Long-Term Clinical Outcomes of Care Management for Chronically Depressed Primary Care Patients: A Report From the Depression in Primary Care Project

Dr. Klinkman and his colleagues are to be commended for conducting one of the multi-faceted Depression in Primary Care (DPC) projects sponsored by the Robert Wood Johnson (RWJ) Foundation, and for being among the few such research groups to analyze and report their findings in meaningful detail. The RWJ-sponsored DPCs were not designed as randomized controlled trials, but rather were intended to capture needed patient and programmatic information of both the quantitative and qualitative types. Klinkman, et al detail the virtues and limitations of this complex research strategy. The reader is then left to weigh the merits and biases of outcome findings generated in this manner. We agree with the authors that sustained clinical remission is produced by applications of the Clinical Care Model and that its outcome is superior to that achieved by the primary care physician’s Usual Care. We are then left with the following two questions: (1) Which elements of Care Management are particularly therapeutic, and which can be omitted without impairing the model’s effectiveness? (2) Approximately half of the patients provided depression care management failed to remit at each of the three follow-up assessments. What interventions do we now design for them?

Competing interests:   None declared

I'd like to thank Dr. Schwartz for raising the important issue of monitroring for suicidality in depressed patients. His comment also allows me to clarify some details of our protocol that did not make it to print.

We were indeed constrained by University of Michigan Health System legal counsel at the outset of the study from including the ninth item on the PHQ on our outcome monitoring package. This was decided on the grounds that "it is possible that a positive answer on [the ninth item] might not be immediately addressed if the patient completes the PHQ via IVR call", although an automated protocol was in place to notify care managers if this occurred. The decision was made over the objections of all study personnel, and our own arguments were not unlike those submitted by Dr. Schwartz. Later in the study we were allowed to add the ninth item back in to our outcome monitoring protocol and did so.

We regret that we did not have sufficient space in the manuscript to describe all the details of the care process. Here they are. Clinically, suicidality for program enrollees was monitored carefully. Care managers assessed for suicidal thoughts during calls, although this was not standardized as it is with the PHQ-9, and care managers were trained to notify referrring clinicians immediately for concerns of suicidality. These occurrences were tracked. This process closely followed the STAR*D protocol for suicidality assessment and notification, and was approved by our institution's IRB. Primary care clinicians also carried out suicidality assessment during routine care as they would for all their depressed patients. One of the educational interventions delivered to study clinicians was a brief primer on suicidal assessment, distributed through the M-DOCC newsletter.

In the end we did not include the ninth PHQ item in our analyses, as it was available for only some enrollees. Over the full term of the study, we identified only 2 occasions on which the care manager needed to immediately contact the referring clinician for suicidal concern, although there were many occasions on which care managers contacted referring clinicians with concerns about increasing severity of depressive symptoms.

I agree with Dr. Schwartz that clinicians - and care support programs - need to monitor for suicidal ideation, Our experience with our own legal experts highlights the fact that medicolegal constraints can work at cross -purposes to good patient care in this area. I hope that era is behind us.

Competing interests:   None declared

Thanks to Drs Korsen and Bogner for their comments on our study. The common thread in both comments is the potential for this type of depression care support to extend to other chronic health problems, to address multimorbidity, and to eventually be a core part of the patient- centered medical home.

I agree completely. We developed the MDOCC program to test this model of flexible, efficient care support for a single identified condition, but our plan has always been to extend the model to cover other chronic problems once we demonstrated effectiveness and achieved sustainability. We think we have shown effectiveness, but sustainability has been a real challenge. It has proven exceedingly difficult in our setting to find operational support for a program that in essence is paid for by practices while benefits accrue to insurers and employers. Despite pay-for-performance contracting and growing interest in PCMH in Michigan, the parties have until now not been in alignment.

We are now seeing things open up. We are beginning two demonstrations of a more comprehensive care support program, one in our academic health system and one in a nearby community. Both are supported by a coalition of insurers and health systems, and the community program is seen as a cornerstone of a more comprehensive approach to build a community-based accountable care organization. So, for us at least, there may be a tipping-point coming soon. Our challenge in these demonstrations - and I believe the next real challenge for the field - will be to learn the balance between disease-specific care support (still useful for some patients) and more general care support, to learn how to integrate both into the workflow of primary care practices, and above all, to keep this as simple as possible for both patients and providers.

Competing interests:   None declared

I was appalled to see this depression study forgo suicide screening, gutting the PHQ9 and concocting their own instrument, "the PHQ8... omitting the ninth (self harm) item, which we were legally constrained from asking early in the study," without clearly stating that they screened or monitored for suicidal ideation in any other way.

I'd rather not screen for suicidal ideation either. In fact, the entire field of psychiatry shares this universal (countertransference) as clearly indicated by making suicide the 9th and last DSM criteria, implicitly colluding with clinicians who are "too busy" to ask all nine questions. Furthermore, I am sure our risk managers and legal departments would like us to stay as far away from potential mortality as possible.

Protecting and helping our patients is, however, our mission. Studies have shown that as many as 15% of patients with severe depression lasting at least 1 month eventually die by suicide. They have also shown us that a remarkable percentage of patients who kill themselves come to see physicians in the preceding month (Murphy Ann Int Med 1975, Hirschfeld NEJM 1997, Goldman Psychiatry for Primary Care Physicians 1998). And studies have allayed our irrational concern that we will put the idea of suicide in patients’ heads by asking, including our impressionable adolescents (Gould JAMA 2005).

I cannot imagine that the authors failed to screen and monitor for suicidal ideation, but they certainly left their audience with that impression.

Charles E. Schwartz, MD Associate Professor, Psychiatry, Medicine, Family Medicine Albert Einstein College of Medicine Bronx, NY

Competing interests:   None declared

Dr. Klinkman and his colleagues make an important contribution to what is known about care management for depression by demonstrating that a low intensity care management intervention offered to an unselected population of patients with depression, many of whom have chronic depression and comorbid medical or mental health conditions, can lead to clinical improvement.

One interesting aspect of this study is that it reinforces the fact that people with complicated, chronic depression are being seen and treated in primary care. As primary care moves towards a team-based medical home model, it is important to include mental health clinicians and care managers as members of these teams to ensure that primary care practices have the resources needed to care for this population.

Neil Korsen, MD, MSc

Competing interests:   None declared