Forty percent of the patients enrolled in the Sequenced Treatment Alternatives to Relieve Depression study (STAR*D) are being studied in their primary care settings and managed by their primary care clinicians (1). The clinic setting—whether primary or specialty—seems to have little bearing on the presentation or severity of the patient’s illness, the clinician’s fidelity to the treatment protocol, or the outcome. Both primary care clinicians and psychiatrists are taking their patients through a complex set of treatment steps, in an effort to achieve remission. The participation of primary care clinicians in this trial provokes three questions for primary care clinicians and for psychiatrists: 1) Through how many steps beyond initial recognition and treatment can primary care clinicians effectively and safely manage their patients’ depression? 2) What level of care are primary care clinicians actually delivering in their practice? 3) How should primary care clinicians and psychiatrists work together for better care of depressed patients?
Once the dust settles from STAR*D and we have evidence-based practice guidelines, then primary care clinicians such as myself should be able to safely move several steps further along the management algorithm before consulting a psychiatrist. Today, when most primary care clinicians diagnose depression, they initiate treatment with an antidepressant. Even with conscientious follow-up and aggressive dosage management, the patient may well not achieve remission. The primary care physician then wonders whether to refer the patient for psychotherapy, to switch to another antidepressant (of a similar or different class), perhaps to augment with another antidepressant, or even to augment with T3, lithium, or another medication—all questions addressed by STAR*D. At this point, the primary care physician generally either asks for help or first tries one of these options and then asks for help, if help is available. This is not because the decision is difficult or the management itself is dangerous or complicated; it is just because the evidence on how to proceed is incomplete and disorganized. STAR*D and a few confirmatory studies should make it possible for primary care clinicians themselves to push care decisions to the third or fourth step before seeking a consultation. This would result not in fewer consultations for mental health consultants—it is a good bet that primary care clinicians will seek as much mental health consultation as they can conveniently get—but, rather, in the seeking of consultations for more complicated patients and even more difficult problems.
Most likely this change in primary care practice will not happen, at least not until primary care clinicians secure a few additional resources for their practice operations and adopt a somewhat different model of care for conditions such as depression. For primary care physicians, there is a chasm between what we know and what we do. We have been inundated over the past decade with information, exhortation, and recommendations regarding depression and its management. By now, pretty much everyone in primary care knows that depression is common, disabling, expensive, and treatable. Most clinicians are familiar with the diagnostic criteria and can tell you four or five effective drugs, and their doses, for its treatment. I do not wish to argue that the level of knowledge about this condition is sufficient in primary care, because it is not, but lack of information is not the deepest problem.
As of this writing, hundreds of primary care practices have participated in various depression trials, many of which have demonstrated effectiveness and cost-effectiveness. But the vast majority have been plagued by nonsustainability. Almost without exception, these practices, after demonstrating to themselves and to the world that a particular form of depression treatment works, fail to sustain that improvement. Shortly after the resources used to conduct the study are withdrawn, the practice returns to its old ways. Practices have an implicit but highly rational and sophisticated set of rules about how they will function, how they will use their collective time and resources, and which problems among the overwhelming swarm coming in every day they will attend to. It turns out that new knowledge about how to render better care simply does not “stick” unless one of two conditions is met. Either it must fit into practice patterns as they already are or, if a change is required, it must come with additional resources to effect and sustain it.
It is becoming apparent that primary care practices are structured to handle very well the panoply of common acute problems. Our system of financing primary care, such as it is, supports and maintains this structure. But the most important clinical problems in primary care today are chronic, not acute. Diabetes, heart disease, chronic lung disease, arthritis, and, of course, depression are fundamentally different from acute otitis media, acute cystitis, acute bronchitis, upper respiratory infection, and minor injury, and they require a fundamentally different kind of care. Unfortunately, primary care practices have neither the equipment nor incentives to manage chronic diseases well. Wagner et al. (2) have produced an elegant and detailed formulation showing how chronic diseases can be managed in the primary care setting. The so-called “chronic disease management” model involves the development of a registry of all patients with a given condition, the use of a care manager, the consistent use of an evidence-based care plan or management algorithm, standardized monitoring, and access to an expert consultant for unanticipated problems and complicated patients. With the advent of electronic health records, registries, and monitoring systems, chronic disease management programs are easier to implement. Financing the care managers has been the most difficult part under our current system of care, but it is one for which a number of interesting solutions are emerging (3).
Unfortunately, not all these innovations have penetrated the world of practice. In the July 2006 issue of the Journal, Wang et al. produced a clever comparison from the two National Comorbidity Surveys that illuminates a trend toward more mental health care occurring in the primary setting, with less specialist consultation, even for more severe conditions (4). This is unwelcome news and suggests that our health care system, such as it is, has become less collaborative in regard to mental health problems, despite the aforementioned studies documenting the value of collaborative care. Thus, even as the algorithms become more ramified and complex, and the therapeutic regimens become more effective, the partnerships on which successful regimens are based become more attenuated.
How should primary care clinicians and psychiatrists collaborate on the care of patients with mental health problems? When researchers in the field of mental health services first tumbled to the fact that much of the care rendered to “their” patients was in fact rendered as part of primary care (5), they responded in an interesting and constructive way. They could have promulgated evidence-based recommendations and guidelines for primary care clinicians that were not relevant or feasible for this setting, given the limited resources, competing priorities, and patient preferences. The world of primary care has experienced a lot of this—over the years, primary care clinicians have been flooded with guidelines that make no sense in primary care and have no chance of implementation. Alternatively, mental health professionals could have insisted that all such patients be referred for specialty care, another doomed recommendation that is at odds with patient preferences, clinical efficiencies, and the precepts of comprehensive care. A third way was found instead. These professionals responded by going into primary care settings and working with primary care clinicians, investigators, and patients to translate efficacious treatments into management strategies that would work in primary care. This produced a very interesting partnership and gave birth to a whole new understanding of how primary care practice worked. Twenty years later, as we come to terms with the new Roadmap initiative of the National Institutes of Health, with its emphasis on translation of research findings to community settings, we find that the mental health services researchers are at the forefront of many elegant study innovations and promising models of care. In simply trying to improve the care of depressed patients, mental health services researchers have helped develop such design innovations as the use of heterogeneous study samples with comorbid conditions, usual-care comparison groups, nested designs with group randomization, multistep/multilevel interventions, conduct of studies in primary care settings because patients insist on staying there, and mixed-methods assessments. They have also been quick to adopt the collaborative chronic care model, which, among other things, defines the relationship between the generalist and the specialist in working together for patients with chronic diseases such as depression. These innovations have given rise to effectiveness research as a major pathway to improvements in clinical care, and along the way they have utterly transformed the nature of the randomized clinical trial. The design of STAR*D benefits from this tradition, even as it contributes to it. STAR*D lays a path for an enormous advance in the clinical management of depression in all settings, if it can be matched by corresponding system-level changes, and a way will open for collaborations that permit us to reverse the sorry state of affairs that Wang et al. (4) have documented.