|Home | About | Journals | Submit | Contact Us | Français|
To develop and test a modular psychotherapy protocol in older primary care patients with anxiety disorders.
Randomized, controlled pilot study.
University-based geriatric medicine clinics.
31 elderly primary care patients with Generalized Anxiety Disorder or Anxiety Disorder Not Otherwise Specified.
Modular form of psychotherapy compared to enhanced community treatment.
Self-reported, interviewer-rated, and qualitative assessments of anxiety, worry, depression, and mental health-related quality of life.
Both groups showed substantial improvements in anxiety symptoms, worry, depressive symptoms, and mental health-related quality of life. Most individuals in the enhanced community treatment condition reported receiving medications or some other form of professional treatment for anxiety. Across both conditions, individuals who reported major life events or stressors and those who used involvement in activities as a coping strategy made smaller gains than those who did not.
Results suggest that modular psychotherapy and other treatments can be effective for anxiety in older primary care patients. Results further suggest that life events and coping through increased activity may play a role in the maintenance of anxiety in older adults.
Anxiety disorders are common in the elderly, with community prevalence estimates as high as 14%, and higher rates in medical settings (1). Generalized Anxiety Disorder (GAD) appears to be the most prevalent condition in community samples (2), but Anxiety Disorder Not Otherwise Specified (ADNOS) is the diagnosis most often assigned in primary care settings where older adults with anxiety typically seek treatment (3).
GAD and other anxiety syndromes are associated with many negative outcomes, including disability, increased health service use, poorer quality of life, and risk of developing major depression (4, 5, 6). Prospective research in the elderly has demonstrated that anxiety symptoms are associated with medical illness such as coronary heart disease (7), self-reported mobility limitations (8), and higher levels of health care utilization, including more primary care visits and more time spent during each visit (3).
Data suggest that anxiolytic medications are effective for geriatric anxiety (9, 10). Many older adults, however, prefer nonpharmacological treatments for mood symptoms (11). Although cognitive-behavioral therapy (CBT) has shown some benefit for late-life anxiety (12), the evidence suggests that conventional CBT may be less effective and associated with higher attrition rates than in younger adults (13), supporting the need for the development of new treatment protocols, especially ones that can be implemented in primary care.
A relatively new approach in psychotherapy research involves modular treatment, in which different components are used depending upon the patients’ presenting problems or symptoms (14, 15). This approach is consistent with the increasing emphasis on personalized medicine. The rationale for modular treatment of anxiety in older primary care patients is two-fold: first, anxiety in medical settings may be more heterogeneous than anxiety in the specialty mental health sector, suggesting that “one-size-fits-all” packages may be less effective in this setting (16). Second, the current cohort of older adults is, on average, less socialized to psychotherapy and more sensitive to the stigma associated with mental illness than are younger adults (17). Tailoring the treatment to the specific symptoms and needs of older participants can ensure that the psychotherapeutic process is seen as responsive and clinically relevant as a way of reducing attrition and enhancing engagement.
In the present study, we developed a modular psychotherapy protocol to treat older primary care patients with GAD and ADNOS. We tested the protocol against an enhanced community treatment condition in which patients and their primary care providers were given information about the anxiety diagnosis. We conceived of this project as a pilot effectiveness study to examine how well this novel approach to psychotherapy would work for patients seen in Geriatric Medicine. We hypothesized that the modular psychotherapy intervention would result in greater decreases in anxiety and depression and greater increases in quality of life than community treatment.
Patients were 31 adults at least 60 years old with a principal (e.g., most severe) or co-principal diagnosis of GAD or ADNOS diagnosed according to the Anxiety Disorders Interview Schedule for DSM-IV (ADIS; 18) administered by a Ph.D. level clinician. Recruitment was conducted from August 2004 through August 2005 using a primary care screening procedure and self-referrals. A screening questionnaire consisting of the two GAD items from the Patient Health Questionnaire was mailed to 1,000 patients selected at random from a list of 5,132 patients who had received services from the UCSD Geriatrics clinics within the previous year. Two-hundred and three questionnaires were returned, 89 of them with at least one item endorsed. Fifty-eight of these patients were screened by telephone using the Mini International Diagnostic Interview (MINI; 19). Of those, 24 received an in-person diagnostic interview using the ADIS, which was conducted within approximately two weeks of the phone screening. Seventeen of these patients were enrolled; one additional patient went into the hospital shortly after his interview and had to be withdrawn from the study.
Thirty-one individuals contacted the study team after seeing recruitment flyers or advertisements. Of those, 11 did not qualify for an in-person diagnostic interview. Of the 20 who completed the ADIS, an additional five did not qualify. The remaining 15 patients were enrolled into the study.
Results from the MINI and ADIS concurred in 24 of the 31 cases; in those cases in which the raters disagreed, a third Ph.D.-level clinician viewed the videotape of the ADIS, and diagnosis was determined by consensus. The majority of patients who did not qualify for the study did not meet criteria for GAD or ADNOS (84%). Other exclusion criteria were cognitive impairment (n = 1), substance use disorders within the past 6 months (n = 1), serious medical conditions (n = 2), and language barrier (n = 1). Written informed consent was obtained from all participants, and the study was conducted in compliance with the Human Research Protections Program of the University of California, San Diego and a Data Safety Monitoring Board.
Demographic and clinical information about the participants is presented in Table 1. Almost all were diagnosed with GAD. Overall, they were a well-educated, affluent group. Those with higher levels of education reported lower levels of anxiety symptoms at baseline, r (df=29) = -.39, p = .03. Although cognitive performance as measured by the MMSE was high on average, we found a significant inverse relationship between MMSE scores and baseline anxiety symptoms, r (df=29) = -.45, p = .01, indicating that those with poorer cognitive functioning reported higher levels of anxiety. This pattern is consistent with other research on cognitive functioning and anxiety (20). Axis I comorbidities included major depression (4 patients), panic disorder (4 patients), dysthymia (2 patients), agoraphobia without a history of panic disorder, social phobia, and specific phobia (1 patient each).
Participants were permitted to take psychotropic medications or use concurrent mental health services provided that they had been on a stable regimen for at least two months prior to enrollment. We permitted concurrent treatment in order to allow better generalization of the results to primary care patients, many of whom receive pharmacotherapy or other treatments in that setting. Data on mental health treatment at baseline and changes over the course of the study are presented in Table 2. At baseline, 26.7% of the modular psychotherapy group and 50.0% of the community treatment group were taking daily prescribed psychotropic medication, which was not significant according to Fisher’s exact test, p = .17. There were no differences in anxiety or worry symptoms between individuals taking medications and those who were not. An additional person in the modular psychotherapy group and three in the community treatment group were taking prescribed psychotropic medications on an as-needed basis.
Participants were randomly assigned to receive either 12 weekly, individual sessions of modular psychotherapy (n = 15) or enhanced community treatment (n = 16). Twelve sessions were chosen based on the average number of sessions delivered in other late-life GAD treatment studies; patients took on average 14 weeks to complete these sessions (11). Randomization was performed without blocking or stratification, following the baseline assessment and diagnostic case conference, using a coin toss.
Modules were included in the intervention protocol based on empirical evidence of efficacy from previous investigations and theoretical relevance to GAD, anxiety symptoms, and frequently comorbid conditions in older medical patients (for additional information about the protocol, see 21; manual also available upon request). Modules included: 1) education about anxiety and symptom monitoring; 2) relaxation training, including diaphragmatic breathing, progressive muscle relaxation, and imagery; 3) cognitive restructuring; 4) thought-stopping and scheduled worry; 5) exposure through systematic desensitization; 6) behavioral activation, consisting of pleasant events scheduling; 7) sleep hygiene guidelines; 8) problem-solving skills training, which involved learning to specify a problem and brainstorm, evaluate, and implement solutions; 9) life review, a structured journaling exercise designed to change long-standing negative beliefs; 10) acceptance, which included mindfulness exercises, discussion of values, and goal-setting; 11) assertiveness training; 12) time management; 13) pain management; and 14) relapse prevention.
Most participants received seven or eight of the 14 possible modules. The decision about which modules to use for each patient was based on a problem list generated by the patient and comorbid diagnoses or symptoms (e.g., those with a comorbid depression diagnosis were targeted to receive the behavioral activation module). Both therapist and patient agreed on modules that would target the patient’s most distressing problems and symptoms. Of the 12 patients who completed the modular psychotherapy intervention, all received education, relaxation, and relapse prevention; 11 each received problem-solving and acceptance; 8 received thought-stopping and scheduled worry; 7 received sleep hygiene; 6 received behavioral activation; 5 received assertiveness training; 3 received life review; 2 each received cognitive therapy and time management; and 1 received pain management.
Patients randomized to enhanced community treatment and their health care providers received information about the patients’ anxiety disorder diagnosis. Patients were contacted after their in-person assessment and informed of their diagnosis. Additionally, a letter was sent to their health care provider documenting the diagnosis and instructing them to continue to treat the patient as they otherwise would. Patients were not given specific referrals or other information about enhanced community treatment, only that they should continue to be followed by their chosen provider. Patients received a second in-person assessment but no other contact or referrals from study staff. Table 2 displays treatment received during the course of the study; most participants received psychotropic medications from their community health care providers.
Psychotherapy was performed by M.A. and Ph.D.-level clinicians supervised by the first author. All therapy sessions were videotaped and reviewed in individual supervision sessions to maintain fidelity to the treatment protocol. Review of a random sample of four tapes (one per therapist) for adherence and competence (each rated on a 0-4 scale) by an expert rater indicated that all therapists achieved a score of 4 (“excellent”) on adherence, and three scored 4 on competence (the remaining therapist scored 3, “good”). Participants were given daily homework assignments. They had an 82% average completion rate, defined as the proportion of days on which the patient reported completing at least some homework, as documented on forms turned in weekly to the therapists; the rate excluding data from one participant who completed no homework was 95%. Assessments were conducted by two research assistants who were not informed of the patients’ treatment conditions. All patients completed assessments at baseline and after approximately three to four months (following treatment for the modular psychotherapy patients).
The Hamilton Anxiety Rating Scale (HAMA; 22) is a 14-item interviewer-rated measure of anxiety primarily assessing somatic symptoms. It is considered the “gold standard” outcome measure in studies of anxiety pharmacotherapy treatment. It has been validated in samples of older GAD patients and normal community volunteers (23). Interrater reliability as measured by the interclass correlation coefficient was .94.
The Penn State Worry Questionnaire (PSWQ; 24) is a 16-item self-report instrument designed to measure trait worry. It has been validated in samples of older adults diagnosed with GAD (25). It was chosen for this investigation because it is the most widely used self-report measure of pathological worry. Cronbach’s alpha in this study was .77.
The Beck Depression Inventory-II (BDI) is a 21-item self-report scale listing common symptoms of depression that the respondent may have experienced in the past two weeks (26). The scale is a revision of one that has been used extensively in research and clinical settings since its development in 1961 and has been used with geriatric populations (27). It is the most widely used self-report depression measure in clinical samples. Cronbach’s alpha in this sample was .86.
Mental health-related quality of life and functioning was evaluated using the Mental Component Score of the Medical Outcomes Study 36-item short form self-report health survey (SF-36; 28). This variable is reported as a T-score in which higher numbers represent better quality of life. It has been validated in large, mixed-age medical and psychiatric samples and used with older GAD patients (29).
Data were collected on medical conditions using the Cumulative Illness Rating Scale - Geriatrics (30), and use of health services including primary care visits, medical specialist visits, mental health visits, and self-help group attendance using the Cornell Service Index (31). In keeping with the pilot nature of this project, patients assigned to modular psychotherapy were asked to rate the helpfulness of the individual modules they received. Patients also rated their satisfaction with the overall treatment using the 8-item Client Satisfaction Questionnaire (32).
Finally, we conducted brief, semi-structured interviews with 26 patients, 15 in the community treatment condition and 11 in the modular psychotherapy condition, by telephone to inquire about their subjective perceptions of improvement in anxiety symptoms (rated as “better,” “about the same,” or “worse”). Only one patient reported worsening symptoms, so ratings were dichotomized into improved vs. not improved. They were also asked about coping strategies, life events, and other factors that may have influenced change in anxiety levels. Patients generated their own list of life events and ways of coping in response to a standard set of open-ended questions; this method could have been biased by retrospective recall. This procedure was added to the study after it was in progress, so calls occurred in some cases as long as 12 months after the four-month assessment; by that point, five individuals had dropped out or could not be reached. Interviews were transcribed and responses were coded by two independent raters.
Codes were developed by the first rater based on interview content. The kappa statistic was used to evaluate agreement between raters on each code; kappas ranged from 0.77 to 1.00, representing good to excellent agreement. For analysis, we grouped the responses into the broad categories of professional help-seeking (e.g., psychotherapy, talking to a health care provider about anxiety), self-help strategies (e.g., meditation), life events or stressors (e.g., “milestone” birthday, medical problems), and involvement in activities (e.g., “keeping busy”).
Analyses were conducted using SPSS release 11.5. Groups were compared on clinical and demographic variables using t-tests and chi-square tests. Outcomes were evaluated using repeated measures analysis of variance (RMANOVA). Because of the high proportion of participants receiving concurrent daily pharmacotherapy, this variable was included as a covariate in the RMANOVA models. Data were analyzed on an intent-to-treat basis; all patients completed both assessments. The assumption of equality of the variance-covariance matrices was fulfilled for all of the RMANOVAs (Box’s M values ranging from 8.19, p = .06, to .503, p = .93). No dependent variable displayed a significant departure from normality based on Kolmogorov-Smirnov tests (z values ranging from .59 to .76).
Results from the outcome analyses are presented in Table 3. We found substantial and comparable decreases in anxiety, worry, and depressive symptoms and improvement in mental health-related quality of life in both the community treatment and modular psychotherapy groups. This pattern of results did not change when medication use was removed from the models. The within-group mean effect size (weighted average Hedges’ g) for change in worry and anxiety symptoms for the modular psychotherapy condition (g = 1.23) is somewhat larger than in most previous studies of psychotherapy with anxious older adults (e.g., 29, 33). The within-group effect size for the community treatment condition (g = 1.67) is, to our knowledge, higher than has ever been reported for a minimal contact condition in a GAD treatment study (14).
The attrition rate for patients who received modular psychotherapy was 20% (although all patients in both groups returned for post-treatment assessments). The three patients who dropped out of therapy did so after 2, 5, and 6 sessions; one cited health problems, one a lack of time, and the third reported that she was doing much better and no longer required help with anxiety. Participants who received modular psychotherapy reported higher levels of satisfaction than those who did not, t(26) = 2.11, p < .05. For modules received by at least 3 participants, the mean rankings of helpfulness, from most to least helpful, were as follows: 1) relaxation, 2) pleasant activities, 3) psychoeducation and monitoring, 4) acceptance, 5) worry control, 6) relapse prevention, 7) sleep hygiene, 8) problem-solving, 9) assertiveness training, and 10) life review.
Table 2 displays data on changes in mental health service use for patients in the modular psychotherapy and community treatment conditions. There were no statistically significant differences between the conditions in use of “as-needed” psychotropic medications, over-the-counter medications, professional mental health services, or self-help groups. More individuals in the community treatment group were taking prescribed medications for anxiety daily at the four-month assessment, 56.3% vs. 26.7%; this difference approached statistical significance according to Fisher’s exact test, p = .096. Across both groups, those taking such medications (n = 13) showed less change in their worry symptoms than those who were not (n = 18), t(29) = 2.34, p = .03.
A significantly higher proportion of the modular psychotherapy group than the community treatment group reported subjective improvement in anxiety, 90.9% vs. 40.0%, p = .01 according to Fisher’s exact test. Those who reported improvement, however, did not show greater change on objective measures of anxiety or worry than those who reported no improvement. Specific self-help strategies are displayed in Table 2. Across both conditions, individuals who reported receiving professional help for anxiety (n = 15) and those who used self-help strategies like bibliotherapy or meditation (n = 18) showed no greater change in anxiety or worry symptoms than those who did not (n = 11 and n = 8, respectively). Individuals who reported major life events, either positive or negative, or stressors (n = 11) showed less change in anxiety symptoms than those who did not report such events (n = 15): t(24) = 2.44, p = .02. Those who reported increasing their engagement in activities (e.g., keeping busy, getting more involved in groups or organizations; n = 10) showed less change in anxiety symptoms than those who did not (n = 16), t(24) = 2.26, p = .03.
In a comparison between a modular form of psychotherapy and enhanced community treatment (in which patients and their primary care providers were given information about their anxiety diagnosis), both groups experienced equivalent symptomatic improvement. Substantial, statistically significant improvements were found in anxiety symptoms, worry, depressive symptoms, and mental health-related quality of life in both groups.
Individuals in both conditions used a wide variety of anxiety management strategies, including medications, psychotherapy, and various self-help techniques. Despite randomization, a higher proportion of individuals in the community treatment group were using psychotropic medications; medication use was therefore included as a covariate in the analyses. Those who used medications showed less improvement in worry symptoms than those who did not, suggesting that medication use might have represented a marker for treatment resistance rather than a successful treatment strategy in this sample.
Professional services and self-help techniques were not associated with greater improvement in anxiety. Although individuals reported pleasant activities as the second most helpful module, those who increased their involvement in activities showed smaller declines in their anxiety symptoms than those who did not. “Keeping busy” may be an unhelpful avoidance strategy for anxious older adults. These results are interesting given that behavioral activation has been an effective treatment for geriatric depression. Major life events, both positive and negative, and stressors were also associated with less improvement in anxiety symptoms. This finding is consistent with much research documenting that stress is associated with both positive and negative events, and further suggests that factors beyond an individual’s control may influence anxiety outcomes regardless of treatment.
Overall, these data provide no evidence that a modular psychotherapy protocol for anxiety in older primary care patients is more effective than enhanced community treatment, when the latter involves aggressive pharmacotherapy and multiple other services and strategies. Although the within-group effect size for the modular psychotherapy intervention was quite high relative to other late-life anxiety trials, the effect size in the community treatment condition was substantially higher than has been reported in the past for a minimal contact condition. Thus, results from this study should not be used to support an argument that services available in most primary care practices are sufficient to treat chronic anxiety in older adults.
Other forms of psychotherapy currently under investigation in younger adults with GAD and similar disorders include an integration of CBT and interpersonal therapy (34), acceptance and commitment therapy (35), attention training (36), and therapies targeting intolerance of uncertainty (37), emotion dysregulation (38), and metacognition (39). Only two have been tested with older adults: Ladouceur and colleagues reported success treating eight older GAD patients using therapy based on the intolerance of uncertainty model (37), and Papageorgiou and Wells published a case series of three geriatric patients with hypochondriasis who responded to attention training (36).
Age-related changes may make some of these alternatives particularly appropriate for anxious older adults. For example, some evidence suggests that older adults regulate emotion better than do younger adults (40), making an emotion regulation-based approach potentially appealing for use in geriatric populations. With the high prevalence and negative consequences associated with anxiety in late life, testing such innovative protocols with older anxiety disorder patients should be a high priority.
The chief clinical implication is that many types of interventions and strategies can lead to improvements in symptoms and quality of life for older people with anxiety. This is a positive message for older adults with this chronic and distressing condition. It is important to note, however, that the level of care provided in this study across both the modular psychotherapy and community treatment conditions is not representative of care for late-life anxiety disorders in most primary care settings. These results should be interpreted with some caution given that this sample was Caucasian, well educated, and wealthy. Participants had the means and access to alternative treatments such as yoga and meditation classes. Older persons without the protective factors and resources of subjects in our sample may not have shown improvements despite positive attitudes about care.
Scientifically, it is of interest to discover why the types of psychotherapy for geriatric anxiety investigated thus far appear to work less well than medications (11) or than psychotherapy for anxiety in younger individuals (14), whereas the same is not true for geriatric depression (41). Additional research on psychosocial, biological, and cognitive factors associated with anxiety disorders in late life may ultimately inform new and more effective behavioral treatment protocols.
Research supported by NIMH Grant No. MH067643. The authors gratefully acknowledge the contributions of Georgia Birchler, Debora Goodman, Dilip V. Jeste, M.D., Laura Otis, Ph.D., and Murray B. Stein, M.D., M.P.H.