Integration of Care: Integrated Treatment of Schizophrenia

The need for integrated treatment is supported by current theories of the pathophysiology of schizophrenia, which are summarized in Figure 1. The widely held stress-diathesis model (11) proposes that symptoms arise from a combination of internal and external factors. Persons who are at risk of developing schizophrenia carry an underlying biologic vulnerability, possibly related to genetic factors or early insult, such as prenatal infections or obstetric complications. Symptoms arise from the interaction of this vulnerability with environmental influences, such as situational stressors and substance abuse. The fact that the onset of more severe psychopathology usually occurs during adolescence implies that maturational changes may also facilitate the development of symptoms.

The course is most often chronic, characterized by intermittent relapses of acute psychotic symptoms in a context of ongoing functional impairment of widely varying severity (12). Research associating long-term outcomes with duration of untreated psychosis or number of relapses has raised interest in the possibility that early intervention—perhaps even in the prodromal phase—may be a way of improving long-term outcomes. It has also discouraged the use of targeted or intermittent treatment, because concern has grown about potential long-term adverse effects associated with relapses.

The impact of multiple factors on symptoms also opens the door to multiple ways of positively affecting the course of the illness. Underlying biologic vulnerabilities related to abnormalities in the function of neurotransmitters such as dopamine can be addressed with antipsychotic medications. Situational stressors that exacerbate symptoms can be lessened through environmental supports such as case management and family interventions. Individual and group therapy can also help people with schizophrenia to better understand and manage their own patterns of stress and response. Substance abuse can be treated through dual diagnosis programs. Different types of rehabilitative interventions, such as social skills training, vocational rehabilitation, and cognitive remediation, can be targeted at each patient's individual needs.

The deinstitutionalization of persons with schizophrenia has made it necessary to develop new ways of providing mental health care that can accommodate the increased complexity of community-based treatment. Institutions inherently provided a form of integrated treatment, albeit of varying quality and at a high cost to individual freedom. Community-based treatment requires a more explicit determination of the key components of integrated treatment and of how to make these components accessible to patients and families in a way that is flexible enough to accommodate their varying needs. Evidence-based practice has arisen as a way of incorporating research findings into clinical practice. However, a significant difficulty in using evidence-based practice for guiding integrated treatment of schizophrenia is the limited number of studies that used appropriate methodology, including measurement of dosages of both pharmacologic and psychosocial aspects of treatment.

Most research on the treatment of schizophrenia has focused on either pharmacologic intervention or psychosocial intervention. Other potential concurrent treatments, such as use of medication among patients participating in a trial of a psychosocial intervention, have been typically designated as "standard treatment" and not systematically controlled. Comparatively little research has been done on integrated treatment as such. Studies that simultaneously control for multiple aspects of treatment can be difficult to implement but have the potential to provide information that is not otherwise available. Such information can include the clinical relevance of possible interactions of different types of treatment on different symptoms or functional domains.

The first efforts to examine both pharmacologic and psychosocial treatment were early studies that compared the two types of treatment and established the preeminent role of antipsychotic medications in the treatment of positive symptoms. May (13) placed inpatients with moderately severe schizophrenia into five treatment groups: milieu therapy, insight-oriented individual psychotherapy, antipsychotic medication, electroconvulsive therapy, and a combination of antipsychotic medication and individual psychotherapy. Both of the groups that received medication had significantly better outcomes, followed by the group that received electroconvulsive therapy. The group that received combination treatment did not do better than the group that received medication alone.

A collaborative National Institute of Mental Health (NIMH) study that compared individual supportive psychotherapy with a control condition among patients who were receiving chlorpromazine or placebo had a different result. Although medication was still necessary for a positive outcome, the incorporation of the psychotherapeutic intervention had a positive effect, indicating that the approaches could be additive (14). This finding was supported by a later study in which patients who received the depot form of fluphenazine had a better response to psychotherapy than those who received the less reliably delivered oral form of the same medication (15). These studies supported the combination of supportive rather than insight-oriented therapy with pharmacologic treatment.

Brown and Rutter (16) developed the concept of expressed emotion in an attempt to understand why some patients who were stable in the hospital rapidly relapsed after returning home. They described a family's level of expressed emotion according to an empirically derived index of three factors, including the frequency of critical comments, hostility, and "emotional overinvolvement." The recognition from studies of expressed emotion that the family environment could have an impact on relapse rates led investigators to question whether patients could be successfully treated with lower dosages of medication if family stressors were also addressed. This question had added urgency because of the significant dose-related neurologic side effects of the typical antipsychotics available at the time.

Goldstein and colleagues (17) examined the effect of a six-week crisis-oriented family intervention on relapse rates among recently discharged patients receiving either standard-dosage (25 mg) or low-dosage (6.25 mg) fluphenazine decanoate every two weeks. They found that relapse rates at the end of the six-week family intervention period ranged from zero among patients who received both standard-dosage fluphenazine and the family intervention to 48 percent among patients who received only the low-dosage fluphenazine. At six weeks both groups showed a beneficial effect from the addition of psychotherapy. However, at six months this effect had persisted only among the patients who received the higher dosage of fluphenazine, indicating that the effect of the family intervention was not sustained in the absence of adequate medication.

Hogarty and colleagues (18) examined the relationship of rates of relapse to household levels of expressed emotion among patients who were treated with a standard dosage (25 mg every two weeks) or a minimal dosage (20 percent of the standard dosage) of depot fluphenazine. They found that patients who were receiving the lower dosage were more likely to experience minor relapses in environments with high levels of expressed emotions but that at the end of the two-year study period the lower-dosage group showed more improvement in measures of social adjustment. This finding supported the hypothesis that lower medication dosages could be sufficient for patients living in households with lower levels of expressed emotions and that minimizing dosages might be advantageous for other areas of functioning.

A study by Marder and colleagues (19) assessed the relative impacts of social skills training (problem-solving model) and supportive therapy over a two-year period among patients who received a low dosage of fluphenazine decanoate (5 to 10 mg every two weeks). Patients were also treated with supplemental active oral fluphenazine or placebo if they started to manifest prodromal symptoms suggestive of relapse. At the end of the study, the group that received the social skills training showed a small but significant advantage over the group that received supportive therapy in two of six measures of social adjustment. Social skills training did not appear to affect relapse rates. On the other hand, the supplemental medication decreased rates of relapse during the second year but did not affect measures of social adjustment. The authors concluded that this finding could be evidence that the different interventions affected different areas of outcome—that is, medications for relapse and social skills training for social functioning, with the best overall results among patients who received both (20).

The large multicenter NIMH Treatment Strategies in Schizophrenia Study assigned patients who were in the maintenance phase of treatment to three different dosage strategies of fluphenazine decanoate: standard dosage, low dosage, and targeted treatment—that is, placebo plus supplemental medication if the patient developed prodromal symptoms. All patients could receive a higher dosage of antipsychotic as rescue medication if their symptoms worsened. The study compared the interaction of each medication strategy with two different types of family intervention. The family intervention took the form of monthly multifamily support groups or a combination of these support groups with more intensive and individualized behavioral family treatment.

The study found higher rates of relapse in the targeted and low-dosage medication groups. No statistically significant difference in relapse rates was found between the two family treatment groups (21). Measures of patients' social functioning, family attitudes, and family burden were also examined. Families who received the intensive behavioral intervention showed significantly less rejecting attitudes toward patients over the two-year period. However, patients' social functioning and family burden were not related to family treatment or medication (22). The investigators speculated that the lack of a substantial differential effect between the two family treatments may have been due to a positive effect of the less intensive family treatment, because the rate of rehospitalization for both family treatment groups was lower than what had been reported in other studies for usual community care. They also questioned whether patients who met the criteria for participation in the maintenance phase were already so stable that the differences in relapse rates that may have been seen in a less stable population were lost. Such issues illustrate the complexity of conducting research on integrated psychosocial and pharmacologic treatment.

The recent finding of the importance of cognitive functioning to overall outcome among patients with schizophrenia has stimulated interest in whether the effects of antipsychotic medications on cognitive measures correlate with response to psychosocial treatments. Rosenheck and colleagues (23) compared participation in psychosocial treatment between patients with treatment-resistant illness who were receiving clozapine or haloperidol. They found that patients receiving clozapine were more likely to actively participate in rehabilitation.

In a different study of patients with treatment-resistant schizophrenia, Liberman and colleagues (24) assessed the relative improvement in activities of daily living among patients who received risperidone or haloperidol while residing on an inpatient unit with a well-developed behavioral rehabilitation program. They found that although all patients experienced significant improvement, there was no difference between patients receiving the different medications.

These studies suggest that the increased efficacy seen with clozapine in this population can improve patients' ability to make use of other elements of treatment but that for patients receiving other currently available antipsychotics, a sufficiently robust psychosocial intervention may have much more of an effect on clinically significant outcome measures.

Several key points can be drawn from these studies of integrated approaches. The chief message is that medication and psychosocial interventions are efficacious for different outcomes. Although medication is essential for preventing relapse, psychosocial treatments affect other domains that are important for the quality of life. In addition, a synergistic aspect is evident. Medication can facilitate participation in psychosocial treatment, and psychosocial interventions such as family therapy can decrease relapse rates beyond what is seen for a given dosage of medication alone.

An essential aspect of guiding community-based treatment is determining which outcomes should be prioritized. This decision is particularly complicated for an illness such as schizophrenia, which can affect so many areas of functioning and involves multiple stakeholders with potentially different priorities—for example, patients, families, clinicians, administrators, and law enforcement personnel. Lehman (25) and Attkisson and colleagues (26) have suggested a framework for evaluating outcomes in schizophrenia, based on recommendations of an NIMH expert panel. This framework divides outcomes into four general domains: clinical, rehabilitative, humanitarian, and public welfare. The clinical domain includes issues such as psychopathology, symptoms, and treatments. The rehabilitative domain emphasizes the individual's strengths and ability to function adaptively, both socially and vocationally. The humanitarian domain includes concerns such as quality of life and subjective sense of well-being, and the public safety domain focuses on issues relating to the balance between the rights of the patient and the community to both liberty and to personal safety and well-being.

Lehman also proposed categorizing outcomes as either proximal or distal outcomes, according to the likely temporal relationship between an intervention and its outcome. For example, an intervention with an antipsychotic medication is likely to have an effect on easily observable clinical symptoms—and within a short period after the medication is administered—in addition to improving long-term quality of life. Results of a social skills program may not be as quickly detectable but may have a significant impact on longer-term outcomes such as ability to maintain employment. Such a framework can help to clarify the different effects of an intervention and their possible interactions. It can also help to keep a focus on more distal outcomes or those not directly within the clinical domain, which may be more difficult to study but are nonetheless important.

Effective community-based treatment also takes into account the importance of maintaining a collaborative relationship between patients, families, and treatment providers. Patients and families are more likely to remain engaged with treatment if they perceive it as being relevant to their personal concerns. The consumer movement and associated recovery model of mental illness have helped empower patients and their families to articulate their goals and the means they find most helpful to reach them (27).

A strong therapeutic alliance based on collaborative engagement and attention to multiple outcome domains is the foundation of integrated treatment and should be maintained regardless of the stage of illness (28). Different specific interventions can then be called upon within this framework for particular clinical situations.

The preceding discussion focuses on what is known about the efficacy of integrated treatment approaches. However, it has become increasingly recognized that there are several necessary steps between identifying an efficacious intervention and changing practice in community clinics to reflect those findings (65). These steps include giving attention to the difference between efficacy, or the potential of an intervention to result in a particular outcome in a research setting, and effectiveness, defined as what the outcomes are likely to be when an intervention is transplanted into a real-world setting with a more heterogeneous patient population and more obstacles in maintaining fidelity to a given treatment protocol.

The next step is determining whether the potential intervention is an efficient use of available resources. Finally, questions of how easy or difficult it may be to implement the intervention must be considered. A growing body of literature about dissemination of changes shows that education alone is seldom effective and that a multipronged approach—including strong administrative support, hands-on exposure, redesign of professional incentives to support new methods, and ongoing support and feedback—are necessary to make a sustained difference in the actual practice of mental health care delivery (37,66,67,68). Such changes may represent a significant burden to already beleaguered mental health care systems and must be planned well ahead of time if proposed changes are to succeed.

Schizophrenia arises from a complex interaction between genetic and environmental factors that affect multiple functional domains in a heterogeneous fashion. The ways in which these different domains contribute to the psychopathology of schizophrenia are gradually being delineated. It has become recognized that pharmacologic and psychosocial interventions both have a necessary place in the treatment of schizophrenia, and the past decades of research on these different forms of intervention are beginning to clarify how each affects specific types of symptoms and areas of functioning.

The importance of cognitive deficits in psychosocial outcomes emphasizes the need to systematically examine other variables besides relapse, hospitalization, and positive symptoms. As our understanding of schizophrenia continues to become more sophisticated, it will be important to conduct research studies that are designed to clarify the ways in which different treatment methods and patterns of psychopathology interact (20). Such studies will provide the evidence needed to guide the integration of these interventions in the service of patients, their families, and communities.

The authors are affiliated with the department of psychiatry at the University of New Mexico School of Medicine, 2400 Tucker N.E., Albuquerque, New Mexico 87131 (e-mail, [email protected]). This paper is part of a special section on integrated care for persons with mental illness.