Adherence to Depression Treatment in Primary Care: A Randomized Clinical Trial

Key Points

Question  Can a brief psychosocial intervention (ie, Treatment Initiation and Participation Program) targeting medication barriers improve antidepressant adherence among middle-aged and older adults with a newly initiated depression treatment by primary care physicians?

Findings  In this randomized clinical trial, the Treatment Initiation and Participation Program participants were 3 times more likely to be at least 80% adherent to their antidepressant pharmacotherapy at both 6 and 12 weeks. The Treatment Initiation and Participation Program did not have a sustained effect on depression; however, participants who reported 80% adherence at both 6 and 12 weeks showed greater improvement in depression severity by 24 weeks.

Meaning  The Treatment Initiation and Participation Program improved early adherence at weeks 6 and 12 combined and participants showed an early reduction in depressive symptoms at 6 weeks; participants in the Treatment Initiation and Participation Program and treatment as usual groups who were adherent at weeks 6 and 12 had a greater improvement in depressive symptoms.

Importance  Nonadherence to antidepressant medication is common and leads to poor outcomes. Early nonadherence is especially problematic.

Objective  To test the effectiveness of a psychosocial intervention to improve early adherence among older patients whose primary care physician newly initiated an antidepressant for depression.

Design, Setting, and Participants  The Treatment Initiation and Participation Program (TIP) was offered in a 2-site randomized clinical effectiveness study between January 2011 and December 2014 at primary care practices in New York, New York, and Ann Arbor, Michigan. Analyses began in February 2016. All participants were middle-aged and older adults (aged ≥55 years) who received newly initiated depression treatment by their primary care physician and recruited within 10 days of their prescription. Analyses were intention-to-treat.

Interventions  Participants were randomly assigned to the intervention (TIP) or treatment as usual. Participants in the TIP group identified and addressed barriers to adherence, including stigma, misconceptions, and fears about treatment, before developing a personalized adherence strategy. The Treatment Initiation and Participation Program was delivered in three 30-minute contacts scheduled during a 6-week period just after the antidepressant was prescribed.

Main Outcomes and Measures  The primary outcome was self-reported adherence on the Brief Medication Questionnaire, with adequate early adherence defined as taking 80% or more of the prescribed doses at 6 and 12 weeks. The secondary outcome was depression severity.

Results  In total, 231 middle-aged and older adults (167 women [72.3%] and 64 men [27.7%]) without significant cognitive impairment were randomly assigned to the TIP intervention (n = 115) or treatment as usual (n = 116). Participants had a mean (SD) age of 67.3 (8.4) years. Participants in the TIP group were 5 times more likely to be adherent at 6 weeks (odds ratio, 5.54; 95% CI, 2.57 to 11.96; χ²₁ = 19.05; P < .001) and 3 times more likely to be adherent at both 6 and 12 weeks (odds ratio, 3.27; 95% CI, 1.73 to 6.17; χ²₁ = 13.34; P < .001). Participants in the TIP group showed a significant earlier reduction (24.9%) in depressive symptoms (95% CI, 13.9 to 35.9; t₃₃₇ = 4.46; adjusted P < .001). In both groups, participants who were 80% adherent at weeks 6 and 12 had a 15% greater improvement in depressive symptoms from baseline over the course of treatment (95% CI, −0.2 to −30; t₃₆₉ = 1.93; P = .051).

Conclusions and Relevance  The Treatment Initiation and Participation Program is an effective intervention to improve early adherence to pharmacotherapy. Improved adherence can promote improvement in depression.

Trial Registration  clinicaltrials.gov Identifier: NCT01301859

Quiz Ref IDNonadherence to antidepressant therapy remains a major challenge to quality depression care. Rates of nonadherence among older adults range from 29% to 40% in the United States.^1,2 Most depression management is provided in primary care,³ and higher rates of antidepressant nonadherence are documented in primary care compared with psychiatry sites.^4,5 Although nonadherence is problematic throughout treatment, the first 6 weeks of treatment is a particularly critical period to promote adherence, with increased risks of treatment dropout, relapse,⁶ medication discontinuation,^7,8 vulnerability to suicide,⁹ and greater economic burden¹⁰ among those who show early nonadherence to antidepressants. There is evidence that early adequate antidepressant dosing and good adherence are both associated with recovery from depression^11,12 and may improve the long-term outcomes of patients with depression.¹⁰

Quiz Ref IDKey risk factors for nonadherence include age, comorbid conditions, beliefs about treatment, and concerns over adverse events.¹⁰ In the elderly population, additional risk factors may include patient variables such as the lack of a medication routine, retaining discontinued medications, combining prescriptions, and multiple storage locations.¹³ Medication adherence problems also increase with the total number of drugs prescribed.¹⁴ The average older American adult takes 3 prescription and 4 over-the-counter medications daily, and those with depression typically take more medications.¹⁵ Even in successful clinical intervention studies for older adults with depression, nonadherence is a challenge.¹⁶

Although older adults face structural barriers (eg, costs of medications, distance from home to physician’s office), often negative attitudes are the most important factors affecting adherence.¹⁷ Perceived stigma predicts poorer medication adherence and treatment discontinuation among elderly people with depression^7,8,18 Older adults worry about having a diagnosis of depression, especially older adults of African descent.^2,19,20 Low perceived symptom severity is associated with poorer adherence,^1,7,17,21 and even when distress is acknowledged, many older adults feel they should not need mental health help.²² In a community-based study of older adults, when perceived costs outweighed perceived benefits, nonadherent behaviors were more likely.²³

The Treatment Initiation and Participation Program (TIP)²⁴ is a brief psychosocial intervention designed to improve adherence to pharmacotherapy for patients with depression. The Treatment Initiation and Participation Program is informed by the theory of reasoned action²⁵ and targets individual-level, modifiable factors such as psychological barriers (eg, stigma, self-efficacy), fears about antidepressants, misattributions about causes of depression, and the lack of an adherence strategy. The program helps patients address barriers, identify treatment benefits, and feel empowered to manage their medication regimen and communicate with the physician effectively.²⁶

For this 2-site randomized clinical effectiveness trial of TIP, we targeted the critical period of early adherence among middle-aged and older adults with a newly initiated antidepressant treatment for depression by their primary care physician (PCP). We hypothesized TIP participants would be more likely to be at least 80% adherent to their antidepressant medication at both 6 and 12 weeks after the prescription was provided compared with participants who received treatment as usual. In addition, we hypothesized that TIP participants would have a greater reduction of depressive symptoms compared with participants receiving treatment as usual. Finally, we explored whether greater adherence in the TIP and treatment as usual groups was associated with decreased depression at 24 weeks.

The trial protocol (available in the Supplement) was approved by the institutional review boards at Weill Cornell Medicine and the University of Michigan. Potential participants were identified at the participating primary care sites through daily electronic medical record system reports on antidepressant prescriptions. Trained staff reviewed each medical record to verify inclusion eligibility (eg, a newly prescribed antidepressant for treatment of depression) prior to contacting patients. To capture early adherence decisions among a wide range of patients, all eligible patients were cold-called by staff to gauge their interest in the study within 10 days of their prescription date. Patients who expressed interest were met in person by research staff for a more in-depth discussion of the study, written informed consent, and baseline assessment. The meetings were held at either the primary care office or in nearby research offices. Participants who met eligibility criteria during the baseline assessment were randomized 1 to 1 (stratified by site) to receive either the TIP intervention in addition to the antidepressant monitoring provided by their PCP or treatment monitoring as usual provided by the PCP. The TIP sessions were delivered within the first 6 weeks of pharmacotherapy. Research assessments were collected at 6, 12, and 24 weeks to evaluate the impact of TIP on adherence and depressive symptoms. Participants in both groups were compensated $40 for the longer baseline and $25 for each of the 3 follow-up research visits. The 2 site principal investigators (J.A.S. and H.C.K.) oversaw the conduct and data analysis of the trial and data with an annual review by an independent data and safety monitoring committee.

Participants were consecutively recruited beginning in January 2011 through December 2014, and analyses began in February 2016 (Figure). Patients 55 years and older were eligible; this age cutoff was designed to include middle-aged adults who might be more similar to older adults in terms of medical burden²⁷ or socioeconomic disadvantage.²⁸ All patients had been newly prescribed an antidepressant by their PCP specifically for depressive symptoms. Given the high rates of prescription of antidepressants among older adults,²⁹ this target population was chosen to test the effectiveness of TIP among patients typically treated by PCPs according to guidelines from the National Institute for Health and Care Excellence.³⁰ Patients were excluded if they had (1) active suicidality; (2) current substance abuse, bipolar disorder, or psychosis; (3) significant cognitive impairment (Mini-Mental State Examination score ≤23); (4) terminal illness or current chemotherapy; or (5) an inability to communicate in English.

Quiz Ref IDThe TIP intervention included 5 steps: (1) review symptoms and antidepressant regimen and conduct a barriers assessment; (2) define a personal goal that could be achieved with adherence; (3) provide education about depression and antidepressant therapy; (4) collaborate to address barriers to treatment participation; and (5) create an adherence strategy and empower the older adult to talk directly with the PCP about treatment. The study counselor documented each TIP intervention meeting on a contact sheet that helped structure and standardize the TIP intervention. For this study, the TIP intervention was delivered in three 30-minute meetings during the first 6 weeks of pharmacotherapy. The 3-session format was designed to allow for an alliance between the TIP counselor and participants while ensuring brevity and decreasing potential dependency.

The intervention was delivered by 2 master’s-level social workers trained in the TIP protocol by 1 of us (J.A.S.) during a full-day training, with 6 months of weekly supervision. Treatment Initiation and Participation Program competency certification was based on audiotape ratings of practice cases by the principal investigator. After certification and the initial 6 months, TIP supervision was conducted bimonthly. Two randomly selected audiotapes from each counselor were reviewed quarterly to evaluate drift. If there were indications of drift, corrective interventions were integrated into supervision.

Research assessments were conducted at baseline and at 6, 12, and 24 weeks by trained research assistants blinded to group assignment. The structured clinical interview from DSM-IV was administered to confirm a diagnosis of depression.³¹ The 24-item Hamilton Depression Rating Scale (HAM-D) was used to assess the severity of depressive symptoms.³² The use of services and number of medical visits were recorded at baseline and at 12 and 24 weeks using the Cornell Services Index.³³

Early adherence was measured using the Brief Medication Questionnaire, which is a self-report measure validated against an electronic bottle cap data (MEMS Cap; AARDEX Group Ltd) and used extensively with older adults.^34,35 Adequate antidepressant adherence was defined as 80% adherence to assure that most doses are taken to achieve a biological effect (eg, therapeutic level).^26,36,37 For this project, the goal was a sustained period of early adherence defined as self-reported 80% adherence or greater at both the 6- and 12-week assessments. Self-reported adherence was selected both to capture acknowledged nonadherence and to minimize the effects of the adherence measure (such as MEMs Caps) on participants’ behavior.

Depression outcomes were measured based on percentage reductions in depressive symptoms on the HAM-D. Because the inclusion criteria were PCP initiation of a new treatment of depression with antidepressant medication, rather than an official diagnosis, depression severity was expected to be heterogeneous.

A sample size of 130 participants per treatment arm was planned to have 90% power to detect a 70% (TIP) vs 50% (treatment as usual) adherence rate (≥80%) at weeks 6 and 12 for a 2-sided test at α = .05. Sociodemographic, functional, or clinical differences were evaluated using t and χ² tests comparisons. In separate logistic regression models, we examined group differences on achieving 80% adherence to antidepressants at week 6 and week 12 and a combined measure of adherence at both 6 and 12 weeks. Percentage improvements in depressive symptoms (HAM-D scores) from baseline to weeks 6, 12, and 24 were modeled in an intent-to-treat mixed-effects regression model with a participant-specific random intercept and fixed effects for study site, treatment, time, and treatment by time interaction. To examine the effect of adherence on depressive symptoms, a similar mixed-effects regression model was used, which included additional fixed effects of adherence, adherence by treatment, adherence by time, and adherence by treatment by time interactions. As none of the interactions were significantly different from 0, they were dropped from the model. These analyses were repeated in the subgroup of participants (n = 175) with mild-to-moderate depressive symptoms recorded during the baseline (HAM-D score, ≥10). Post hoc tests were conducted using the least-square means, and P values were adjusted for multiple comparisons using Bonferroni-Holm step-down procedure. All tests conducted used a 2-sided alternative at a significance level of α = .05, and all statistical analyses were conducted using SAS software, version 9.4 (SAS Institute Inc).

To target early adherence, we identified patients with a newly prescribed antidepressant for depression (no antidepressant in the past 3 months). Of the 607 patients who met initial study inclusion criteria, 277 (45.6%) consented and 231 were randomized to receive either the TIP intervention in addition to their antidepressant and PCP care (115 [49.8%]) or treatment as usual with the PCP (116 [50.2%]) (Figure). All participants were recruited from either Weill Cornell Internal Medical Associates in New York, New York (114 [49.4%]), or Michigan Medicine outpatient primary care clinics in Ann Arbor (117 [50.6%]). Primary outcome data were collected at 6 weeks from 211 participants (91.3%); 200 at 12 weeks (86.6%); and 199 at 24 weeks (86.1%). There were no significant differences in retention rates by treatment group, sex, race/ethnicity, age, site, or education level (Table 1).

Cross-site interrater reliability was calculated based on audiotapes of participants’ assessments compared with a criterion standard rater (clinical psychologist) for the structured clinical interview and HAM-D depression severity ratings. Raters maintained good reliability throughout the study (Spearman r = 0.81-0.95). Treatment Initiation and Participation Program intervention delivery was monitored by the study principal investigator (J.A.S.) who listened to 18 intervention audiotapes provided by the TIP counselors to assure continued competency and offered ongoing supervision. Of 116 participants, 111 (95.6%) who were referred to the TIP intervention completed all 3 TIP meetings.

Quiz Ref IDRates of nonadherence were higher among participants in the treatment as usual group throughout the study period. Participants in the TIP group were 5 times more likely to be adherent to their antidepressant at week 6 (odds ratio [OR], 5.54; 95% CI, 2.57 to 11.96; χ²₁ = 19.05; P < .001), and nearly 3 times more likely at week 12 after controlling for study site (OR, 2.84; 95% CI, 1.47 to 5.50; χ²₁ = 9.60; P = .002) compared with participants in the treatment as usual group. A combined measure of 80% adherence at 6 and 12 weeks showed that TIP participants as a group were 3 times more likely to be adherent at both 6 and 12 weeks combined (OR, 3.27; 95% CI, 1.73 to 6.17; χ²₁ = 13.34; P < .001) (Table 2). The TIP effect on the combined adherence measure was also higher than treatment as usual (OR, 2.2; 95% CI, 1.1 to 4.3; χ²₁ = .8; P = .03) within the subgroup of individuals with mild-to-moderate baseline depression (HAM-D score, ≥10).

Quiz Ref IDImprovement in depressive symptoms from baseline changed significantly with time in both TIP and treatment as usual groups after adjusting for study site (time: F_2,377 = 5.05, P = .01). The group rates of symptom improvement for the total sample did not differ significantly over time (time by treatment: F_2,377 = 1.03, P = .36). However, TIP participants showed a significant improvement (24.9%) in depressive symptoms at 6 weeks (95% CI, 13.9 to 35.9; t₃₃₇ = 4.46; adjusted P < .001), whereas treatment as usual participants showed a less robust, nonsignificant improvement (10.7%) (95% CI, −0.01 to 21.4; t₃₃₇ = 1.96; adjusted P = .05). Participants in the TIP group showed additional improvements in depressive symptoms that were not statistically significant over the course of the study. Treatment as usual participants showed a 12% improvement at 12 weeks (95% CI, 2 to 23; t₃₃₇ = 2.32; adjusted P = .04), with no additional reduction in depressive symptoms. In a post hoc analysis of participants with mild-to-moderate depression severity (HAM-D score, ≥10) at the baseline assessment, both TIP and treatment as usual participants improved over the course of the study with no significant differences between groups. However, at 6 weeks, TIP participants had improved by 30.2% from baseline (95% CI, 21.7 to 38.8), whereas among treatment as usual participants, the improvement was less robust (23.8%) (95% CI, 15.5 to 32.1). Improvement in depressive symptoms, measured as percentage change from baseline, did not change significantly as a function of 80% adherent status (at weeks 6 and 12) and treatment arm (time by treatment by adherence interaction: F_2,363 = 2.28, P = .01). However, participants in both groups who were 80% adherent at weeks 6 and 12 had a 15% greater improvement in depressive symptoms over the course of treatment (95% CI, −0.2 to 30; t₃₆₉ = 1.93; P = .051) after controlling for study site.

In this community-based effectiveness trial among adults 55 years and older who had been prescribed pharmacotherapy for depression in primary care, patients who participated in TIP were more likely to be adherent to their medication than participants who received treatment as usual. The 5-fold increase in adherence during the first 6 weeks of care supports the clinical usefulness of TIP to improve early adherence. Higher adherence among TIP participants was further sustained throughout the first 3 months of care. Although TIP did not significantly improve overall depression, the TIP group showed a significant early reduction in depressive symptoms. These findings were replicated in the participants with mild-to-moderate depression severity at baseline.

To our knowledge, few prior interventions have effectively improved adherence among adults with depression, despite its importance in quality care and potential to reduce the personal and societal costs of depressive disorders.¹⁰ Research has documented that rates of antidepressant prescriptions in primary care are increasing, and antidepressant medication adherence is associated with decreased mortality³⁸ and a lower likelihood of suicidal ideation at 1-year follow-up.³⁹ In a 2015 cluster randomized trial, shared decision making was found to be useful in selecting an antidepressant and improving patient satisfaction but was not effective in improving adherence.⁴⁰ Programs focused on increasing patient activation to maximize participation and engagement in care have had limited impact on retention.⁴¹ In 2014, the most recent update to the Cochrane review of adherence research concluded that additional interventions for adherence are needed given its bearing on treatment outcomes.⁴² To our knowledge, TIP is the only brief intervention that targets early adherence specifically by working directly with the patient to reduce barriers, increase perceived benefits, and promote a personalized antidepressant adherence strategy. This patient-centered approach may help to address the informal cost-benefit analysis resulting in meaningful increases in adherence. These findings also support the relationship between early adherence and depression improvement.

In the present study, we did not find a significant impact of TIP on depression compared with treatment as usual over the course of the study. There was an early depression response among the TIP group, but both groups improved over time. There is emerging evidence of the differential impact of antidepressant medications on depression trajectory based on symptom clusters⁴³ as well as the benefit of looking beyond symptom severity to predict treatment response.⁴⁴ Targeted treatments and improved adherence could improve depression outcomes as adequate dosing remains a challenge for depression treatment⁴⁵ especially in primary care, which continues to be the largest provider of mental health services for older adults.⁴⁶ Recent research has documented that depression remains poorly managed in primary care, especially when compared with other chronic diseases.⁴⁷

There are limitations to this study. First, we cannot account for factors outside of the primary care setting that may affect adherence. Second, this study uses a self-report measure of adherence. Although the adherence measure has been validated, is reliable, and does not increase attention to adherence or serve as an intervention in itself, we recognize that different measures (eg, electronic monitoring, pill counts, self-report) yield different information. Individuals who deny nonadherence or are unintentionally nonadherent could have been missed. In addition, by focusing on acknowledged adherence, we could not include adults with cognitive impairments who may have higher rates of unintentional nonadherence. Third, as a new intervention, we did not use an independent rater to cross-validate intervention fidelity. Finally, by focusing on early adherence, we recruited only patients whose physicians diagnosed them as having depression and who could be contacted within 10 days of their prescription to capture early decisions. Although this criterion is stringent for community-based clinical research, it misses those patients with depression who refused to accept a prescription. We chose to exclude patients with treatment for refractory depression from an earlier antidepressant trial as their views of medication, barriers to adherence, and overall adherence strategies may differ. Future implementation of TIP will target all patients in primary care to increase its usefulness.

This 2-site randomized clinical trial evaluated the effectiveness of TIP on early adherence among adults prescribed antidepressant therapy for depression in primary care. The Treatment Initiation and Participation Program uses strategies to identify barriers to adherence and develop a personalized adherence strategy. The TIP intervention was more effective in helping patients achieve adequate adherence during the critical early adherence period. Adequate antidepressant adherence was of 6-month depression outcomes. Given the high rates of depression, nonadherence, and use of primary care as a mental health service, wide-scale implementation of a targeted adherence intervention, such as TIP, could have a meaningful impact on public health. In addition, future work can extend TIP to other patient groups with depression (eg, patients with treatment for refractory depression from an earlier antidepressant trial) or patients with comorbid clinical conditions (eg, cognitive impairment).

Corresponding Author: Jo Anne Sirey, PhD, Department of Psychiatry, Weill Cornell Medical College, 21 Bloomingdale Rd, White Plains, NY 10605 (jsirey@med.cornell.edu).

Accepted for Publication: August 6, 2017.

Published Online: September 27, 2017. doi:10.1001/jamapsychiatry.2017.3047

Author Contributions: Drs Sirey and Kales had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Sirey, Bruce, Turnwald, Blow, Kales.

Acquisition, analysis, or interpretation of data: Sirey, Banerjee, Marino, Bruce, Halkett, Chiang, Liles, Artis, Kales.

Drafting of the manuscript: Sirey, Banerjee, Halkett, Blow, Kales.

Critical revision of the manuscript for important intellectual content: Sirey, Banerjee, Marino, Bruce, Halkett, Turnwald, Chiang, Liles, Artis, Kales.

Statistical analysis: Banerjee, Halkett, Chiang, Liles, Artis.

Obtained funding: Sirey, Kales.

Administrative, technical, or material support: Sirey, Marino, Halkett, Turnwald, Liles, Blow.

Study supervision: Sirey, Banerjee, Marino, Bruce.

Conflict of Interest Disclosures: None reported.

Funding/Support: The project was supported by grants R01 MH087562 (principal investigator: Dr Sirey) and R01 MH087557 (principal investigator: Dr Kales) from the National Institute of Mental Health.

Role of the Funder/Sponsor: The National Institute of Mental Health had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, and approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: We thank the many patients who gave their time to participate in this study. We also thank the staff and physicians at the practices who endorsed the research. No community research can be done without a strong partnership.