OVERVIEW

The convergence of historical trends in the research, education, and clinical practice of psychiatry ¹ over the last century mirror concerns and developments in the more general United States health care system² that call for system redesign to provide safe, personal, cost-effective, high-quality health care. This includes innovative approaches to the psychiatric care of patients in the general medical setting, which remains the locus of care that most patients prefer, and the only available resource for many. Advances in psychopharmacology have greatly facilitated the development of such new models, designed to address not only quality, but also cost-containment and allocation of limited resources. Psychiatric care provided to medically ill patients was primarily hospital-based, but the advent of ever-shorter inpatient stays means that now such care predominantly occurs in the outpatient setting.³ This coincides with the trend for shorter inpatient psychiatric hospitalizations, without a concomitant increase in community mental health resources,⁴ that leaves primary care providers (PCPs) to treat more acute psychiatric illness of patients in their outpatient practices. Psychiatrists must collaborate with their medical colleagues to develop and implement pragmatic, cost-effective, outpatient models of high-quality psychiatric care delivered in the primary care/general medical outpatient venue.

The realization of limited health care resources and rapid escalation of health care expense have forced a change in focus from patient- to population-based care.⁵ Although inherently painful for a society that so highly values the individual, this transition has served to expose the tremendous fiscal burden of psychiatric morbidity. The psychiatrically-disordered population experiences increased physical health care utilization, work absenteeism, unemployment, subjective disability,⁶^–⁸ and mortality rates. Though more difficult to demonstrate, there is also evidence of the cost-offset value of appropriate and timely psychiatric treatment.⁹^–¹¹

Changes in health care reimbursement have resulted in conflicted PCP incentives to recognize and treat psychiatric problems.¹² On the one hand, prepaid, provider-risk plans, such as health maintenance organizations (HMOs) and other capitated programs, have exposed the expensive use of general medical services by patients with untreated or poorly managed psychiatric illness. This statistic serves as an incentive for the PCP to initiate treatment for the more common psychiatric problems seen in primary care. On the other hand, the PCP gate-keeper system, which evolved to manage the expense of specialty care, can be a disincentive to the recognition of more serious mental illness or any mental condition the PCP is not comfortable treating. Managed care organizations (MCOs) often carve out “behavioral health care” to managed behavioral health organizations (MBHOs)¹³ that may have limited referral networks not inclusive of the PCP’s psychiatric colleagues. This situation makes the referral process a time-consuming disincentive and complicates future communication and collaboration between mental health and physical health providers. Limited formularies, which generally vary by plan, often with onerous or time-consuming prior authorization requirements, further complicate and thus deter initiation of treatment.

While many of the MBHOs have spearheaded initiatives to promote treatment of common psychiatric problems in the primary care setting, most do not credential or contract with nonpsychiatric physicians (which essentially shifts the cost of the PCP’s treatment of psychiatric problems to the [medical] MCOs).

EPIDEMIOLOGY

The Epidemiologic Catchment Area Study of the early 1980s was an attempt to quantify the prevalence of psychiatric problems in the general United States population. The study found that during a 6-month interval, roughly 7% of community residents sought help for a mental health problem. More than 60% of these individuals never saw a mental health professional, but sought care in a medical setting, such as the emergency department (ED) or from their PCP.¹⁴ Even among those who met full criteria for a diagnosable psychiatric disorder, 75% were seen only in the general medical (rather than the mental health or psychiatric) setting.¹⁵ Psychiatric distress therefore was exceedingly common among primary care populations. About half of general medical outpatients had some psychiatric symptoms. The use of structured diagnostic interviews detected a prevalence of 25% to 35% for diagnosable psychiatric conditions in this patient population. However, roughly 10% of primary care patients had significant psychiatric symptomatology that did not meet criteria for a recognizable psychiatric disorder.¹⁶ Of the full criteria disorders, the vast majority were mood disorders (80%), with depression being the most prevalent (60%) and anxiety a distant second (20%). The more severe disorders (e.g., psychotic disorders) were more likely to be treated by mental health professionals.¹⁵

Conducted from 1990 to1992, the National Comorbidity Survey (NCS) found a lifetime prevalence of one or more psychiatric disorders in United States adults to be about 50%, with a 30% 1-year prevalence of at least one disorder.¹⁷ Alcohol dependence and major depression were the most common disorders.

In 2001 and 2002, a rigorous replication of the NCS (NCS-R) was undertaken to incorporate measures of severity, clinical significance, overall disability, and role impairment.¹⁸ The NCS-R revealed the risk of major depression to be relatively low until early adolescence, when it begins to rise in a linear fashion. The slope of that line has increased (i.e., it became steeper) for each successive birth cohort since World War II. The lifetime prevalence of significant depression is 16.2%; the 12-month prevalence is 6.6%. Two findings, however, are of particular interest. One is that 55.1% of depressed community respondents seeking care now receive that care in the mental health sector. The other significant finding, attributable to advances in pharmacotherapy and educational efforts, is that 90% of respondents treated for depression in any medical setting now receive medication. While this suggests improvement in the community treatment of depression, it is tempered by the further finding that only 21.6% of patients receive what recent, evidence-based guidelines (American Psychiatric Association [APA], Agency for Healthcare Research and Quality [AHRQ]) would consider minimally adequate treatment (64.3% treated by mental health providers, and 41.3% of those treated by general medical providers). This takes into account that almost half (42.7%) of patients with depression still receive no treatment.¹⁸

Although past studies documented PCPs’ failure to diagnose over half of the full criteria mental disorders present in their patients,^19,²⁰ later studies demonstrate PCPs’ better recognition of more seriously depressed²¹ or anxious²² patients. These studies also demonstrate that higher-functioning, less severely symptomatic primary care patients have relatively good outcomes, even with short courses of relatively low-dose medication. This highlights the diagnostic difficulty in the primary care setting. Primary care patients are different from those who seek specialty care (which is the population in whom most psychiatric research is done). Primary care patients may seek treatment earlier in the course of their illness, since they have an established relationship with their PCP that is not dependent on their having a psychiatric disorder. They frequently present with somatic complaints, rather than psychiatric symptoms. Since the soma is the rightful domain of the PCP, this further obscures the diagnosis. Primary care patients also may have acute psychiatric symptoms that clear relatively quickly, even before medications have had time to reach therapeutic levels, which suggests that such individuals might benefit as much from watchful waiting and from the empathic support of their PCP. There is a high noise-to-signal ratio in psychiatrically distressed primary care patients. That is, as many as one-third of these significantly distressed patients have subsyndromal disorders that do not meet full, DSM-IV criteria for a diagnosable mental disorder. This diagnostic ambiguity in the general medical setting is paralleled by the relatively good outcomes that primary care patients experience on what most psychiatrists would consider subtherapeutic doses and inadequate durations of pharmacotherapy.^19,²³ These factors are cause to reconsider the significance of the PCP’s failure to diagnose. Much of the angst of primary care patients resolves spontaneously, either with resolution of an initiating event, expressed concern of caregivers, or the placebo effect of a few days of medication, and it may be attributable to an adjustment disorder.

Barriers to Treatment

Although necessary, symptom recognition is not sufficient to ensure treatment of psychiatric problems in the primary care setting.²⁴ Even when PCPs are informed of the results of standardized screening tests, they may not initiate treatment. PCP, patient, and system factors collude to inhibit the discussion necessary to promote treatment (“don’t ask/don’t tell”).²⁵

Physician factors, the “don’t ask” part of the equation, include the failure to take a social history or to perform a mental status examination (MSE).²⁶ This behavior has been attributed to deficits in training of medical students and residents,²⁷ time and productivity pressures, and to personal defenses (such as identification, denial,²⁸ or isolation of affect). PCPs are experienced and more comfortable addressing their patients’ physical complaints. Some PCPs fear that their patients will leave their practice if asked about mental health issues. The PCP, like many patients, may not believe treatment will help. Not having a ready response or approach to a problem is a major deterrent to identification of a new problem within the context of a 15-minute primary care visit. Denial or avoidance may prevail when the time-pressured PCP feels unsure of how, or whether, to treat or to refer.

Stigma, which is prevalent among patients and providers, is a major patient deterrent to initiating a discussion of psychiatric symptomatology. Often patients “don’t tell” because of shame or embarrassment. They may believe psychiatric problems are a personal weakness, and they may perceive that their PCP shares this belief. Patients may not know they have a diagnosable or treatable mental disorder.²⁹ These are among the reasons that primary care patients more frequently have physical complaints. Somatic complaints also increase the diagnostic complexity³⁰ since medical disorders may simulate psychiatric disorders, psychiatric disorders may lead to physical symptoms, and psychiatric and medical disorders may co-exist.

System factors include the ever-changing health care finance and reimbursement climate. This encompasses managed care, “carve-outs,” provider risk, capitation, fee-for-service, free care, coding nuances, differential formularies, and prior authorization, all of which promote financial imperatives to contain cost and to increase efficiency. This systemic instability, confusion, and administrative time-creep easily overshadows the impulse to pursue the treatment of a possibly self-limited condition. Mental health carve-outs have either eliminated, or greatly complicated, the possibility of reimbursing PCP treatment of mental disorders. Prepaid plans, such as HMOs, significantly decrease incentives to offer anything “extra.”³¹ The necessity to increase productivity has excessively shortened the “routine visit,” now often less than 15 minutes, while the excessive burden of required documentation further erodes clinically available time. (The implementation in some practices of the electronic medical record [EMR] has standardized and improved screening, documentation, and follow-up,³² but has introduced further time-consumption issues.) The care-promoting advent of new, safer, more tolerable psychotropic medications has been offset by soaring pharmacy costs and by restrictive (and possibly short-sighted³³) formularies. The practice of primary care has reached a crisis point where the pressures are so overwhelming that few PCPs can sustain full-time clinical practice.

THE GOALS OF COLLABORATION

Now that clinically proven and effective treatments exist, access and quality of care remain significant issues, best addressed through the collaboration of psychiatry and primary care. The four major goals of collaboration are to improve access, treatment, outcomes, and communication.

Access

Collaborative care in the primary care setting addresses both physician and patient factors that limit the patient’s access to appropriate assessment and treatment. Most patients are not new to the general medical setting and therefore feel relatively comfortable and unstigmatized in that setting. Conversely, they may believe the mental health clinic is for “crazy people,” and they do not identify with that (perceived) clientele. A specified mental health wing or floor in the primary care setting may hold that same stigma and act as a disincentive for patients to access treatment. Most patients do not know of a psychiatrist or how to access care from one and may not feel certain that they need one. The unaided decision to foray into the mental health arena may be fraught with shame and anxiety, powerful deterrents to making that first call. Calling the PCP’s office and making an appointment for fatigue, sleep problems, weight loss, or palpitations is infinitely less threatening.

An established relationship between the PCP and a trusted, accessible psychiatric consultant removes the onus of recognizing, treating, or referring patients with mental disorders. With this availability of expert opinion and clinical backup, PCPs more readily identify psychiatric distress and initiate treatment for their patients.

Treatment

In the past, PCPs often prescribed insufficient doses of medications (e.g., amitriptyline 25 mg) for major depression.³⁴ Since the advent of safer, well-tolerated medications, such as selective serotonin reuptake inhibitors (SSRIs), the selection of medications by PCPs has improved,^35,³⁶ although the dose chosen often remains suboptimal. Benzodiazepines have been prescribed by PCPs more frequently than any other class of psychotropic medication, even for major depression,³⁷ although they now are appropriately surpassed by antidepressant prescriptions.³⁶ Collaboration with the consultation psychiatrist can improve the choice, dose, and management of psychotropic medications. Collaboration is also helpful when the medication with which the PCP might be most familiar is off-formulary on a given patient’s pharmacy plan. Such a treatment deterrent may instead become an opportunity for brief, pragmatic education.

Outcomes

Several studies have demonstrated better outcomes for seriously depressed primary care patients treated collaboratively by their PCP and a psychiatrist.³⁸^–⁴⁰ Cost-offset, however, is difficult to demonstrate because of the hidden costs of psychiatric disability.^7,^41,⁴² Nonetheless, there is some evidence for a decrease in total health care spending when mental health problems are adequately addressed.¹¹ Even if this were not so, the case for cost-effectiveness could be made.^9,