink et al., 2008). Furthermore, it provides timely information about GAD on the Australian context. Role of funding source Daniela Gonçalves was supported by a scholarship from the Fundação para a Ciência e Tecnologia, Portugal (SFRH/BD/30226/2006) and financial support from an anonymous donor. Gerard Byrne was supported by grants from the Australian National Health and Medical Research Council (456182) and the US Alzheimer's Association (IIRG-07-59015). None of the institutions had further role in study design; in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication. Conflict of interest No conflict declared. References Andrews, G., Kemp, A., Sunderland, M., Von Korff, M., Ustun, T.B., 2009. Normative data for the 12 item WHO Disability Assessment Schedule 2.0. PLoS ONE 4, e8343. APA, 1994. Diagnostic and Statistical Manual of Mental Disorders: DSM-IV. American Psychiatric Association, Washington, DC. Australian Bureau of Statistics, 2007. National Survey of Mental Health & Well Being, Basic CURF, CD-ROM. Barger, S.D., Sydeman, S.J., 2005. Does generalized anxiety disorder predict coronary heart disease risk factors independently of major depressive disorder? J. Affect. Disord. 88, 87–91. Barsky, A.J., Wyshak, G., Klerman, G.L., 1992. Psychiatric comorbidity in DSM-III-R hypochondriasis. Arch. Gen. Psychiatry 49, 101–108. Beekman, A.T.F., Bremmer, M.A., Deeg, D.J.H., van Balkom, A.J.L.M., Snut, J.H., de Beurs, E., van Dyck, R., van Tilburg, W., 1998. Anxiety disorders in later life: a report from the Longitudinal Aging Study Amsterdam. Int. J. Geriatr. Psychiatry 13, 717–726. Beesdo, K., Pine, D.S., Lieb, R., Wittchen, H.U., 2010. Incidence and risk patterns of anxiety and depressive disorders and categorization of generalized anxiety disorder. Arch. Gen. Psychiatry 67, 47–57. Brenes, G.A., Guralnik, J.M., Williamson, J.D., Fried, L.P., Simpson, C., Simonsick, E.M., Penninx, B., 2005. The influence of anxiety on the progression of disability. J. Am. Geriatr. Soc. 53, 34–39. Bryant, C., Jackson, H., Ames, D., 2008. The prevalence of anxiety in older adults: Methodological issues and a review of the literature. J. Affect. Disord. 109, 233–250. Byers, A.L., Yaffe, K., Covinsky, K.E., Friedman, M.B., Bruce, M.L., 2010. High occurrence of mood and anxiety disorders among older adults: the National Comorbidity Survey Replication. Arch. Gen. Psychiatry 67, 489–496. Byrne, G.J., Pachana, N.A., 2010. Anxiety and depression in the elderly: do we know any more? Curr. Opin. Psychiatry 23, 504–509. Carstensen, L.L., Isaacowitz, D.M., Charles, S.T., 1999. Taking time seriously — a theory of socioemotional selectivity. Am. Psychol. 54, 165–181. Carstensen, L.L., Pasupathi, M., Mayr, U., Nesselroade, J.R., 2000. Emotional experience in everyday life across the adult life span. J. Pers. Soc. Psychol. 79, 644–655. Cerda, M., Sagdeo, A., Johnson, J., Galea, S., 2010. Genetic and environmental influences on psychiatric comorbidity: a systematic review. J. Affect. Disord. 126, 14–38. Culpepper, L., 2009. Generalized anxiety disorder and medical illness. J. Clin. Psychiatry 70, 20–24. Flint, A.J., 2005. Anxiety and its disorders in late life — moving the field forward. Am. J. Geriatr. Psychiatry 13, 3–6. 229 Frasure-Smith, N., Lesperance, F., 2008. Depression and anxiety as predictors of 2-year cardiac events in patients with stable coronary artery disease. Arch. Gen. Psychiatry 65, 62–71. Goldberg, D., 2008. Towards DSM-V: the relationship between generalized anxiety disorder and major depressive disorder. Psychol. Med. 38, 1671–1675. Hettema, J.M., Neale, M.C., Kendler, K.S., 2002. A review and meta-analysis of the genetic epidemiology of anxiety disorders. Am. J. Psychiatry 5, 92. Hosmer, D.W., Lemeshow, S., 2000. Applied Logistic Regression. John Wiley & Sons, Inc., USA. Hunt, C., Issakidis, C., Andrews, G., 2002. DSM-IV generalized anxiety disorder in the Australian National Survey of Mental Health and WellBeing. Psychol. Med. 32, 649–659. Jeste, D.V., Blazer, D.G., First, M., 2005. Aging-related diagnostic variations: need for diagnostic criteria appropriate for elderly psychiatric patients. Biol. Psychiatry 58, 265–271. Kendler, K.S., Neale, M.C., Kessler, R.C., Heath, A.C., Eaves, L.J., 1992. Major depression and generalized anxiety disorder: same genes, (partly) different environments. Arch. Gen. Psychiatry 49, 716–722. Kessler, R.C., Ustun, T.B., 2004. The World Mental Health (WMH) Survey Initiative Version of the World Health Organization (WHO) Composite International Diagnostic Interview (CIDI). Int. J. Meth. Psychiatr. Res. 13, 93–120. Mackintosh, M.A., Gatz, M., Wetherell, J.L., Pedersen, N.L., 2006. A twin study of lifetime generalized anxiety disorder (GAD) in older adults: genetic and environmental influences shared by neuroticism and GAD. Twin Res. Hum. Genet. 9, 30–37. Manela, M., Katona, C., Livingston, G., 1996. How common are the anxiety disorders in old age? Int. J. Geriatr. Psychiatry 11, 65–70. Mantella, R.C., Butters, M.A., Dew, M.A., Mulsant, B.H., Begley, A.E., Tracey, B., Shear, M.K., Reynolds III, C.F., Lenze, E.J., 2007. Cognitive impairment in late-life generalized anxiety disorder. Am. J. Geriatr. Psychiatry 15, 673–679. Mather, M., Carstensen, L.L., 2003. Aging and attentional biases for emotional faces. Psychol. Sci. 14, 409–415. Mathers, C., Vos, T., Stevenson, C., 1999. The Burden of Disease and Injury in Australia. AIHW, Canberra. McLaughlin, K.A., Behar, E., Borkovec, T.D., 2008. Family history of psychological problems in generalized anxiety disorder. J. Clin. Psychol. 64, 905–918. O'Connor, D.W., Parslow, R.A., 2010. Mental health scales and psychiatric diagnoses: responses to GHQ-12, K-10 and CIDI across the lifespan. J. Affect. Disord. 121, 263–267. Palmer, B.W., Jeste, D.V., Sheikh, J.I., 1997. Anxiety disorders in the elderly: DSM-IV and other barriers to diagnosis and treatment. J. Affect. Disord. 46, 183–190. Pinquart, M., Duberstein, P.R., 2007. Treatment of anxiety disorders in older adults: a meta-analytic comparison behavioral and pharmacological interventions. Am. J. Geriatr. Psychiatry 15, 639–651. Porensky, E.K., Dew, M.A., Karp, J.F., Skidmore, E.R., Rollman, B.L., Shear, M.K., Lenze, E.J., 2009. The burden of late-life generalized anxiety disorder: effects on disability, health-related quality of life, and healthcare utilization. Am. J. Geriatr. Psychiatry 17, 473–482. Ramsawh, H.J., Raffa, S.D., Edelen, M.O., Rende, R., Keller, M.B., 2009. Anxiety in middle adulthood: effects of age and time on the 14-year course of panic disorder, social phobia and generalized anxiety disorder. Psychol. Med. 39, 615–624. Regier, D.A., Boyd, J.H., Burke, J.D., Rae, D.S., Myers, J.K., Kramer, M., Robins, L.N., George, L.K., Karno, M., Locke, B.Z., 1988. One month prevalence of mental disorders in the United States: based on 5 epidemiologic catchment area sites. Arch. Gen. Psychiatry 45, 977–986. Rubio, G., Lopez-Ibor, J.J., 2007. Generalized anxiety disorder: a 40-year follow-up study. Acta Psychiatr. Scand. 115, 372–379. Ruscio, A.M., Chiu, W.T., Roy-Byrne, P., Stang, P.E., Stein, D.J., Wittchen, H.U., Kessler, R.C., 2007. Broadening the definition of generalized anxiety disorder: effects on prevalence and associations with other disorders in the National Comorbidity Survey Replication. J. Anxiety Disord. 21, 662–676. Schoevers, R.A., Deeg, D.J.H., van Tilburg, W., Beekman, A.T.F., 2005. Depression and generalized anxiety disorder — co-occurrence and longitudinal patterns in elderly patients. Am. J. Geriatr. Psychiatry 13, 31–39. Simon, N.M., 2009. Generalized anxiety disorder and psychiatric comorbidities such as depression, bipolar disorder, and substance abuse. J. Clin. Psychiatry 70, 10–14. Slade, T., Johnston, A., Browne, M.A.O., Andrews, G., Whiteford, H., 2009. 2007 National Survey of Mental Health and Wellbeing: methods and key findings. Aust. N. Z. J. Psychiatry 43, 594–605. Smit, F., Comijs, H., Schoevers, R., Cuijpers, P., Deeg, D., Beekman, A., 2007. Target groups for the prevention of late-life anxiety. Br. J. Psychiatry 190, 428–434. 230 D.C. Gonçalves et al. / Journal of Affective Disorders 132 (2011) 223–230 StataCorp, 2009. Statistical Software: Release 11.0. Stata Corporation, College Station, TX. Steffens, D.C., McQuoid, D.R., 2005. Impact of symptoms of generalized anxiety disorder on the course of late-life depression. Am. J. Geriatr. Psychiatry 13, 40–47. Vink, D., Aartsen, M.J., Schoevers, R.A., 2008. Risk factors for anxiety and depression in the elderly: a review. J. Affect. Disord. 106, 29–44. Wetherell, J.L., Maser, J.D., van Balkom, A., 2005. Anxiety disorders in the elderly: outdated beliefs and a research agenda. Acta Psychiatr. Scand. 111, 401–402. Wolitzky-Taylor, K.B., Castriotta, N., Lenze, E.J., Stanley, M.A., Craske, M.G., 2010. Anxiety disorders in older adults: a comprehensive review. Depress. Anxiety 27, 190–211. Zimmerman, M., Chelminski, I., 2003. Generalized anxiety disorder in patients with major depression: is DSM-IV's hierarchy correct? Am. J. Psychiatry 160, 504–512.