Geriatric Nursing xx (2014) 1e4
Contents lists available at ScienceDirect
Geriatric Nursing journal homepage: www.gnjournal.com
Feature Article
Testing the psychometric properties of the Cognitions Checklist, a measure to differentiate anxiety and depression among older adults Catherine R. Ayers, PhD, ABPP a, b, c, *, John H. Riskind, PhD d a
Research Service, VA San Diego Healthcare System, USA Psychology Service, VA San Diego Healthcare System, USA c Department of Psychiatry, University of California, San Diego School of Medicine, USA d George Mason University, USA b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 9 January 2014 Received in revised form 10 May 2014 Accepted 12 May 2014 Available online xxx
Considerable debate has been waged in the field about whether anxiety and depressive cognitions can be discriminated, and whether they can discriminate anxiety and depression symptoms. The current study examined a standard measure of cognitions, the Cognitions Checklist (CCL) that has yielded mixed results when tested in older age samples. A community sample of older adults (N ¼ 169; mean age ¼ 75.70; SD ¼ 8.55) completed a series of self-report questionnaires, including the CCL as well as measures of anxiety and depression symptoms. The CCL, which yielded a three-factor structure rather than the typical two-factor structure, did not cognitively discriminate anxiety from depression. The results have implications for understanding cognitive factors that differentiate between anxiety and depression symptoms in older adults and suggest the importance of assessing cognitions that are tailored to the concerns of this population. Ó 2014 Mosby, Inc. All rights reserved.
Keywords: Geriatric Cognitions Checklist Content specificity hypothesis
Introduction Because symptoms of anxiety and depression may often manifest in similar ways, it is imperative to use assessments that can reliably and validly differentiate between the two types of symptom presentation. This has historically been done through the evaluation of emotions (i.e., the two-factor model of affect)1 or cognitions (i.e., the cognitive content-specificity hypothesis).2 Advancing our knowledge of the distinct cognitive mechanisms related to anxiety and depression has important implications for prevention and alleviation of these symptoms in late life. One challenge to this, however, is that considerable debate has been waged in the field about whether anxiety and depressive cognitions can be discriminated, and whether they can discriminate the symptoms of anxiety and depressive disorders,3 particularly in
Role of funding source: The research was supported by a Career Development Award (CSRD-068-10S) from the Clinical Science R & D Program of the Veterans Health Administration. The contents do not reflect the views of the Department of Veterans Affairs or the United States Government. Declaration of interest: There are no conflicts to report. * Corresponding author. 3350 La Jolla Village Drive 116B, San Diego, CA 92161, USA. Tel.: þ1 858 5528585x2976. E-mail address: [email protected] (C.R. Ayers). 0197-4572/$ e see front matter Ó 2014 Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.gerinurse.2014.05.002
older adults. Attempts to disentangle anxiety cognitions from depression cognitions have had mixed success, in part reflecting the general difficulty in disentangling anxiety and depression symptoms in both older4,5 and younger populations.3 An added challenge is the growing evidence that the anxiety and depression symptoms witnessed in younger adults may be different that the symptoms which present in older adults.6e8 Cognitions are of particular interest when studying the differentiation of anxiety and depression in older adults because cognitions are less susceptible to age related factors than are somatic or behavioral mechanisms. According to Beck’s cognitive content specificity hypothesis, anxiety has its own unique disorder-specific content that differentiates it from depression.2,9 Basic to this model is the assumption that anxiety is concerned with the harm appraisal of potential future threat, whereas depression is concerned with past loss, defeat, and failure.2,9,10 Cognitions related to perception of danger are typically associated with anxiety symptoms, whereas cognitions of loss or failure are more characteristic of depression. While these automatic thoughts are not specific symptoms of anxiety or depressive disorders, having automatic thoughts related to danger or failure is associated with the diagnosis of anxiety or depressive disorders, respectively. Beck and colleagues11 developed the Cognition Checklist (CCL) to assess the frequency of automatic thoughts specific to anxiety and depression in order to facilitate the
2
C.R. Ayers, J.H. Riskind / Geriatric Nursing xx (2014) 1e4
Measures
differentiation of anxiety and depressive symptoms and thus the diagnosis of anxiety and depressive disorders. Across age groups, the ability to discriminate anxiety from depression with the CCL or similar instruments has received mixed support.3,12e14 For example, Beck and Perkins3 performed a metaanalysis and found evidence of cognitive content specificity for depression but not for anxiety. In keeping with cognitive content specificity, their meta-analysis confirmed that depression has distinct cognitive content (e.g., hopelessness) that is not as strongly related to anxiety. Their results yielded little evidence for discriminating cognitive features or specific cognitive contents of anxiety (e.g., threat cognitions or worry) that are not equally correlated with depression. Studies of the CCL in older adult samples have likewise failed to substantiate a clear distinction between anxious and depressive cognitions.13,14 In a clinical sample of older adults diagnosed with generalized anxiety disorder (GAD), Beck and colleagues13 found the expected two-factor structure of the CCL (anxiety & depression cognitions) but they did not show specificity to symptoms. Results indicated that the CCL-Depression factor correlated with depression and the CCL-Anxiety factor did not uniquely correlate with anxiety. Shapiro et al14 found three factors (Anxious, Social Loss, & Worthlessness) in the CCL in their work with an older community sample. Taken together, it is clear that gaps remain in our understanding of the disorder-specific cognitions in older adults. The primary purpose of this investigation was to explore the reliability and validity of a previously established measure by further evaluating the factor structure and discriminant validity of the standard CCL11 in an older adult population. The importance of examining the psychometric properties is twofold. First, this would help to cognitively differentiate anxiety and depression in older adults. Further, this would shed light on the utility of this measure in an older adult psychiatric population. Eventually, these results could lead to further development of accurate measures of anxiety and depressive cognitions in older adults.
Basic demographic factors including age, gender, and ethnicity were included. Marital status, educational level, and total number of self-reported physician diagnosed medical illnesses were also assessed. There was no formal measurement of health status. The Adult Manifest Anxiety Scale e Elderly Version (AMAS-E)15 is a 44-item self-report inventory that measures the level and nature of anxiety in older adults based on the Diagnostic and Statistical Manual e IV-TR (DSM-IV-TR).16 The AMAS-E includes three subscales focused on Worry/Oversensitivity, Physiological Anxiety, and Fear of Aging. All items are answered dichotomously (yes or no), with affirmative answers corresponding to endorsement of anxiety thoughts, feelings, or actions. The current study looked only at the Total Anxiety score of the AMAS-E, in which higher scores suggest higher levels of anxiety. Internal consistency of the AMAS-E Total Anxiety score is good (a ¼ 0.90), as is the test-retest reliability (r ¼ 0.83).15 The Clinical Assessment Scales for the Elderly (CASE-SF)17 is a self-report measure designed to screen for Axis I disorders in older adults based on the DSMe IV-TR.16 The clinical scales include: anxiety, cognitive competence, depression, fear of aging, mania, obsessive-compulsive, paranoia, psychoticism, somatization, and substance abuse. The Cognition Checklist (CCL)11 is a 26-item self-report measure of cognitions typically present in individuals with depressive or anxiety symptoms. The CCL consists of two subscales: Depressive Cognition subscale (CCL-D) and the Anxious Cognition subscale (CCL-A). Both subscales have demonstrated high internal consistency (CCL-D: a ¼ 0.90; CCL-A: a ¼ 0.92), test-retest reliability over six-weeks (CCL-D: r ¼ 0.76, p < 0.001; CCL-A: r ¼ 0.79, p < 0.001).7 Scores on the CCL have been found to correlate with depressive and anxiety symptoms in both clinical and non-clinical populations in young and middle-aged adults.
Methodology
Data analysis
Participants
Univariate analyses of variance were conducted on the entire sample to determine the effects of demographic factors (age, ethnicity, sample origin, number of health conditions, and education) on measures of cognitions, anxiety, and depression. Next, exploratory factor analysis was conducted on the CCL. Following the analyses of Shapiro and colleagues,14 principalaxis factoring with varimax (orthogonal) rotation was utilized to explore the items of the CCL and a scree plot was used to examine the potential factor solution. To determine the number of stable factors present in each measure, factors with eigenvalues over one were examined. Factor Loadings of 0.30 or greater were considered to be stable factors. Internal reliability, concurrent, and discriminant validity were explored. Finally, the relationships between the three observed subscales of the CCL and anxiety and depression were examined with Pearson and partial correlations.
A total of 169 older adult participants were given a series of selfreport questionnaires, including measures of anxiety and depression symptoms. Recruitment took place at a local university educational course for retired older adults as well as in a continuing care retirement facility that contained over 300 residents via posted flyers. A total of 110 from the educational course and all residents from the continuing care were invited to participate. To be included in the study, participants had to be over the age of 60. All participants consented to the Institutional Review Board approved project and no participants were found ineligible. Participants were 113 female (66.9%) and 56 male (33.1%) older adults with an average age of 75.70 (SD ¼ 8.55). This sample consisted mainly of Caucasian participants (98.3% Caucasian, 1.1% African American, and 0.6% Other). Ninety-four participants were recruited from a continuing care retirement facility and 75 participants were recruited from community dwelling seniors. Although continuing care residents reported significantly higher numbers of health conditions (M ¼ 4.74, SD ¼ 2.65) than those from the community ((M ¼ 3.11, SD ¼ 2.59), F (1,175) ¼ 19.37, p < 0.001), there were no other significant differences found between these two samples. Over one half of the participants were married (56.5%) while 32.8% were widowed, 6.2% were divorced, 3.4% were never married, and 1.1% were separated. Educational levels were high (44.1% received a graduate school degree, 20.3% received a college degree, 15.3% received some college, 8.5% received a high school degree, 1.2% received some high school, and 8.5% gave no response).
Results Age and gender differences on the CCL, AMAS-E, and CASE were explored. Although no gender differences were found on the CCL, females were significantly higher than males on both the AMAS-E total anxiety scale (females: M ¼ 46.52, SD ¼ 9.02; males: M ¼ 43.04, SD ¼ 7.19), F (1,167) ¼ 6.26, p < 0.05, and depression on the CASE depression subscale (females: M ¼ 44.82, SD ¼ 5.51; males: M ¼ 42.37, SD ¼ 3.80; F (1,157) ¼ 8.53, p < 0.05. Age significantly predicted number of health conditions (R2 ¼ 0.06, F (1,168) ¼ 10.74, p < 0.01) but was not related to other variables.
C.R. Ayers, J.H. Riskind / Geriatric Nursing xx (2014) 1e4
Principal-axis factoring yielded seven factors with eigenvalues greater than 1. The scree plot suggested a 3-factor solution, as this showed a fairly clear distinction between the first three factors and factors 4e7 appeared to be uninterpretable. Factor analysis was then rerun specifying the extraction of exactly 3 factors, which accounted for 46.1% of the variance. Factor 1 accounted for 39.72% of the total variance and was comprised of primarily the original CCL anxiety items. Factors 2 and 3 were comprised of the original CCL depressive cognitions and accounted for 10.3% and 6.09% of the variance of the original CCL depressive cognitions. Internal reliability of the CCL was high (a ¼ 0.90), as were the factors (anxiety ¼ 0.85, social loss ¼ 0.80, and worthlessness ¼ 0.80). Reliability did not improve with the deletion of any items. Factor 1 included items of general anxiety and future oriented threats of harm to self or others but also included three that had depressive themes (e.g., I’m not worthy of people’s attention or affection) (Table 1). Factor 2 contained cognitions concerning current feelings of social inferiority, loss of social status, and negative social comparison (e.g., I’ll never be as good as other people). Factor 3 contained cognitions concerning current feelings of worthlessness and hopelessness (e.g., No one cares if I live or die). To examine the concurrent and discriminant validity of the set of cognitive measures, Pearson correlations were conducted between the cognition factors, anxiety, and depression (Table 2). Concurrent validity was established between constructs related to anxiety (AMAS-E, CCL-Anxiety) and depression (CASE-DEP, CCLSocial Loss, CCL-Worthlessness). Due to the significant positive correlation between symptoms of anxiety and depression (r ¼ 0.50,
3
Table 2 Correlations among anxiety, depression, and cognitions in non-psychiatric older adults. Measure
a
1
2
3
4
5
M
SD
1. 2. 3. 4. 5.
0.88 0.86 0.88 0.80 0.80
e
0.55* e
0.32* 0.10 e
0.11 0.11 0.08 e
0.09 0.13 0.05 0.08 e
45.51 43.98 9.69 2.87 1.40
8.66 5.11 6.98 3.05 4.86
AMAS-E CASE-DEP CCL-A CCL-S CCL-W
AMAS-E ¼ Adult Manifest Anxiety Scale e Elderly Version, CASE-DEP ¼ Clinical Assessment Scales for the Elderly e Depression Subscale, CCL-A ¼ Cognition Checklist e Anxiety Factor, CCL-S ¼ Cognition Checklist e Social Loss Factor, CCL-W ¼ Cognition Checklist e Worthlessness Factor. * p < 0.01. AMAS-E and CASE-DEP scores used were partial correlation residuals.
p < 0.001), anxiety was covaried out from depression and depression was covaried out from anxiety, which created a “pure anxiety” and “pure depression” residual using partial correlations on the AMAS-E and CASE variables. These residuals for the AMAS-E and CASE were used in the remainder of the correlational analyses. Anxiety symptoms (using the AMAS-E partial correlation residual) were significantly related to the CCL-Anxiety. Depression symptoms (using the CASE-DEP partial correlation residual) were correlated with the CCL-Worthlessness factor. No significant correlations were found between the CCL-Anxiety, CCL-Social Loss, and CCL-Worthlessness factors. Discussion
Table 1 Varimax-rotated Factor Loadings for Cognition Checklist (CCL) items. Cognition Checklist items
Anxiety
I’m going to have an accident. (A) I am going to be injured. (A) What if no one reaches me in time to help? (A) I will never overcome my problems. (D) Something awful is going to happen. (A) I am not a healthy person. (A) I might be trapped. (A) There’s something very wrong with me. (A) I’m going to have a heart attack. (A) Something might be happening that will ruin my appearance. (A) Nothing ever works out for me anymore. (D) What if I get sick and become an invalid? (A) Something will happen to someone I care about. (A) I have become physically unattractive. (D) I’ll never be as good as other people are. (D) People don’t respect me anymore. (D) I’m not worthy of people’s attention or affection. (D) I’m a social failure. (D) No one cares whether I live or die. (D) I don’t deserve to be loved. (D) I’ve lost the only friends I’ve had. (D) There’s no one left to help me. (D) Life isn’t worth living. (D) I’m worthless. (D) I’m worse off than they are. (D)
0.67 0.66 0.66
Social loss
Worthlessness
Factor 1 Factor 2 Factor 3
0.63 0.63 0.62 0.61 0.55 0.53 0.50 0.49
0.45
0.46 0.44 0.41 0.72
0.32
0.61 0.60 0.58 0.54 0.47 0.31
0.30
0.49 0.79 0.57 0.56 0.55 0.34
Factor Loadings of less than 0.30 were omitted. (A) ¼ CCL e Anxiety subscale, (D) ¼ CCL e Depression subscale (Beck et al, 1987).
The factor structure derived from the current study resembled the three factor solution of Shapiro et al.14 These three CCL factors (anxiety over health issues, loss of social role, and self-evaluation worthlessness) were similar to those found by Shapiro and colleagues.14 However, factor one, CCL-Anxiety, included the additional cognitions of “I will never overcome my problems” (from Factor 3 in Shapiro et al14), “I have become physically unattractive” (from Factor 3 in Shapiro et al14), “I’m not worthy of people’s attention or affection” (from Factor 2 in Shapiro et al14), “I’m worse off than they are” (from Factor 2 in Shapiro et al14). One additional item appeared in Factor 2, “I’ll never be as good as other people,” which was not included in the three factors of Shapiro et al14 because it loaded less than 0.30. Similarly, two cognitions appeared in Factor 2, CCL-Social Loss, in the present study that did not load on the three factors in Shapiro’s study, “I’m worse off than they are” and “I’ve lost the only friends I’ve had.” It is also noted that four out of the eleven items that were found in the CCLWorthlessness factor of Shapiro et al14 were found in the present study. Despite these differences, the overarching themes presented in both studies’ factor results were similar. There has been considerable debate about whether anxiety and depressive cognitions can be discriminated, and whether they can discriminate anxiety and depression symptoms,3 particularly in older populations. Our goal in the present study was to assess a standard measure of cognitions, the CCL,11 which has yielded mixed results when tested in older adult samples. Our results supported the notion that anxiety and depressive symptoms may not be as clearly delineated in late life samples as they can be in younger populations. Our objective was to further evaluate the cognitive specificity of the standard CCL in older adults. The present findings largely supported a three-factor structure for the CCL found by Shapiro14 with a community sample of older adults, rather than a two-factor structure (anxious & depressive cognitions) found by Beck et al13 in older adults with generalized anxiety disorder. Beck and colleagues13 did
4
C.R. Ayers, J.H. Riskind / Geriatric Nursing xx (2014) 1e4
note that CCL-Anxiety scale did not uniquely correlate with anxiety measures. Much like Shapiro, our results indicated that the “anxiety” factor derived from the original CCL contained both anxiety and depression cognitions, whereas the “depressive” factor derived from the original CCL divided into two discrete factors (social loss and worthlessness). Thus, anxiety and depressive cognitions may not be able to easily differentiate in later life. Clinicians should be aware that the relationships between cognitions and anxiety or depressive symptoms may not be as clear in older adults as they are in middle-aged individuals, which may decrease the ability of clinicians to make appropriate diagnoses. According to the most recent edition of the Diagnostic and Statistical Manual (DSM-5),18 the cognitions associated with generalized anxiety disorder (GAD) and major depressive disorder (MDD) should present differently. Cognitions associated with GAD are presented more in the context of worry or “apprehensive expectation. about a number of events or activities,” whereas the cognitions associated with MDD are presented as “feelings of worthlessness or excessive or inappropriate guilt.”18 Clinicians basing their diagnoses of anxiety or depression disorders on the DSM-5 criteria may be more likely to assume that the cognitions will be equally differentiating in older adults as in middle-aged adults. The current study provided an example of the limitations of the DSM-5 to meet the needs of atypical psychiatric populations, such as older adults. Clinicians may benefit from using the criteria listed in the DSM-5 as more of a guideline and not necessarily as the definitive word on empirical criteria for all populations. This study holds several important clinical implications as these findings suggested that the CCL may not clearly delineate anxiety and depression in older adults. While the CCL is not susceptible to misinterpreted medical symptoms or behaviors, which are likely more age related (e.g., decreased activity due to change in social status) than cognitions, the CCL may not accurately capture distress cognitions in later life. There are benefits to utilizing measures across the lifespan for comparative purposes. However, important nuances of cognition may be missed in older adults. Given life events (retirement, widowhood, etc.) and developmental processes (social changes, physical changes, medical conditions), it is not surprising that automatic thoughts change across the lifespan. It is possible that it is the cognitions examined e or even just the phrasing of the cognitions e in the CCL that affected the relationship between cognitions and anxiety and depression symptoms. A cognitions measure that is specific to the automatic thoughts of older adults may elicit different results. Clinicians should be wary of using the CCL because of the bleed over of cognitions between late life anxiety and depression, which inhibits the utility of current diagnostic standards. Several limitations of the present study should be recognized. First, the study relied on a non-psychiatric sample of older adults and the generalizability of the findings to older adults with anxiety or mood disorders is unclear. Second, the study used a relatively homogenous sample that was largely female, Caucasian, and relatively well educated. We were also not able to obtain test-retest results of our measures to establish reliability. Nonetheless, the findings have both theoretical and practical relevance given that adults who do not meet criteria for anxiety disorders often show disability that is similar to those who pass symptom thresholds for such disorders.19e21 Another limitation of the study is that it used a cross-sectional design, and a prospective longitudinal design would
be necessary for any causal inferences about the influence any of the cognitive factors on the development of anxiety or depressive symptoms. Future work should examine other measures of anxiety and depression cognitions in older adult psychiatric sample as it has direct implications on cognitive-behavioral treatment. Finally, the study was limited by the use of exploratory factor analysis instead of confirmatory factor analysis. Exploratory factor analysis is traditionally used to reveal possible subscales or to reduce extraneous items in a measure, whereas confirmatory factor analysis is used to determine if the observed factor structure fits across samples. Future studies may want to examine the CCL in older adults using a confirmatory factor analysis to further evaluate the model fit of the subscales of the CCL. References 1. Watson D, Clark LA, Carey G. Positive and negative affectivity and their relation to anxiety and depressive disorders. J Abnorm Psychol. 1988;97:346e353. 2. Beck AT. Cognitive Theory and the Emotional Disorders. New York: International University Press; 1976. 3. Beck R, Perkins TS. Cognitive content-specificity for anxiety and depression: a meta-analysis. Cogn Ther Res. 2001;25:651e663. 4. Kvaal K, McDougall FA, Brayne C, et al. Co-occurrence of anxiety and depressive disorders in a community sample of older people: results from the MRC CFAS (Medical Research Council Cognitive Function and Ageing Study). Int J Geriatr Psychiatry. 2008;23:229e237. 5. Schoevers RA, Beekman AT, Deeg DJ, et al. Comorbidity and risk-patterns of depression, generalized anxiety disorder and mixed anxiety-depression in later life: results from the AMSTEL study. Int J Geriatr Psychiatry. 2003;18:994e1001. 6. Mather M. The emotion paradox in the aging brain. Ann N Y Acad Sci. 2012;1251:33e49. 7. Miloyan B, Byrne GJ, Pachana NA. Age-related changed in generalized anxiety disorder symptoms. Int Psychogeriatr. 2014;26:565e572. 8. Sütterlin S, Paap MC, Babic S, et al. Rumination and age: some things get better. J Aging Res. 2012;2012:267327. http://dx.doi.org/10.1155/2012/267327. 9. Clark DA, Beck AT, Stewart B. Cognitive specificity and positiveenegative affectivity: complementary or contradictory views on anxiety and depression? J Abnorm Psychol. 1990;99:148e155. 10. Riskind JH, Williams NL, Gessner TL, et al. The looming maladaptive style: anxiety, danger, and schematic processing. J Pers Soc Psychol. 2000;79: 837e852. 11. Beck AT, Brown G, Steer RA, et al. Differentiating anxiety and depression: a test of the cognitive content-specificity hypothesis. J Abnorm Psychol. 1987;96(3): 179e183. 12. Barlow DH. Unraveling the mysteries of anxiety and its disorders from the perspective of emotion theory. Am Psychol. 2000;55:1247e1263. 13. Beck JG, Novy DM, Diefenback GJ, et al. Differentiating anxiety and depression in older adults with generalized anxiety disorder. Psychol Assess. 2003;15(2): 184e192. 14. Shapiro AM, Roberts JE, Beck JG. Differentiating symptoms of anxiety and depression in older adults: distinct cognitive and affective profiles. Cogn Ther Res. 1999;23:53e74. 15. Lowe PA, Reynolds CR. Examination of the psychometric properties of the adult manifest anxiety scale-elderly version scores. Educ Psychol Meas. 2006;66(1): 93e115. 16. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. text rev. 4th ed. Washington, DC: American Psychiatric Publishing; 2000. 17. Reynolds CR, Bigler ED. Clinical Assessment Scale for the Elderly; Professional Manual for the CASE and CASE-SF. Lutz, FL: Psychological Assessment Resources, Inc; 2001. 18. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Health Disorders: DSM-5. 5th ed. Washington, DC: American Psychiatric Publishing; 2013. 19. de Beurs E, Beekman AT, van Balkom AJ, Deeg DJ, van Dyck R, van Tilburg W. Consequences of anxiety in older persons: its effect on disability, well-being and use of health services. Psychol Med. 1999;29:583e593. 20. Brenes GA, Guralnik JM, Williamson JD, et al. The influence of anxiety on the progression of disability. J Am Geriatr Soc. 2005;53:34e39. 21. Brenes GA, Guralnik JM, Williamson J, Fried LP, Penninx BW. Correlates of anxiety symptoms in physically disabled older women. Am J Geriatr Psychiatry. 2005;13:15e22.
Contents lists available at ScienceDirect
Geriatric Nursing journal homepage: www.gnjournal.com
Feature Article
Testing the psychometric properties of the Cognitions Checklist, a measure to differentiate anxiety and depression among older adults Catherine R. Ayers, PhD, ABPP a, b, c, *, John H. Riskind, PhD d a
Research Service, VA San Diego Healthcare System, USA Psychology Service, VA San Diego Healthcare System, USA c Department of Psychiatry, University of California, San Diego School of Medicine, USA d George Mason University, USA b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 9 January 2014 Received in revised form 10 May 2014 Accepted 12 May 2014 Available online xxx
Considerable debate has been waged in the field about whether anxiety and depressive cognitions can be discriminated, and whether they can discriminate anxiety and depression symptoms. The current study examined a standard measure of cognitions, the Cognitions Checklist (CCL) that has yielded mixed results when tested in older age samples. A community sample of older adults (N ¼ 169; mean age ¼ 75.70; SD ¼ 8.55) completed a series of self-report questionnaires, including the CCL as well as measures of anxiety and depression symptoms. The CCL, which yielded a three-factor structure rather than the typical two-factor structure, did not cognitively discriminate anxiety from depression. The results have implications for understanding cognitive factors that differentiate between anxiety and depression symptoms in older adults and suggest the importance of assessing cognitions that are tailored to the concerns of this population. Ó 2014 Mosby, Inc. All rights reserved.
Keywords: Geriatric Cognitions Checklist Content specificity hypothesis
Introduction Because symptoms of anxiety and depression may often manifest in similar ways, it is imperative to use assessments that can reliably and validly differentiate between the two types of symptom presentation. This has historically been done through the evaluation of emotions (i.e., the two-factor model of affect)1 or cognitions (i.e., the cognitive content-specificity hypothesis).2 Advancing our knowledge of the distinct cognitive mechanisms related to anxiety and depression has important implications for prevention and alleviation of these symptoms in late life. One challenge to this, however, is that considerable debate has been waged in the field about whether anxiety and depressive cognitions can be discriminated, and whether they can discriminate the symptoms of anxiety and depressive disorders,3 particularly in
Role of funding source: The research was supported by a Career Development Award (CSRD-068-10S) from the Clinical Science R & D Program of the Veterans Health Administration. The contents do not reflect the views of the Department of Veterans Affairs or the United States Government. Declaration of interest: There are no conflicts to report. * Corresponding author. 3350 La Jolla Village Drive 116B, San Diego, CA 92161, USA. Tel.: þ1 858 5528585x2976. E-mail address: [email protected] (C.R. Ayers). 0197-4572/$ e see front matter Ó 2014 Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.gerinurse.2014.05.002
older adults. Attempts to disentangle anxiety cognitions from depression cognitions have had mixed success, in part reflecting the general difficulty in disentangling anxiety and depression symptoms in both older4,5 and younger populations.3 An added challenge is the growing evidence that the anxiety and depression symptoms witnessed in younger adults may be different that the symptoms which present in older adults.6e8 Cognitions are of particular interest when studying the differentiation of anxiety and depression in older adults because cognitions are less susceptible to age related factors than are somatic or behavioral mechanisms. According to Beck’s cognitive content specificity hypothesis, anxiety has its own unique disorder-specific content that differentiates it from depression.2,9 Basic to this model is the assumption that anxiety is concerned with the harm appraisal of potential future threat, whereas depression is concerned with past loss, defeat, and failure.2,9,10 Cognitions related to perception of danger are typically associated with anxiety symptoms, whereas cognitions of loss or failure are more characteristic of depression. While these automatic thoughts are not specific symptoms of anxiety or depressive disorders, having automatic thoughts related to danger or failure is associated with the diagnosis of anxiety or depressive disorders, respectively. Beck and colleagues11 developed the Cognition Checklist (CCL) to assess the frequency of automatic thoughts specific to anxiety and depression in order to facilitate the
2
C.R. Ayers, J.H. Riskind / Geriatric Nursing xx (2014) 1e4
Measures
differentiation of anxiety and depressive symptoms and thus the diagnosis of anxiety and depressive disorders. Across age groups, the ability to discriminate anxiety from depression with the CCL or similar instruments has received mixed support.3,12e14 For example, Beck and Perkins3 performed a metaanalysis and found evidence of cognitive content specificity for depression but not for anxiety. In keeping with cognitive content specificity, their meta-analysis confirmed that depression has distinct cognitive content (e.g., hopelessness) that is not as strongly related to anxiety. Their results yielded little evidence for discriminating cognitive features or specific cognitive contents of anxiety (e.g., threat cognitions or worry) that are not equally correlated with depression. Studies of the CCL in older adult samples have likewise failed to substantiate a clear distinction between anxious and depressive cognitions.13,14 In a clinical sample of older adults diagnosed with generalized anxiety disorder (GAD), Beck and colleagues13 found the expected two-factor structure of the CCL (anxiety & depression cognitions) but they did not show specificity to symptoms. Results indicated that the CCL-Depression factor correlated with depression and the CCL-Anxiety factor did not uniquely correlate with anxiety. Shapiro et al14 found three factors (Anxious, Social Loss, & Worthlessness) in the CCL in their work with an older community sample. Taken together, it is clear that gaps remain in our understanding of the disorder-specific cognitions in older adults. The primary purpose of this investigation was to explore the reliability and validity of a previously established measure by further evaluating the factor structure and discriminant validity of the standard CCL11 in an older adult population. The importance of examining the psychometric properties is twofold. First, this would help to cognitively differentiate anxiety and depression in older adults. Further, this would shed light on the utility of this measure in an older adult psychiatric population. Eventually, these results could lead to further development of accurate measures of anxiety and depressive cognitions in older adults.
Basic demographic factors including age, gender, and ethnicity were included. Marital status, educational level, and total number of self-reported physician diagnosed medical illnesses were also assessed. There was no formal measurement of health status. The Adult Manifest Anxiety Scale e Elderly Version (AMAS-E)15 is a 44-item self-report inventory that measures the level and nature of anxiety in older adults based on the Diagnostic and Statistical Manual e IV-TR (DSM-IV-TR).16 The AMAS-E includes three subscales focused on Worry/Oversensitivity, Physiological Anxiety, and Fear of Aging. All items are answered dichotomously (yes or no), with affirmative answers corresponding to endorsement of anxiety thoughts, feelings, or actions. The current study looked only at the Total Anxiety score of the AMAS-E, in which higher scores suggest higher levels of anxiety. Internal consistency of the AMAS-E Total Anxiety score is good (a ¼ 0.90), as is the test-retest reliability (r ¼ 0.83).15 The Clinical Assessment Scales for the Elderly (CASE-SF)17 is a self-report measure designed to screen for Axis I disorders in older adults based on the DSMe IV-TR.16 The clinical scales include: anxiety, cognitive competence, depression, fear of aging, mania, obsessive-compulsive, paranoia, psychoticism, somatization, and substance abuse. The Cognition Checklist (CCL)11 is a 26-item self-report measure of cognitions typically present in individuals with depressive or anxiety symptoms. The CCL consists of two subscales: Depressive Cognition subscale (CCL-D) and the Anxious Cognition subscale (CCL-A). Both subscales have demonstrated high internal consistency (CCL-D: a ¼ 0.90; CCL-A: a ¼ 0.92), test-retest reliability over six-weeks (CCL-D: r ¼ 0.76, p < 0.001; CCL-A: r ¼ 0.79, p < 0.001).7 Scores on the CCL have been found to correlate with depressive and anxiety symptoms in both clinical and non-clinical populations in young and middle-aged adults.
Methodology
Data analysis
Participants
Univariate analyses of variance were conducted on the entire sample to determine the effects of demographic factors (age, ethnicity, sample origin, number of health conditions, and education) on measures of cognitions, anxiety, and depression. Next, exploratory factor analysis was conducted on the CCL. Following the analyses of Shapiro and colleagues,14 principalaxis factoring with varimax (orthogonal) rotation was utilized to explore the items of the CCL and a scree plot was used to examine the potential factor solution. To determine the number of stable factors present in each measure, factors with eigenvalues over one were examined. Factor Loadings of 0.30 or greater were considered to be stable factors. Internal reliability, concurrent, and discriminant validity were explored. Finally, the relationships between the three observed subscales of the CCL and anxiety and depression were examined with Pearson and partial correlations.
A total of 169 older adult participants were given a series of selfreport questionnaires, including measures of anxiety and depression symptoms. Recruitment took place at a local university educational course for retired older adults as well as in a continuing care retirement facility that contained over 300 residents via posted flyers. A total of 110 from the educational course and all residents from the continuing care were invited to participate. To be included in the study, participants had to be over the age of 60. All participants consented to the Institutional Review Board approved project and no participants were found ineligible. Participants were 113 female (66.9%) and 56 male (33.1%) older adults with an average age of 75.70 (SD ¼ 8.55). This sample consisted mainly of Caucasian participants (98.3% Caucasian, 1.1% African American, and 0.6% Other). Ninety-four participants were recruited from a continuing care retirement facility and 75 participants were recruited from community dwelling seniors. Although continuing care residents reported significantly higher numbers of health conditions (M ¼ 4.74, SD ¼ 2.65) than those from the community ((M ¼ 3.11, SD ¼ 2.59), F (1,175) ¼ 19.37, p < 0.001), there were no other significant differences found between these two samples. Over one half of the participants were married (56.5%) while 32.8% were widowed, 6.2% were divorced, 3.4% were never married, and 1.1% were separated. Educational levels were high (44.1% received a graduate school degree, 20.3% received a college degree, 15.3% received some college, 8.5% received a high school degree, 1.2% received some high school, and 8.5% gave no response).
Results Age and gender differences on the CCL, AMAS-E, and CASE were explored. Although no gender differences were found on the CCL, females were significantly higher than males on both the AMAS-E total anxiety scale (females: M ¼ 46.52, SD ¼ 9.02; males: M ¼ 43.04, SD ¼ 7.19), F (1,167) ¼ 6.26, p < 0.05, and depression on the CASE depression subscale (females: M ¼ 44.82, SD ¼ 5.51; males: M ¼ 42.37, SD ¼ 3.80; F (1,157) ¼ 8.53, p < 0.05. Age significantly predicted number of health conditions (R2 ¼ 0.06, F (1,168) ¼ 10.74, p < 0.01) but was not related to other variables.
C.R. Ayers, J.H. Riskind / Geriatric Nursing xx (2014) 1e4
Principal-axis factoring yielded seven factors with eigenvalues greater than 1. The scree plot suggested a 3-factor solution, as this showed a fairly clear distinction between the first three factors and factors 4e7 appeared to be uninterpretable. Factor analysis was then rerun specifying the extraction of exactly 3 factors, which accounted for 46.1% of the variance. Factor 1 accounted for 39.72% of the total variance and was comprised of primarily the original CCL anxiety items. Factors 2 and 3 were comprised of the original CCL depressive cognitions and accounted for 10.3% and 6.09% of the variance of the original CCL depressive cognitions. Internal reliability of the CCL was high (a ¼ 0.90), as were the factors (anxiety ¼ 0.85, social loss ¼ 0.80, and worthlessness ¼ 0.80). Reliability did not improve with the deletion of any items. Factor 1 included items of general anxiety and future oriented threats of harm to self or others but also included three that had depressive themes (e.g., I’m not worthy of people’s attention or affection) (Table 1). Factor 2 contained cognitions concerning current feelings of social inferiority, loss of social status, and negative social comparison (e.g., I’ll never be as good as other people). Factor 3 contained cognitions concerning current feelings of worthlessness and hopelessness (e.g., No one cares if I live or die). To examine the concurrent and discriminant validity of the set of cognitive measures, Pearson correlations were conducted between the cognition factors, anxiety, and depression (Table 2). Concurrent validity was established between constructs related to anxiety (AMAS-E, CCL-Anxiety) and depression (CASE-DEP, CCLSocial Loss, CCL-Worthlessness). Due to the significant positive correlation between symptoms of anxiety and depression (r ¼ 0.50,
3
Table 2 Correlations among anxiety, depression, and cognitions in non-psychiatric older adults. Measure
a
1
2
3
4
5
M
SD
1. 2. 3. 4. 5.
0.88 0.86 0.88 0.80 0.80
e
0.55* e
0.32* 0.10 e
0.11 0.11 0.08 e
0.09 0.13 0.05 0.08 e
45.51 43.98 9.69 2.87 1.40
8.66 5.11 6.98 3.05 4.86
AMAS-E CASE-DEP CCL-A CCL-S CCL-W
AMAS-E ¼ Adult Manifest Anxiety Scale e Elderly Version, CASE-DEP ¼ Clinical Assessment Scales for the Elderly e Depression Subscale, CCL-A ¼ Cognition Checklist e Anxiety Factor, CCL-S ¼ Cognition Checklist e Social Loss Factor, CCL-W ¼ Cognition Checklist e Worthlessness Factor. * p < 0.01. AMAS-E and CASE-DEP scores used were partial correlation residuals.
p < 0.001), anxiety was covaried out from depression and depression was covaried out from anxiety, which created a “pure anxiety” and “pure depression” residual using partial correlations on the AMAS-E and CASE variables. These residuals for the AMAS-E and CASE were used in the remainder of the correlational analyses. Anxiety symptoms (using the AMAS-E partial correlation residual) were significantly related to the CCL-Anxiety. Depression symptoms (using the CASE-DEP partial correlation residual) were correlated with the CCL-Worthlessness factor. No significant correlations were found between the CCL-Anxiety, CCL-Social Loss, and CCL-Worthlessness factors. Discussion
Table 1 Varimax-rotated Factor Loadings for Cognition Checklist (CCL) items. Cognition Checklist items
Anxiety
I’m going to have an accident. (A) I am going to be injured. (A) What if no one reaches me in time to help? (A) I will never overcome my problems. (D) Something awful is going to happen. (A) I am not a healthy person. (A) I might be trapped. (A) There’s something very wrong with me. (A) I’m going to have a heart attack. (A) Something might be happening that will ruin my appearance. (A) Nothing ever works out for me anymore. (D) What if I get sick and become an invalid? (A) Something will happen to someone I care about. (A) I have become physically unattractive. (D) I’ll never be as good as other people are. (D) People don’t respect me anymore. (D) I’m not worthy of people’s attention or affection. (D) I’m a social failure. (D) No one cares whether I live or die. (D) I don’t deserve to be loved. (D) I’ve lost the only friends I’ve had. (D) There’s no one left to help me. (D) Life isn’t worth living. (D) I’m worthless. (D) I’m worse off than they are. (D)
0.67 0.66 0.66
Social loss
Worthlessness
Factor 1 Factor 2 Factor 3
0.63 0.63 0.62 0.61 0.55 0.53 0.50 0.49
0.45
0.46 0.44 0.41 0.72
0.32
0.61 0.60 0.58 0.54 0.47 0.31
0.30
0.49 0.79 0.57 0.56 0.55 0.34
Factor Loadings of less than 0.30 were omitted. (A) ¼ CCL e Anxiety subscale, (D) ¼ CCL e Depression subscale (Beck et al, 1987).
The factor structure derived from the current study resembled the three factor solution of Shapiro et al.14 These three CCL factors (anxiety over health issues, loss of social role, and self-evaluation worthlessness) were similar to those found by Shapiro and colleagues.14 However, factor one, CCL-Anxiety, included the additional cognitions of “I will never overcome my problems” (from Factor 3 in Shapiro et al14), “I have become physically unattractive” (from Factor 3 in Shapiro et al14), “I’m not worthy of people’s attention or affection” (from Factor 2 in Shapiro et al14), “I’m worse off than they are” (from Factor 2 in Shapiro et al14). One additional item appeared in Factor 2, “I’ll never be as good as other people,” which was not included in the three factors of Shapiro et al14 because it loaded less than 0.30. Similarly, two cognitions appeared in Factor 2, CCL-Social Loss, in the present study that did not load on the three factors in Shapiro’s study, “I’m worse off than they are” and “I’ve lost the only friends I’ve had.” It is also noted that four out of the eleven items that were found in the CCLWorthlessness factor of Shapiro et al14 were found in the present study. Despite these differences, the overarching themes presented in both studies’ factor results were similar. There has been considerable debate about whether anxiety and depressive cognitions can be discriminated, and whether they can discriminate anxiety and depression symptoms,3 particularly in older populations. Our goal in the present study was to assess a standard measure of cognitions, the CCL,11 which has yielded mixed results when tested in older adult samples. Our results supported the notion that anxiety and depressive symptoms may not be as clearly delineated in late life samples as they can be in younger populations. Our objective was to further evaluate the cognitive specificity of the standard CCL in older adults. The present findings largely supported a three-factor structure for the CCL found by Shapiro14 with a community sample of older adults, rather than a two-factor structure (anxious & depressive cognitions) found by Beck et al13 in older adults with generalized anxiety disorder. Beck and colleagues13 did
4
C.R. Ayers, J.H. Riskind / Geriatric Nursing xx (2014) 1e4
note that CCL-Anxiety scale did not uniquely correlate with anxiety measures. Much like Shapiro, our results indicated that the “anxiety” factor derived from the original CCL contained both anxiety and depression cognitions, whereas the “depressive” factor derived from the original CCL divided into two discrete factors (social loss and worthlessness). Thus, anxiety and depressive cognitions may not be able to easily differentiate in later life. Clinicians should be aware that the relationships between cognitions and anxiety or depressive symptoms may not be as clear in older adults as they are in middle-aged individuals, which may decrease the ability of clinicians to make appropriate diagnoses. According to the most recent edition of the Diagnostic and Statistical Manual (DSM-5),18 the cognitions associated with generalized anxiety disorder (GAD) and major depressive disorder (MDD) should present differently. Cognitions associated with GAD are presented more in the context of worry or “apprehensive expectation. about a number of events or activities,” whereas the cognitions associated with MDD are presented as “feelings of worthlessness or excessive or inappropriate guilt.”18 Clinicians basing their diagnoses of anxiety or depression disorders on the DSM-5 criteria may be more likely to assume that the cognitions will be equally differentiating in older adults as in middle-aged adults. The current study provided an example of the limitations of the DSM-5 to meet the needs of atypical psychiatric populations, such as older adults. Clinicians may benefit from using the criteria listed in the DSM-5 as more of a guideline and not necessarily as the definitive word on empirical criteria for all populations. This study holds several important clinical implications as these findings suggested that the CCL may not clearly delineate anxiety and depression in older adults. While the CCL is not susceptible to misinterpreted medical symptoms or behaviors, which are likely more age related (e.g., decreased activity due to change in social status) than cognitions, the CCL may not accurately capture distress cognitions in later life. There are benefits to utilizing measures across the lifespan for comparative purposes. However, important nuances of cognition may be missed in older adults. Given life events (retirement, widowhood, etc.) and developmental processes (social changes, physical changes, medical conditions), it is not surprising that automatic thoughts change across the lifespan. It is possible that it is the cognitions examined e or even just the phrasing of the cognitions e in the CCL that affected the relationship between cognitions and anxiety and depression symptoms. A cognitions measure that is specific to the automatic thoughts of older adults may elicit different results. Clinicians should be wary of using the CCL because of the bleed over of cognitions between late life anxiety and depression, which inhibits the utility of current diagnostic standards. Several limitations of the present study should be recognized. First, the study relied on a non-psychiatric sample of older adults and the generalizability of the findings to older adults with anxiety or mood disorders is unclear. Second, the study used a relatively homogenous sample that was largely female, Caucasian, and relatively well educated. We were also not able to obtain test-retest results of our measures to establish reliability. Nonetheless, the findings have both theoretical and practical relevance given that adults who do not meet criteria for anxiety disorders often show disability that is similar to those who pass symptom thresholds for such disorders.19e21 Another limitation of the study is that it used a cross-sectional design, and a prospective longitudinal design would
be necessary for any causal inferences about the influence any of the cognitive factors on the development of anxiety or depressive symptoms. Future work should examine other measures of anxiety and depression cognitions in older adult psychiatric sample as it has direct implications on cognitive-behavioral treatment. Finally, the study was limited by the use of exploratory factor analysis instead of confirmatory factor analysis. Exploratory factor analysis is traditionally used to reveal possible subscales or to reduce extraneous items in a measure, whereas confirmatory factor analysis is used to determine if the observed factor structure fits across samples. Future studies may want to examine the CCL in older adults using a confirmatory factor analysis to further evaluate the model fit of the subscales of the CCL. References 1. Watson D, Clark LA, Carey G. Positive and negative affectivity and their relation to anxiety and depressive disorders. J Abnorm Psychol. 1988;97:346e353. 2. Beck AT. Cognitive Theory and the Emotional Disorders. New York: International University Press; 1976. 3. Beck R, Perkins TS. Cognitive content-specificity for anxiety and depression: a meta-analysis. Cogn Ther Res. 2001;25:651e663. 4. Kvaal K, McDougall FA, Brayne C, et al. Co-occurrence of anxiety and depressive disorders in a community sample of older people: results from the MRC CFAS (Medical Research Council Cognitive Function and Ageing Study). Int J Geriatr Psychiatry. 2008;23:229e237. 5. Schoevers RA, Beekman AT, Deeg DJ, et al. Comorbidity and risk-patterns of depression, generalized anxiety disorder and mixed anxiety-depression in later life: results from the AMSTEL study. Int J Geriatr Psychiatry. 2003;18:994e1001. 6. Mather M. The emotion paradox in the aging brain. Ann N Y Acad Sci. 2012;1251:33e49. 7. Miloyan B, Byrne GJ, Pachana NA. Age-related changed in generalized anxiety disorder symptoms. Int Psychogeriatr. 2014;26:565e572. 8. Sütterlin S, Paap MC, Babic S, et al. Rumination and age: some things get better. J Aging Res. 2012;2012:267327. http://dx.doi.org/10.1155/2012/267327. 9. Clark DA, Beck AT, Stewart B. Cognitive specificity and positiveenegative affectivity: complementary or contradictory views on anxiety and depression? J Abnorm Psychol. 1990;99:148e155. 10. Riskind JH, Williams NL, Gessner TL, et al. The looming maladaptive style: anxiety, danger, and schematic processing. J Pers Soc Psychol. 2000;79: 837e852. 11. Beck AT, Brown G, Steer RA, et al. Differentiating anxiety and depression: a test of the cognitive content-specificity hypothesis. J Abnorm Psychol. 1987;96(3): 179e183. 12. Barlow DH. Unraveling the mysteries of anxiety and its disorders from the perspective of emotion theory. Am Psychol. 2000;55:1247e1263. 13. Beck JG, Novy DM, Diefenback GJ, et al. Differentiating anxiety and depression in older adults with generalized anxiety disorder. Psychol Assess. 2003;15(2): 184e192. 14. Shapiro AM, Roberts JE, Beck JG. Differentiating symptoms of anxiety and depression in older adults: distinct cognitive and affective profiles. Cogn Ther Res. 1999;23:53e74. 15. Lowe PA, Reynolds CR. Examination of the psychometric properties of the adult manifest anxiety scale-elderly version scores. Educ Psychol Meas. 2006;66(1): 93e115. 16. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. text rev. 4th ed. Washington, DC: American Psychiatric Publishing; 2000. 17. Reynolds CR, Bigler ED. Clinical Assessment Scale for the Elderly; Professional Manual for the CASE and CASE-SF. Lutz, FL: Psychological Assessment Resources, Inc; 2001. 18. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Health Disorders: DSM-5. 5th ed. Washington, DC: American Psychiatric Publishing; 2013. 19. de Beurs E, Beekman AT, van Balkom AJ, Deeg DJ, van Dyck R, van Tilburg W. Consequences of anxiety in older persons: its effect on disability, well-being and use of health services. Psychol Med. 1999;29:583e593. 20. Brenes GA, Guralnik JM, Williamson JD, et al. The influence of anxiety on the progression of disability. J Am Geriatr Soc. 2005;53:34e39. 21. Brenes GA, Guralnik JM, Williamson J, Fried LP, Penninx BW. Correlates of anxiety symptoms in physically disabled older women. Am J Geriatr Psychiatry. 2005;13:15e22.
