Intended for healthcare professionals

Research

Long term outcomes from the IMPACT randomised trial for depressed elderly patients in primary care

BMJ 2006; 332 doi: https://doi.org/10.1136/bmj.38683.710255.BE (Published 02 February 2006) Cite this as: BMJ 2006;332:259
  1. Enid M Hunkeler, senior research scientist (emh{at}dor.kaiser.org)1,
  2. Wayne Katon, professor of psychiatry2,
  3. Lingqi Tang, statistician3,
  4. John W Williams Jr, professor of medicine4,
  5. Kurt Kroenke, professor of medicine5,
  6. Elizabeth H B Lin, scientific investigator6,
  7. Linda H Harpole, head, global health outcomes7,
  8. Patricia Arean, associate professor8,
  9. Stuart Levine, medical director9,
  10. Lydia M Grypma, internist10,
  11. William A Hargreaves, professor emeritus of psychology11,
  12. Jürgen Unützer, professor of psychiatry12
  1. 1 Kaiser Permanente, Division of Research, 2000 Broadway, 2nd Floor, Oakland, CA 94612, USA
  2. 2 Department of Psychiatry and Behavioral Sciences, University of Washington, 1959 NE Pacific St, Seattle, WA 98195-6560, USA
  3. 3 Health Services Research Center, Neuropsychiatric Institute, University of California at Los Angeles, Los Angeles, CA 90024-6505, USA
  4. 4 Health Services Research and Development, Duke University Medical Center, Hock Plaza, 2424 Erwin Rd, Suite 1105, Durham, NC 27705, USA
  5. 5 Regenstrief Institute, 1050 Wishard Blvd, RG6, Indianapolis, IN 46202, USA
  6. 6 Center for Health Studies, Group Health Cooperative, 1730 Minor Avenue, Suite 1600, Seattle, WA 98101, USA
  7. 7 GlaxoSmithKline, 3030 Cornwallis Rd, MAI.B.530, RTP, NC 27709, USA
  8. 8 Department of Psychiatry, Langley Porter Psychiatric Institute, University of California, San Francisco, 401 Parnassus Avenue, Box ADM-0984, San Francisco, CA 94143-0984, USA
  9. 9 SCAN Health Plan, 3800 Kilroy Airport Way, Long Beach, CA 90801, USA
  10. 10 Kaiser Permanente, Primary Care, 8080 Parkway Drive, La Mesa, CA 91942, USA
  11. 11 University of California, San Francisco and Kaiser Permanente Division of Research, 2000 Broadway, 2nd Floor, Oakland, CA 94612, USA
  12. 12 University of Washington, Department of Psychiatry, Box 356560, Seattle, WA 98195, USA
  1. Correspondence to: E M Hunkeler
  • Accepted 8 November 2005

Abstract

Objective To determine the long term effectiveness of collaborative care management for depression in late life.

Design Two arm, randomised, clinical trial; intervention one year and follow-up two years.

Setting 18 primary care clinics in eight US healthcare organisations.

Patients 1801 primary care patients aged 60 and older with major depression, dysthymia, or both.

Intervention Patients were randomly assigned to a 12 month collaborative care intervention (IMPACT) or usual care for depression. Teams including a depression care manager, primary care doctor, and psychiatrist offered education, behavioural activation, antidepressants, a brief, behaviour based psychotherapy (problem solving treatment), and relapse prevention geared to each patient's needs and preferences.

Main outcome measures Interviewers, blinded to treatment assignment, conducted interviews in person at baseline and by telephone at each subsequent follow up. They measured depression (SCL-20), overall functional impairment and quality of life (SF-12), physical functioning (PCS-12), depression treatment, and satisfaction with care.

Results IMPACT patients fared significantly (P < 0.05) better than controls regarding continuation of antidepressant treatment, depressive symptoms, remission of depression, physical functioning, quality of life, self efficacy, and satisfaction with care at 18 and 24 months. One year after IMPACT resources were withdrawn, a significant difference in SCL-20 scores (0.23, P < 0.0001) favouring IMPACT patients remained.

Conclusions Tailored collaborative care actively engages older adults in treatment for depression and delivers substantial and persistent long term benefits. Benefits include less depression, better physical functioning, and an enhanced quality of life. The IMPACT model may show the way to less depression and healthier lives for older adults.

Introduction

Among the one in 10 older primary care patients with depression,1 2 only a small fraction receives adequate treatment in primary care3 4 or sees a mental health specialist.2 5 6 Although treatment of depression in primary care has improved,713 few improvements deal with the specific needs of elderly patients.

In 2002, one year findings were published from the “Improving Mood Promoting Access to Collaborative Care Treatment” programme (IMPACT), an intervention designed specifically to tackle unmet needs of elderly depressed patients.3 Patients randomised to collaborative care received more antidepressants and counselling, experienced less depression, functioned better, had better lives, and were more satisfied with their care than patients receiving usual care. Bruce et al also reported less depression in older patients receiving collaborative care.14 However, since depression is chronic and recurrent, and many patients experience residual symptoms and relapse, understanding the long term effects of collaborative care is important.

We report the long term (18 month and 24 month) results of the IMPACT trial. Our hypothesis was that, even a year after the intervention ended, intervention patients would experience more enduring health benefits than controls.

Methods

The IMPACT trial was conducted at 18 diverse primary care clinics across the United States. The 1801 study participants were self referred or referred by their doctor, or identified by systematic depression screening.3 15 Patients were aged 60 or older and met criteria for major depression or dysthymia, or both, according to the structured clinical interview for DSM-IV axis I disorders (SCID).16 We excluded patients with current drinking problems, bipolar disorder or psychosis, severe cognitive impairment, acute risk of suicide, or ongoing psychiatric treatment.3 15 After a structured baseline interview, patients were randomised to IMPACT collaborative care or usual care.

Intervention

Acute depression care

IMPACT care3 15 was delivered by a team: a depression care manager (usually a primary care nurse), the patient's primary care doctor, a consulting psychiatrist, and a liaison primary care doctor. For 12 months, IMPACT patients received proactive depression treatment in primary care.

Treatment options included pharmacotherapy, behavioural activation, and problem solving treatment (a brief, behaviour based psychotherapy). Treatment was adjusted according to a stepped care protocol3 15 during weekly reviews by the depression care manager and psychiatrist. Consulting psychiatrists saw about 10% of patients, typically treatment non-responders. During the intervention year, 73% (618) of IMPACT patients used antidepressants, 70% (589) used psychotherapy or specialty mental health care, and 90% (753) received either anti-depressants or psychotherapy.

Relapse prevention

Patients whose depression improved (a decrease of 50% or more in the severity of depression and fewer than two DSM-IV depression symptoms) during the intervention period created a relapse prevention plan with their depression care manager. It stressed continuing care, observing early warning signs (such as changes in sleep or appetite or isolation) and coping strategies such as discussing problems with others, taking medications as prescribed, reducing stress, scheduling positive activities, and contacting their depression care manager or primary care doctor if symptoms re-emerged. The depression care manager and patient then had monthly telephone appointments during the remainder of the 12 month intervention and a final meeting in the 12th month to review the relapse prevention plan. This plan focused on maintaining gains made during the intervention and was shared with the patient's primary care doctor.

Usual care

Usual care patients received notification, along with their primary care doctor, that they met study criteria for depression. They could use all treatments available in usual care (antidepressants, counselling by the doctor, and referral to specialty mental health care). During the initial year, 57% (470) of these patients used antidepressants, 25% (211) reported using psychotherapy or specialty mental health care, and 62% (517) reported using either antidepressants or psychotherapy.

Data collection and outcome measures

Interviewers, blinded to treatment assignment, conducted interviews in person at baseline and by telephone at follow-up. Outcome variables included self-reported use of antidepressants and psychotherapy, satisfaction with care, SCL-20 depression scores, treatment response (≥ 50% decrease from baseline SCL-20 score), remission (SCL-20 score < 0.5), quality of life, overall functional impairment, and physical functioning (physical component score, PCS-12) from the short form 12 (SF-12) health related quality of life measure.3 15

Statistical analyses

We used all randomised subjects who were alive and interviewed at 12, 18 or 24 months to compute intention to treat analyses of unadjusted group differences, with t test or χ2 statistics and 95% confidence intervals. When outcomes are proportions, differences are reported as odds ratios. The clinical effect of IMPACT can be expressed in terms of the number needed to treat (NNT). It indicates how many patients have to be offered IMPACT to achieve a treatment response (50% reduction from baseline in depression on SCL-20) in one more patient than in usual care. For comparison, we report differences between the IMPACT and usual care groups at baseline and at the end of the intervention, as well as six and 12 months after the end of the intervention. Sensitivity analyses that imputed missing data and controlled for covariates (not reported here) produced similar results.1720

Results

Population

The sample (n = 1801) at baseline3 was 65% female (1168 participants), with an average age of 71.2 (SD = 7.5). About 23% (415) were African American, Hispanic, or from other non-white ethnic backgrounds; 79% (1425) had at least a high school diploma. Seventy per cent (1259) had major depression, 53% (953) major depression with dysthymia, and 30% (542) dysthymia alone. Thirty five per cent (683) had mild cognitive impairment on the basis of a six item screening test21 derived from the mini-mental state examination. The mean SCL-20 depression score was 1.7 (SD = 0.6), indicating moderate to severe depression. On average, patients had 3.2 (SD = 1.7) of 11 chronic medical illnesses. During the previous three months, 43% (769) of the sample took antidepressants and 8% (151) received specialty mental health care or psychotherapy. At base-line the groups did not differ significantly on any of these variables.

Depression care

Table 1 shows depression care at baseline, 12, 18 and 24 months. A significantly higher proportion of IMPACT patients reported taking antidepressants at each follow-up (a difference in the percentages of 18 percentage points at 12 months,3 15 at 18 months, and 14 at 24 months, all P < 0.0001). Differences in use of counselling or specialty mental health care observed during the intervention3 disappeared after the first year. Thus, significantly higher rates of depression treatment at 18 and 24 months were accounted for entirely by pharmacotherapy.

Table 1

Depression care by study group at baseline, 12 months, 18 months, and 24 months (6 and 12 months after the end of the intervention). Values are numbers (percentages) of patients unless otherwise indicated

View this table:

Depression and other outcomes

At all three follow-up times, IMPACT patients fared significantly better than controls on every outcome, except overall functional impairment at 24 months (table 2). The greatest differences were at 12 months. IMPACT patients also reported significantly greater confidence in managing their depression (self efficacy) at 24 months (P < 0.0001). IMPACT patients had significantly lower SCL-20 depression scores at each follow up. Even a year after the intervention ended, a significant difference in SCL-20 scores remained (0.23, t = 6.42, P < 0.0001). The NNT is 4 at 12 months, 6 at 18 months, and 9 at 24 months.

Table 2

Depression outcomes by study group at baseline, 18 months, and 24 months. Values are means with standard deviations for the first six outcomes and numbers with percentages for the last three outcomes

View this table:

We found no significant interactions of intervention status with participating healthcare organisation, recruitment method, age, sex, baseline cognitive impairment, depression severity, or several comorbid medical illnesses.

Discussion

IMPACT collaborative care delivered long term improvements in antidepressant use, patient satisfaction, and clinical outcomes. For two years, IMPACT patients were less depressed, functioned better physically, enjoyed a better quality of life, and were more satisfied with their depression care. This was true for both sexes, all age groups, ethnicities, degrees of depression, and physical comorbidities. Although the benefits of IMPACT attenuated at 18 and 24 months, significant health benefits endured even one year after IMPACT resources were withdrawn.

Possible mechanisms of action

IMPACT actively engaged an often reluctant22 population in effective depression treatments. Patients appreciated receiving medical and psychological care in primary care. They benefited from an ongoing, therapeutic relationship with a depression care manager, who followed a stepped care treatment protocol driven by clinical outcomes. Perhaps because of IMPACT's emphasis on relapse prevention, even 12 months after the intervention ended, IMPACT patients reported higher rates of antidepressant use, greater self efficacy in managing their depression, and better depression outcomes than controls. IMPACT's NNT of 4 at 12 months shows a strong clinical effect. The later numbers needed to treat—6 at 18 months, and 9 at 24 months—although higher, show that the clinical effect was substantial and sustained.

Other studies

Two studies showed that IMPACT's success may hinge on its seamless integration into primary care and, perhaps more importantly, the patient's relationship with the depression care manager. The PRISMe study22 showed that older patients are more likely to engage in depression treatment when it is offered in primary care. The nurse telehealth study8 showed the importance of a therapeutic relationship with a skilled and empathetic care manager in primary care. Two trials in mixed aged populations11 12 23 show that collaborative care can reduce depression over the long term, two to five years. IMPACT extends these findings to elderly populations. IMPACT also improved physical functioning and quality of life.

What is already known on this topic

People over age 60 often receive inadequate treatment for depression in primary care

Organised, multifaceted, and tailored depression treatment programmes are promising. IMPACT produced favourable results during the one year intervention

It is not known if these promising results endure

What this study adds

Tailored collaborative care actively engages people over age 60 in depression treatment and delivers important benefits that persist at least one year after the completion of the intervention programme

IMPACT may show the way to less depression and healthier lives for millions

Limitations

Study limitations include self report of antidepressant, psychotherapy, or specialty mental health use and the possibility that IMPACT improved treatment of usual care participants. The latter is because the same primary care doctors treated patients in the intervention and usual care, both groups were repeatedly surveyed about depression and depression care, and patients found at high risk for suicide received additional clinical attention regardless of group assignment.3 Finally, our study design makes it impossible to determine which intervention components led to enduring health benefits.

In the US, the President's New Freedom Commission on Mental Health designated IMPACT to be a model programme.24 Researchers are currently disseminating IMPACT in diverse health care organizations with support from the John A Hartford Foundation.

Footnotes

  • We thank the patients, primary care providers, and staff at the coordinating centre, and at all of the participating study sites for their contributions and support We would like to thank the investigators and the IMPACT study advisory board for their scientific and other contributions. We thank Christopher Langston, senior programme officer from the John A Hartford Foundation for all of his contributions and support throughout the entire study period. We thank Sabine Oishi for her project coordination efforts. We acknowledge Diane Fraser for editorial assistance. We thank Bruce Fireman and Joe Selby for review and comments on the manuscript. We acknowledge Carla Tillman for preparation of the manuscript. We also thank Claudia Cruz and Yolanda Yarbough for their help with preparation of the manuscript.

  • Contributors JU and WJK conceived the study and were the primary designers, although all authors contributed to the design. JU coordinated the study. EMH, WJK, JWW, KK, EHBL, LHH, PA, SL, LMG, and JU each conducted the study at a different participating site. This included provision of the intervention, data collection and extraction. LT, EMH, JU, WJK, and WAH developed the analysis plan, and LT conducted the analyses. EMH and LT wrote the original draft of the article. JU, WJK, KK, WAH, EHBL, PA, JWW, and LHH made substantial contributions to the manuscript. SL and LMG made more minor revisions to the manuscript. All coauthors contributed to and approved the final version of the manuscript. EMH and JU are the guarantors.

  • Funding This study is supported by grants from the John A. Hartford Foundation, the California Healthcare Foundation, the Hogg Foundation, and the Robert Wood Johnson Foundation.

  • Competing interests EMH has received research funding from Eli Lilly and Company, Merck, Wyeth-Ayerst, Pfizer, Solvay, and GlaxoSmithKline. EMH received a consulting fee regarding women's health from Eli Lilly. JWW has received consulting fees from Wyeth-Ayerst, Pfizer, and GlaxoSmithKline for participating in advisory boards that address depression care. LH has served as a consultant to Wyeth-Ayerst and is currently employed by GlaxoSmithKline. KK has received research funding and consulting fees from Pfizer, Eli Lilly, and Wyeth-Ayerst.

  • Ethical approval The study protocol was formulated independently by the investigators and approved by institutional review boards at all participating study sites and the UCLA study coordinating centre.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
View Abstract