Depression Agency-Based Collaborative (Dep-ABC): Effect of Problem Solving Therapy on Risk of Common Mental Disorders in Older Adults with Home Care Needs

Abstract

Background:

Interventions to prevent depression in older adults have mainly focused on young-old ambulatory adults, not on the old-old with disabilities who receive supportive services in their homes.

Objective:

The Depression Agency-Based Collaborative (Dep-ABC) is a single-blind, pilot randomized trial assessing the effect of an intervention-development strategy using problem solving therapy (PST) on the risk of common mental health disorders in this vulnerable population.

Methods:

The intervention involved 6–8 sessions of PST over 12 weeks. Participants were followed up to 12 months post-intervention.

Results:

Dep-ABC randomized 104 participants, 68.4% of eligible and 17.5% of all older adults screened. The proportion of participants with incident Major Depressive Disorder or Generalized Anxiety Disorder was 11.4% in PST and 14.3% in the enhanced usual care control arm. A test of the interaction between time and intervention for anxiety symptoms favored the PST arm (p = 0.04).

Conclusions:

PST did not lower the risk of incident common mental illness but lowered anxiety symptom burden. Apart from low power, the effects of PST may have been blunted by referral for medical and aging services in the enhanced usual care group.

Introduction

In problem solving therapy (PST), participants work with a therapist to break problems down to manageable components and generate strategies to address concerns (1). PST’s mental health effect appears to be related to enhancing behavioral activation and self-efficacy. It offers benefit for older adults with diverse medical conditions (2, 3), across different intervention modalities (4–5), and as part of broader mental health interventions (6, 7).

The risk of depression and anxiety in older adults with supportive care needs is high. In seniors receiving home-delivered meals, for example, 12% met criteria for clinically significant depression (PHQ9 > 9) (8). In seniors receiving home health nursing, 13.5% met criteria for major depression (9). The 1-year incidence of major depression in homebound elders was 3%, and the 1-year incidence for subsyndromal depression 7% (10). Risk of depressive and anxiety symptoms in older adults is strongly related to severity of disability and need for assistance, not to age (11).

Offering PST to older adults with subsyndromal depression or anxiety and supportive care needs may be an effective prevention strategy (12–17). For these reasons, we developed the Depression Agency-Based Collaborative (Dep-ABC), a pilot prevention RCT. The primary outcome was time to onset of incident major depressive disorder (MDD), minor depressive disorder, or anxiety disorder. Secondary outcomes were change in depression and anxiety symptoms. Recruitment, screening, measures, and baseline status of Dep-ABC participants have been reported (18).

Methods

Dep-ABC was registered at ClinicalTrials.gov (NCT1566318). The University of Pittsburgh Institutional Review Board and an external data safety monitoring board reviewed procedures and protocols. All participants provided signed informed consent. Recruitment began in May 2012. Follow-up was completed in December 2015.

Screening and Eligibility

To identify older adults with disability, we approached local care management agencies that assess eligibility for long term care services. We also approached residents of subsidized senior housing. In screening for disability, we required that participants report difficulty in at least one instrumental or personal self-maintenance activity.

Potential participants were screened using the Patient Health Questionnaire (PHQ-9) to identify subthreshold symptoms (PHQ score of at least 1, with one or more of the cardinal symptoms of depression, anhedonia or dysphoria endorsed). (A small number of participants had scores higher than 9 but did not meet criteria for major depressive disorder on more extensive clinical assessment). At baseline, clinical interviewers also completed the SCID with potentially eligible participants.

Inclusion criteria for the trial required participants to be aged 60+; meet criteria for subsyndromal depression; score above 80 on the Modified Mini-Mental Status Exam (3MS) cognitive screen; and report need for assistance in daily tasks. Exclusion criteria included current use of an antidepressant or regular use of anxiolytics (> 4/wk); major depressive episode or anxiety disorder in the past 12 months; history of bipolar disorder, schizophrenia, or other neuropsychiatric disorders; or drug or alcohol treatment in the past 12 months. No participant reported onset of a neuropsychiatric condition over follow up.

Procedures

Participants were randomized 1:1 to receive either PST or enhanced usual care. A permuted block scheme using block sizes of 2, 6 and 8 was used to balance intervention assignments according to prior history of depression.

The usual care control arm was considered “enhanced” because participants in both arms who reported new medical symptoms or service needs were referred to home care agencies for assessment and changes in care plans.

Participants in both arms were assessed at baseline, directly after the intervention period (12 weeks after baseline), and then at 3, 6, 9, and 12 months for a total of 14 months for incident depression (major or minor) and anxiety disorders as defined by MINI/DSM-IV, PHQ-9, and GAD-7. Participants were assessed for outcomes by independent evaluators who were blind to the intervention.

Delivery of Problem Solving Therapy

Participants were randomized to problem-solving therapy (PST) or usual care. The intervention arm received 6–8 sessions of PST over 12 weeks, each session lasting 40–50 min, delivered in people’s homes. Total face-to-face time with therapists was approximately 7–8 hours. Participants in the PST arm received booster sessions at 6 and 9 months after the end of the initial intervention period. Therapists completed a log of participant-generated problems and strategies after each session, along with a rating of participant engagement in therapy. Problems identified by participants could be interpersonal issues but also practical problems related to service needs, such as adequacy of services, need for home modification, and responsiveness of service providers. PST therapists had greater than 10 years of mental health experience. A PST master trainer supervised therapists, who conducted nearly all assessments in participants’ homes, and also reviewed audiotapes of sessions to assess fidelity to training and suggest therapeutic strategies over the course of contact with participants.

Endpoints

Endpoints of this study were time to first onset of minor depressive disorder, major depressive disorder, anxiety disorder, or a conjoint endpoint of diagnoses. Major depression and minor depression/dysthymia were defined by confirming Structured Clinical Interview for Diagnosis (SCID) and Primary Care Evaluation of Mental Disorders-Mini International Neuropsychiatric Interview (PRIME-MD) diagnostic interview. Anxiety disorder was defined by SCID/MINI diagnosis. Secondary endpoints included depression and anxiety symptom burden measured longitudinally over follow up using PHQ-9 and GAD-7 scores.

Covariates

Dep-ABC participants completed an extensive battery of psychiatric, psychosocial, neurocognitive, and physical assessments at each assessment interval. The following baseline covariates were considered likely to affect outcomes and were included in pre-planned analyses: medical burden (Cumulative Illness Rating Scale for Geriatrics) (19), cognitive function (Modified Mini-Mental State Exam) (20), and physical function (Short Physical Performance Battery) (SPPB) (21).

Statistical Methods

All analyses were performed blind to outcomes using the principle of intention to treat, that is, all patients randomized were included over all of their observed time and were analyzed according to the group to which they were randomized regardless of their eligibility or compliance status.

For summarizing event-time data, Kaplan-Meier curves were employed (22). Log-rank tests (23) and Cox models (24) were used to test whether there were differences between interventions in defined endpoints. Formal tests of the proportionality assumption were conducted using the method of Grambsch and Therneau (25). To compare symptom trajectories between treatment arms, we employed a mixed-models approach using treatment assignment, time, assignment by time interaction, and history of previous depression or anxiety disorder.

Results

In this study, 104 participants of 595 screened were randomized at baseline, 68.4% of eligible and 17.5% of all older adults screened (18): 52 to usual care and 52 to problem solving therapy. Participants were mostly women approaching age 80; about 25% were non-white. The two study arms were similar at baseline in sociodemographic, health, and psychiatric measures.

Participant retention and loss to follow-up are shown in Figure 1. Among the 104 randomized patients, 11 did not complete the baseline or had no follow-up (n=3 enhanced usual care, n=8 PST) because of withdrawal of consent (n=3), death (n=7), or loss to follow up (n=1). Among participants with follow-up, 88.6% (39/44) in PST and 91.8% (45/49) in enhanced usual care completed all 6 assessments.

Figure 1.

Participants and Follow-Up, Dep-ABC

Uptake of Problem Solving Therapy

Participants completed a mean of 5.6 sessions (SD=1.9) of the 6–8 scheduled PST sessions. Therapist ratings of participation averaged 4.5 on a scale of 1–6.

Time to Depression, Anxiety, or Conjoint Diagnosis

For major depression, there were nine individuals with events: five (10.2%) in usual care and four (9.1%) in PST. Twenty-three people met criteria for minor depression: 12 in usual care and 11 in PST. For anxiety disorder, five people had events: three in usual care and two in PST. Combining all diagnoses, 33 individuals had at least one event: 18 (36.7%) in usual care and 15 (34.1%) in PST. Participants meeting the conjoint endpoint of depression or anxiety included 14.3% (7/49) in usual care and 11.4% (5/44) in PST. By 1 year 30% in each arm had met at least one diagnostic endpoint, and 10% met the major depression diagnosis endpoint.

Symptom Burden

While the average PHQ-9 score in the sample as a whole had a small but significant decrease over time (Wald t [df=427] = −2.12; p = 0.03), PHQ-9 scores in the two treatment arms did not change at different rates (i.e., the interaction between time and treatment arm was not significant, Wald t [df=427] = 0.68; p=0.49). The average decrease in GAD-7 scores over time was marked (Wald t [df=426] = −3.66; p < 0.01), and a test of interaction between time and intervention was statistically significant (Wald t [df=426] = 2.09; p = 0.04).” The maximum change in scores for PHQ-9 and GAD-7 in the two study arms was −0.8 and −1.2, respectively, favoring the PST group.

Discussion

In Dep-ABC, PST did not differ from enhanced usual care as a prevention strategy for common mental disorders but was associated with a greater diminution in anxiety symptoms as measured by the GAD-7. After 12 months, 30% in each treatment arm had met at least one diagnostic endpoint, and 10% met the major depression diagnosis endpoint. PHQ-9 and GAD-7 symptom burden scores showed a small but significant decrease over time in both treatment arms, and a test of the interaction between time and intervention for anxiety symptoms favored the PST arm. The two arms did not differ in relapse despite use of PST booster sessions.

Lack of a significant difference between study arms may have a number of sources. The relatively small sample size suggests that statistical power limited our ability to detect differences. Yet the absolute difference in the incidence of common mental illness was low between study arms. This would suggest absence of a treatment effect, perhaps due to the many psychosocial challenges this population experiences. An alternative explanation involves the “enhanced” quality of usual care in the comparison group. Referrals for medical and psychiatric care and aging services throughout the trial can be considered a form of case management; thus Dep-ABC can also be seen as a comparison of case management plus PST vs. case management alone. Lack of superiority of PST in the setting of case management for prevention of depression replicates findings for PST depression treatment trials (26–27).

Conclusion

Problem solving therapy did not lower the risk of incident common mental illness in older people with home care needs in the setting of ongoing assessment and referral for medical care and aging services. In keeping with the exploratory and intervention-development nature of this pilot, the sample was small, which likely affected power to detect differences between treatment arms. Future research should consider adding structured measures of referral and other non-protocol interventions. Including such measures in prevention trials may help separate the effects of enhanced care from treatment.

Highlights.

1) What is the primary question addressed by this study. To determine if problem-solving therapy (PST) reduces the risk of common mental health disorders in the old-old who need or receive supportive services because of disability.

2) What is the main finding of this study? The proportion of participants with incident Major Depressive Disorder or Generalized Anxiety Disorder over 12 months did not differ between treatment arms (11.3% in PST and 14.2% in enhanced usual care).

3) What is the meaning of the finding? The effects of PST may have been blunted by referral for medical and aging services in the enhanced usual care group; future trials should consider adding structured measures of referral and other non-protocol interventions.