Clinically significant depression is present in 5–15 % of medical outpatients, and is even more prevalent in those with chronic medical disorders. Depression is among the top health-related reasons for decreased work productivity,1 and by 2020, it is expected to be second only to ischemic heart disease in the number of disability-adjusted life years lost due to disease or injury.2 Moreover, we have brief self−administered tools to screen for depression and monitor response to treatment; effective pharmacological and psychotherapeutic therapies; and evidence-based models of care for improving outcomes. Integration rather than fractionation of depression care is desirable. Depression is too common and too entangled with other medical problems to be treated exclusively outside of primary care. Instead, a collaborative approach is preferable, wherein the optimal site of care and treating clinician is guided by severity and duration of symptoms, patient preferences, psychiatric comorbidity, and clinical response. Comparable depression outcomes can be achieved in primary care and mental health settings, when patients are appropriately managed with monitoring and treatment adjustments.3

The Veterans Health Administration (VHA) has been an excellent health care system in which to implement and evaluate new models of integrating depression care. A single payer providing comprehensive medical and mental health services for eligible beneficiaries, a commitment to the parity of mental with medical disorders, an investment in mental health resources, and a nationwide electronic medical record all reduce some of the barriers that commonly exist in other practice settings. Also, primary care has received extra attention, including routine depression screening (97 % of eligible patients received depression screening in 2010) and the funding of additional mental health resources embedded in primary care. This issue of JGIM includes two reports on the VHA track record of integrating depression care.

Using national VHA data and a retrospective cohort design, Szymanski et al. examined data from a random sample of 36,323 primary care patients with a positive depression screen between October 1, 2009 and September 30, 2010.4 In addition to the large random sample, strengths of this study included selection criteria targeting an inception cohort of patients with newly-diagnosed depression and an elegant strategy to adjust for confounders, as well as three types of sensitivity analyses. The principal aim was to assess whether embedding mental health services in primary care leads to improved initiation of depression treatment. This was in fact confirmed, in that those patients who received same-day mental health services were much more likely to receive a depression diagnosis and initiate psychotherapy and antidepressant medications within 12 weeks than were those who received only primary care services on the screening day. Even when analyses were restricted to those with a depression diagnosis, treatment initiation occurred much more often in those receiving same-day mental health services. Other risk factors for not receiving treatment were older age, male sex, psychiatric comorbidity, and having a service-connected disability rating of ≥ 70 %.

The most important limitation in this study, where patients were not randomized to level of care, is the inability to control for patient acceptance of a depression diagnosis and willingness to be treated. For example, adding a single question about desire for treatment (“Is this something with which you would like help?”) improves both the diagnostic specificity and patient-centeredness of depression screening.5 It is quite possible that a not-insubstantial proportion of patients not receiving same-day mental services had some reluctance in embracing either a diagnosis of depression and/or its treatment.

Chang et al. took a systems-level perspective in documenting VHA integration of depression care.6 In a 2007 cross-sectional survey of leaders from 225 VA primary care (PC) practices, they examined factors related to readiness of PC to implement one of three VA-endorsed models: collocation, Translating Initiatives in Depression into Effective Solutions (TIDES), and Behavioral Health Lab (BHL). Collocation is a simpler arrangement (i.e. having a mental health specialist physically located in PC), whereas TIDES and BHL use standardized assessments and care management based on evidence-based collaborative care principles. Not surprisingly, the simplest solution—collocation—was implemented in a far greater proportion (47.5 %) of clinics compared to TIDES (17.3 %) or BHL (7.6 %).

However, it is not clear that simpler means more effective. More than 40 trials for depression have been conducted, testing various types of collaborative care interventions that have much more in common with TIDES and BHL than with simple collocation.7,8 The cumulative results have conclusively established the superiority of these models over usual care, whereas collocation has been less rigorously evaluated. While Syzmanski et al. provide the first evidence that collocation increases treatment initiation,4 neither the Syzmanski nor Chang studies were designed to examine whether depression outcomes are actually improved. TIDES and BHL use structured assessments, systematic monitoring of depression outcomes, patient self-management tools, and information technology to support depression care managers and mental health professionals, as well as their links to primary care clinicians. Collocated care relies instead on “mental health expertise” for most of these functions. However, many of the specific components in TIDES and BHL have been found to be important elements in improving depression outcomes in collaborative care models, and likely could provide added value to depression care delivered by mental health specialists as well.

Models of integrated depression care have also proven successful in non-VHA systems such as Kaiser, Intermountain HealthCare, and the DIAMOND project in Minnesota. In a 2006 JGIM editorial, Callahan concluded that integrating depression care might come with some added costs.9 However, Unutzer et al. have shown that enhanced depression care may be cost-saving in the long term, at least in certain populations.10 Moreover, the amount that depressed patients are willing to pay for depression treatment is comparable to that reported for the treatment of other chronic medical disorders and higher than the actual cost of depression treatment.11 One key barrier to implementing integrated depression care in many primary care practices is the financial challenge introduced by segregated physical and mental health reimbursement practices.12

Given the substantial evidence that at least some models of integrated care are superior to primary care “as usual” for improving depression outcomes, several steps in implementation warrant consideration:

  1. 1)

    A depression measure should be used not only to support initiation of therapy, but at follow-up to gauge treatment response and adjust therapy. This is no different than adjusting treatment for hypertension and diabetes based upon follow-up blood pressure or blood sugar readings. This can be facilitated by the use of a validated self-report measure, embedding the scores in the medical records, and providing the score to the clinician at the point of care.

  2. 2)

    Measurement-based care should be adopted by both mental health and PC clinicians. For both groups of providers, the type and intensity of treatment should be tailored to optimize a measurable clinical response. Indeed, studies in mental health settings have confirmed the added value of measurement-based depression care.3,13

  3. 3)

    A stepped care approach may be preferable to a model that shifts either too many or too few depressed PC patients to integrated care. Ideally, mental health services would be targeted to those most likely to get added benefits beyond PC-only care. Decisions could be guided by factors present at baseline (e.g., depression severity, psychiatric comorbidity, patient treatment preferences) and/or follow-up (e.g., poor clinical response). Those committed to mental health parity must also be responsible for wisely shepherding resources which, even if enhanced, are not inexhaustible and compete for a limited pool of health care dollars.

  4. 4)

    In-person visits can be complemented and/or reduced through technologies such as telehealth14 and automated depression monitoring,15 which might increase both the cost-effectiveness and patient-centeredness of depression care.

Collectively, the literature convincingly demonstrates that some models of integrated depression care are evidence-based and cost-effective. It is now time to translate and refine what is happening in some systems to primary care practice in general.