The Fate of Integrated Treatment: Whatever Happened to the Biopsychosocial Psychiatrist?

Almost all psychiatrists, and certainly those who are leaders in our field, endorse the notion that psychiatrists are distinct from all other mental health professionals in that their training and expertise allow them to be the ultimate integrators of the biological and psychosocial perspectives underlying diagnostic understanding and treatment. However, the biopsychosocial model made famous by Engel (1) has been relegated to political lip service in our managed care era.

The decline of the biopsychosocial model is perhaps most dramatically illustrated in the waning of emphasis on integrating pharmacotherapy and psychotherapy in the practice of psychiatrists. The irony is striking in this regard, since one survey of practitioners suggests that 55% of patients currently receive both medication and psychotherapy (2), either from one clinician or two, thus making it the treatment that the majority of psychiatric patients receive. In fact, in the most recent Practice Research Network survey (3), only one-third of psychiatric patients did not receive some type of psychotherapy. Yet, until the current edition of the Comprehensive Textbook of Psychiatry(4), this leading textbook of psychiatry did not even have a chapter on combined treatment. The integration of medication and psychotherapy is not systematically taught in most residencies. In her anthropological study of American psychiatry, Luhrmann (5) noted how the juggernaut of managed care has intensified ideological tension between psychosocial and biomedical psychiatry: “These approaches are presented in different lectures, taught by different teachers, associated with different patients, learned in different settings. The new policies have sharply enhanced that separation and severely truncated the psychotherapeutic side” (p. 247). Those residents who learn integrated treatment often have to “bootleg” teaching about the approach from a clinical supervisor. Indeed, a recent informal survey of residency training directors (unpublished 2000 study of J. Kay) revealed that very few programs provide any kind of formal didactic instruction on this subject. Why should we expect trainees to integrate psychotherapy and pharmacotherapy if they are taught as two isolated treatment modalities?

Prominent voices in American psychiatry (6–8) have proposed new models for the field that explicitly call for a dramatic reduction in the provision of psychotherapy by psychiatrists. Under these proposals, psychiatry would be reshaped in a biologically reductionistic direction. Detre and McDonald (6) assert: “Even in those instances in which psychotherapy shows some incremental benefits, given the costs, the following question is increasingly asked: ‘Does the difference really make a difference?’ ” (p. 203). Our answer is a resounding “Yes!”—a response shared by many satisfied patients and by a growing body of psychotherapy research (9–16). Indeed, in many cases the cost is well worth it because the provision of psychotherapy may save money by reducing other costs, such as hospitalization (17, 18).

In this communication we wish to emphasize that combined treatment in many respects is the essence of psychiatric practice and the most obvious exemplar of the biopsychosocial foundation on which treatment decisions are ideally based. Like surgeons, who are trained to know when they should not operate, psychiatrists are trained to know when they should not prescribe, and their knowledge allows them to think about patients from the dual perspective of both biology and psychology in all clinical encounters. We are concerned that the tail may be wagging the dog in our current era because economic factors appear to determine the mode of psychiatric practice. We are suggesting that the economically driven practice of automatically dividing treatment between a psychiatrist-prescriber and a nonpsychiatric psychotherapist may not always be in the best interests of patient care or even cost-effectiveness.

Are Two Treatments Better Than One?

A growing literature supports the notion that patients with a variety of conditions may do better with a combination of psychotherapy and medication. Research comparing combined treatment to either psychotherapy or pharmacotherapy alone is obviously more complex than studying a single treatment approach. Moreover, during the 1950s, 1960s, and 1970s, there was considerable concern that psychotherapy and medication might be incompatible. Medication, the argument went, might interfere with the patient’s capacity to make use of psychotherapy. Those concerns have largely disappeared from the current scene in which most patients are receiving combined treatment (although not always from one psychiatrist).

By 1993, sufficient research had been done that Luborsky and colleagues were able to conduct a meta-analysis of 13 studies of psychodynamic therapy versus other forms of therapy (19). They found that while nonsignificant differences in outcome were true of most of the comparisons, the advantage for combined treatments was striking. Psychotherapy plus pharmacotherapy was clearly superior to psychotherapy or pharmacotherapy alone. While there is still a relative paucity of rigorously controlled trials examining the efficacy of combined treatment versus single treatments, a brief overview of the research that has thus far been conducted on specific diagnostic entities reveals an emerging trend in favor of combined treatment. For persons with schizophrenia, the combination of psychoeducational family therapy and antipsychotic medication produces dramatic improvements in the relapse rate compared to either modality alone (20, 21). A disorder-specific individual treatment, personal therapy, in combination with antipsychotic medication, was superior to both family therapy and individual supportive therapy (also combined with medication) in the treatment of a cohort of 151 patients (22, 23) but only among those patients living with their families. Medication compliance was also significantly improved in the group who received personal therapy.

Combined treatment may be most thoroughly studied in persons with major depression. Keller et al. (24) conducted a randomized, controlled trial that compared an antidepressant (nefazodone), a form of psychotherapy (the cognitive-behavioral-analysis system), and the combination of the two; 519 subjects completed the protocol. The response rates among those who received psychotherapy alone and those who received nefazodone alone were 52% and 55%, respectively, whereas the combined treatment group had a response rate of 85%. The patients in this study suffered from chronic nonpsychotic major depressive disorder, and it may be that combined treatment holds a distinct advantage particularly for patients with chronic forms of depression.

In a 1999 study (25), 107 elderly depressed patients were randomly assigned to treatment with 1) nortriptyline, 2) placebo, 3) monthly maintenance interpersonal therapy and nortriptyline, or 4) monthly maintenance interpersonal therapy with placebo. Over the 3-year period of the study, the recurrence rate was only 20% among those patients who received combined nortriptyline and interpersonal therapy, whereas 43% of the group that received medication alone and 64% of the group that received interpersonal therapy and placebo experienced recurrences. A 1997 “mega-analysis” by Thase et al. (26) compared nonbipolar depressed patients treated with psychotherapy alone to those receiving combined treatment and found that the advantage of the combination was particularly striking in those patients with more severe, recurrent depression.

Some data suggest that psychotherapy and medication may work on different target symptoms and at different rates. In a summary of the literature, Hollon and Fawcett (27) concluded, “Pharmacotherapy appears to provide rapid, reliable relief from acute distress, and psychotherapy appears to provide broad and enduring change, with combined treatment retaining the specific benefits of each” (p. 1232). Also, patients receiving either interpersonal therapy or cognitive behavior therapy report significantly greater capacity to establish and maintain interpersonal relationships and to recognize and understand sources of their depression than patients receiving antidepressant medication or placebo (28). In at least one study of dysthymic patients (29), group cognitive behavior therapy actually augmented the effects of sertraline with respect to certain functional changes. Hence, medication and psychotherapy together may enhance the breadth, the duration, or the magnitude of the treatment response.

Recent research also has shown an advantage of combined treatment for other disorders. Combining cognitive behavior therapy and imipramine for the treatment of panic disorder shows clear advantages at long-term follow-up over either treatment alone (30). While the combination of intensive behavioral treatment with methylphenidate does not necessarily yield significantly greater benefits compared with methylphenidate alone for the core symptoms of attention deficit hyperactivity disorder (ADHD), it provides modest advantages for positive functioning outcomes and for non-ADHD symptoms (31). The combination of brief dynamic psychotherapy and clomipramine improved the relapse rate of patients with panic disorder relative to treatment with clomipramine alone (32). A combined treatment regimen of behavior therapy and imipramine for specific phobia and social phobia showed superior efficacy than either treatment alone (33). In a randomized, controlled trial of 120 women with bulimia nervosa, patients who received both cognitive behavior therapy and an antidepressant experienced greater reduction in binge eating and depression symptoms than did patients who received placebo and psychological treatment or an antidepressant alone (34).

While a survey of the literature suggests that two treatments are often better than one, it also helps answer a more specific question: For whom is combined treatment better? Not all patients may require both modalities. Mild to moderate nonchronic forms of major depressive disorder may respond well to either medication or psychotherapy. In some cases, adding psychotherapy may even detract from treatment outcome (35). For example, in the Hogarty study of personal therapy (22, 23), patients with schizophrenia who lived alone had a higher relapse rate when psychotherapy was added. This literature also suggests that some treatments work synergistically when combined, whereas others may not. The combination of cognitive behavior therapy plus an antidepressant improves the outcome of panic disorder better than either modality alone, but combining cognitive behavior therapy with a benzodiazepine confers no advantage (33). Benzodiazepines may actually interfere with the long-term gains derived from exposure.

The results of clinical trials, however, may not easily generalize to naturalistic settings. Most of the studies cited in this brief survey involve types of psychotherapy for which few psychiatrists have adequate training. Moreover, they also typically use different practitioners for the delivery of psychotherapy and pharmacotherapy. The Residency Review Committee for Psychiatry has now mandated training to a level of competency in five different forms of psychotherapy, including combined psychotherapy and pharmacotherapy. Thus, training programs are beginning to address this deficiency.

Not all disorders have been subject to rigorous trials with combined treatment versus either modality alone. However, emphasis on studies of the combination of formal psychotherapy and pharmacotherapy together may overshadow the more mundane approaches often referred to as “psychotherapeutic management.” Indeed, psychotherapeutic principles should optimally be employed in every interaction with the patient, including the 15-minute medication management format. Successful treatment depends on a solid therapeutic alliance, and psychotherapy techniques enhance the formation of that alliance. The role of the therapeutic alliance in the outcome of 225 depressed outpatients who had enrolled in the Treatment of Depression Collaborative Research Program sponsored by the National Institute of Mental Health (36) was rigorously studied (37). The patients were placed in one of four treatments: imipramine plus clinical management, placebo plus clinical management, 16 weeks of cognitive behavior therapy, or 16 weeks of interpersonal therapy. Using an instrument designed to measure patient-therapist behaviors and videotapes of the interactions, trained raters scored the therapeutic alliance for patients in all four groups. Standard depression rating instruments were also used. The researchers concluded that the therapeutic alliance is just as important for pharmacotherapy as for psychotherapy. In fact, the strength of the therapeutic alliance accounted for more of the variance in treatment outcomes (21%) than the treatment method itself (1%).

Adherence to prescribed medication is a major issue for at least 50% of psychiatric patients (38–40). Many problems with noncompliance to medication regimens can be traced back to the failure to attend to the psychotherapeutic issues, such as resistance, transference, or core beliefs and assumptions about medication. Medication may be imbued with a myriad of meanings, and the psychiatrist who combines treatments may be able to identify and understand those meanings in the course of psychotherapy. Even in nonpsychiatric medical settings, attunement to transference and the therapeutic relationship may be critical in assuring compliance. Ciechanowski et al. (41) applied attachment theory to the problem of poor adherence to diabetic self-management regimens. A dismissing attachment style was associated with higher levels of glycosylated hemoglobin. Moreover, dismissing patients who reported poor communication with their healthcare provider had still higher glycosylated hemoglobin levels than those who reported good communication. In other words, the attachment style forged in early childhood reappears as a particular type of transference to treaters that may require specific communication efforts to establish a collaborative therapeutic alliance that assures adherence (17).

Few studies have provided rigorous data on the effect of psychotherapeutic interventions on adherence to medication. One was the aforementioned study of personal therapy in patients with schizophrenia (22, 23). Another study involved 28 newly admitted outpatients with bipolar disorder (42). They were randomly assigned to standard medication clinic care or medication plus 6 weeks of cognitive behavior therapy. Significantly enhanced compliance was seen at a 6-month follow-up evaluation in those who received cognitive behavior therapy. Cognitive behavior approaches have also been shown to enhance medication compliance in persons with schizophrenia (43). Clarkin et al. (44) studied the relative benefit of adding a structured psychoeducational intervention for married patients with bipolar disorder and their spouses. Patients were randomly assigned to receive medication management alone or medication management plus a marital intervention over an 11-month period. Although symptom levels were not affected when the marital intervention was added, improved medication adherence and significant incremental gains in overall patient functioning were noted with combined treatment. More research is clearly needed.

Are Two Treaters Better Than One?

The one-person treatment model, which involves a psychiatrist who conducts the psychotherapy as well as prescribes medication for the same patient, has many advantages. However, the two-person treatment model, which involves a psychiatrist as a prescriber and another clinician as the psychotherapist, is increasingly becoming the standard. Managed care companies may argue on economic grounds that it is cheaper for a psychiatrist to see the patient for a 15-minute medication management appointment three or four times a year while psychotherapy is relegated to a less expensive, and often less well-trained, nonmedical therapist. Whether this arrangement is truly more cost-effective than the one-person treatment model has not been rigorously studied. Preliminary, although nondefinitive, data suggest that there may be little financial advantage (45, 46).

Beyond issues of cost-effectiveness, however, there is a conceptual price to pay for dividing treatment between a psychiatrist-prescriber and a nonpsychiatrist psychotherapist. Such an arrangement often has the symbolic meaning to all parties of a tacit endorsement of Cartesian dualism that potentially fragments the patient into a “brain” and a “mind.” By contrast, the one-person treatment model implicitly endorses an integration of mind and brain in both the psychiatrist’s and the patient’s perspective (47). The dichotomization of mind and brain has long been associated with a view in psychiatry that psychotherapy is a treatment for “psychologically based” disorders, while “brain-based” disorders should be treated with medication (48). What we commonly refer to as “mind” can be understood as the activity of the brain (49). Psychotherapy must work by its impact on the brain. Kandel (50, 51) has elaborated on how these processes might work at the synaptic and intracellular levels. Psychotherapy can be viewed as one example of how experience with the environment can alter gene expression. He suggests that psychotherapy is a form of learning that produces long-term changes in behavior as a result of altering the strength and anatomical patterns of neuronal connectivity in response to the influence of gene expression.

To say that mind is dependent on brain, of course, is not to say that mental states are easily reducible to neural states. As contemporary mind/body philosophers have stressed (52, 53), the irreducible subjectivity of consciousness defies description in nonmental terms. It is within the scope of a materialistic view to note that the construct of mind has a language of its own because it is introspection based and therefore a “first-person” entity. By contrast, the brain is perception based. It is observable from an outside perspective and is therefore a “third-person” entity. Hence, the language of psychology and the language of biology involve two different levels of discourse when working with a patient (54, 55). The biopsychosocial psychiatrist must be conceptually bilingual.

The one-person treatment model demands that the psychiatrist must think both in terms of a dysfunctional brain and a psychologically distressed human being. Docherty et al. (56) have termed this dual role “bimodal relatedness,” akin to the physicist who must simultaneously think in terms of particles and waves. In one appointment with the patient, the psychiatrist must be capable of shifting from a more or less objective and observational perspective to an empathic, intersubjective (but no less scientific) approach. While this balancing act is challenging, it is also the essence of good medical and psychiatric practice and epitomizes Engel’s biopsychosocial model. The psychiatrist, like any other good physician, treats the whole person.

Many psychiatrists find this integration daunting. It is easy to scapegoat managed care policy as the sole culprit responsible for the demise of integrated treatment. However, biological reductionism may appeal to all of us when immersing ourselves in human suffering is too much to bear. An exclusive focus on dosage adjustment and side effects may provide the psychiatrist with a buffer against painful empathic awareness of the patient’s despair as well as offering an illusion of mastery over the complexities of psychiatric illness.

From a cognitive therapy standpoint, the one-person treatment model has the advantage of allowing close scrutiny of the reasons for noncompliance with medication, which may involve automatic thoughts, general dysfunctional beliefs, or life-long negative, global, rigid, dysfunctional core beliefs held by patients about themselves, their world, and other people (57). The automatic thoughts may be about medication, about the physician prescribing, about the illness, or about oneself or others. Typical dysfunctional beliefs about medication may include that medications are only for “crazy” people or should only be considered as a last resort. Dysfunctional beliefs about oneself and others, about illness in general, and about physicians may also be involved. Basco and Rush (58) suggest that psychiatrists should always assume that obstacles to compliance exist and discuss potential problems and their solutions before they even arise. They have outlined a systematic cognitive behavior approach to maximize treatment adherence in patients with bipolar disorder. They emphasize the need for the establishment of a long-term therapeutic relationship based on trust. Within this context, common obstacles to treatment adherence can be identified. These include cognitive variables, treatment variables, social system variables, and variables involving intrapersonal and interpersonal factors.

From a psychodynamic perspective, the one-person treatment model has the advantage of facilitating a comprehensive view of the patient’s transferences. In other words, transferences that develop to medications and transferences to the psychotherapist can all be dealt with in the framework of one relationship. Patients with borderline personality disorder, for example, may view medication as a transitional object that substitutes for the therapist during weekends or vacation periods (59). Medications may also be viewed as agents of malevolent control or poisonous toxins. A psychiatrist who is immersed in a psychotherapeutic process with the patient may have a better sense of the origins of these transferences to the medications, since many of them reflect patterns of internal object relations that are apparent in the transferences to the therapist (60).

When two clinicians are involved in the treatment, the psychiatrist prescribing the medication is inevitably influenced by the administrative arrangement so that he or she delegates psychotherapeutic understanding to the psychotherapist. Hence, medication easily becomes “split off” from the psychotherapy such that it is artificially conceptualized as outside the sphere of psychodynamic understanding (61, 62). This form of compartmentalization frequently colludes with the patient’s tendency to keep the two processes entirely separate. For many patients with borderline personality disorder, for example, it reinforces a fragmented view of the world often characterized as splitting.

The one-person treatment model also allows much more time in the psychiatrist-patient relationship so that a strong therapeutic alliance can be forged. As previously noted, this alliance itself may be of far greater importance than the impact of a particular treatment modality. The patient’s trust in one primary clinician may facilitate opening up to the psychiatrist about medication concerns that are embarrassing. Sexual side effects of medications, for example, may not be disclosed if the patient only sees the prescribing physician once every 6 months. On the other hand, weekly psychotherapy visits facilitate a more trusting relationship in which even the most embarrassing issues can be forthrightly discussed. Some patients who might simply stop taking medication because of the side effects or specific meanings attributed to the medication may instead choose to bring up their reservations so they can be understood and managed, leading to the potential for greater compliance with the pharmacotherapy regimen.

In a similar vein, a psychiatrist in the one-person treatment model engages the patient’s defensive strategies in considerable depth. The same defenses that appear to deal with painful affect states in the therapy may be employed to deal with the stress about the medication that has been prescribed.

In addition to the advantages previously listed, a one-person treatment model avoids some of the structural problems that occur with split treatment in managed care settings. Under these arrangements, the prescribing psychiatrist and the nonpsychiatrist psychotherapist are often thrown together in what Meyer and Simon (61) have called a “clinical shotgun wedding” (p. 244). The two clinicians may not even know each other, let alone harbor mutual respect. Moreover, because communication between the two of them is rarely reimbursed by third parties, there may be little or no contact, resulting in fragmentation of the treatment (61, 62). Patients and clinicians alike may be confused about who is in charge of what, and the ambiguity may lead to errors or failure to take appropriate responsibility in crises.

The particulars of the two-person treatment model carry with it a specific set of liability risks outlined by MacBeth (63). Seeing a greater number of patients with less frequency carries a greater statistical risk of malpractice suit than a practice involving frequent contact with a smaller number of patients. In addition, some managed care arrangements restrict the psychiatrist’s access to the patient by limiting the number or frequency of visits. As a result, the psychiatrist may not be aware of early warning signs that a patient is becoming suicidal, developing a psychotic decompensation, or experiencing significant medication side effects. Third, patients tend to be more reluctant to sue psychiatrists if an effective therapeutic alliance has been established through frequent contact. When there are only a limited number of personal contacts, there is little opportunity to form a strong therapeutic alliance, increasing the risk that lawsuits may result. Finally, the psychiatrist is often assumed to have overall primary responsibility for the patient and all treatment decisions, even though a poorly trained professional providing psychotherapy may have made a major decision about the patient’s treatment without consulting the psychiatrist.

Future Directions

Psychiatric patients can certainly experience substantial improvements under either a one-person or two-person model of mental health care delivery. The latter may be necessary, for example, in areas that have few or no psychiatrists, thereby placing the primary care physician in the role of medication management. When there is good communication between the prescriber and the psychotherapist, the opportunity for mutual consultation may benefit all parties involved. Systematic research that measures the relative effectiveness of a one-person versus a two-person model of treatment is sorely needed in the field. This lack of knowledge base may, in part, be responsible for the fact that few psychiatric residency training programs systematically teach the indications and contraindications for the one-person model of combined treatment and its application in the clinical setting.

Preliminary efforts have been made to test the assumption within the managed care industry that dividing treatment between a psychiatrist and a nonpsychiatric mental health professional is more cost-effective. Goldman et al. (46) used a quasi-experimental retrospective design to compare claims data from a national managed mental health care organization for 191 patients who received integrated treatment for depression and 1,326 patients who received split treatment. Over an 18-month period, those patients who received integrated treatment had significantly lower treatment costs because fewer outpatient sessions were necessary. The authors stated that their findings did not support the assumption that divided treatment is superior to integrated treatment in cost savings. We must be tentative in drawing conclusions from this study, however, since randomization was not employed, the diagnosis of depression was not validated, and outcomes were not measured. Differences in the characteristics of providers could also have influenced the results.

Another study (45) looked at fee schedules of seven large managed care organizations from 1998 and compared psychotherapy alone, medication alone, and combined treatment provided by a psychiatrist working with a social worker or psychologist. When treatment required both medication and psychotherapy, combined treatment by a psychiatrist cost about the same or less than split treatment with a social worker psychotherapist and was generally less expensive than split treatment with a psychologist doing the therapy. This modest study is also suggestive but not definitive. Nevertheless, both studies point the way to further efforts to gather empirical data that would guide practitioners in assessing which patients will do better with integrated treatment by one psychiatrist and for whom such an approach is more cost-effective.

The type of research required would take a large population base and a mental health care delivery system willing to take some degree of financial risk. Patients could be randomly assigned to a psychiatrist trained in both psychotherapy and pharmacotherapy or to a split-treatment situation involving a psychiatrist and a psychotherapist without psychiatric training. Measures of outcome, length of treatment, and simple measures of cost-effectiveness could be included. For example, the number of hospitalizations, emergency room visits, and appointments with other medical practitioners could be incorporated into a cost-benefit analysis. Only this type of controlled research can put to rest the many questions raised by the managed care industry regarding the optimal treatment of patients and the relative costs of different approaches.

In the absence of systematic data to guide the clinician, we must rely on clinical experience to determine when there might be an advantage for a well-trained psychiatrist to conduct both treatments. These might include such situations as 1) patients with schizophrenia or schizoaffective disorder who are not compliant with prescribed medication; 2) patients with bipolar I disorder who deny illness and who do not cooperate with the treatment plan; 3) patients with serious or unstable medical conditions, when the psychiatrist’s medical knowledge is important in the overall management; 4) patients with severe borderline personality disorder who use splitting in a way that disrupts team treatment; 5) severely suicidal patients who are highly impulsive and may require hospitalization in the course of outpatient treatment; 6) patients with serious eating disorders who present complicated medical management problems; and 7) patients who have an ambiguous clinical picture in which the need for medication is not entirely clear and an ongoing assessment of the relative value of medication is needed as part of a comprehensive treatment plan (47).

While these clinical situations are speculative at this point, they can certainly serve as a starting point for the teaching of integrated treatment to psychiatric residents. Residents should be taught that the biopsychosocial model is crucial in every clinical setting, including hospital treatment, consultation-liaison work, medication management, and managed care. Moreover, they should receive instruction about the patient who is fearful of taking medication, the use of psychotherapy to institute pharmacotherapy in the reluctant patient, and the addition of psychotherapy to medical management.

The educational challenges, as always, will raise many scientific questions that will one day be answered by more systematic research (64). When should psychotherapy precede medication? When should it be added to preexisting pharmacotherapy? Are some forms of psychotherapy more effective for understanding compliance problems? Is it possible that some forms of psychotherapy work better with specific medications than others? What are the limits of brief integrated treatment? Is the one-person or two-person treatment model more cost-effective in the long run? Does the fragmentation in the two-person treatment model lead to adverse outcomes more often than the one-person model? Are there different outcomes in treatment if the meaning of medications is addressed versus ignored?

Our knowledge base limits our efforts to answer all these questions at this time. However, growing data on the neurobiological basis of psychotherapy (49, 65–70) provide glimpses of a new frontier in psychiatry. Indeed, the interface of mind and brain, psychology and biology, and pharmacotherapy and psychotherapy is the cutting edge of neuroscience in our new millennium. As Kay (70) has stated, the unique capacity to conduct integrated treatment provides the clearest distinction between psychiatrists and other physicians on the one hand and psychiatrists and other mental health professionals on the other. It would be a terrible loss to the uniqueness of our profession if our skills in psychotherapy fade because of disuse atrophy. Near the end of Luhrmann’s penetrating analysis of what is wrong in American psychiatry (5), she makes the following observation:

It is a sad commentary that a social scientist outside of psychiatry needs to remind us of what we used to take for granted.

Received April 18, 2001; revision received July 26, 2001; accepted Aug. 8, 2001. From the Department of Psychiatry, Baylor College of Medicine; and the Department of Psychiatry, Wright State University School of Medicine, Dayton, Ohio. Address reprint requests to Dr. Gabbard, Department of Psychiatry, Baylor College of Medicine, One Baylor Plaza (MS 350), Houston, TX 77030; [email protected] (e-mail).