Initial Results from the Schizophrenia PORT Client Survey (continued)
Recommendation 17. Persons who experience persistent and clinically significant associated symptoms of anxiety, depression, or hostility, despite an adequate reduction in positive symptoms with antipsychotic therapy, should receive a trial of adjunctive pharmacotherapy. A trial of a benzodiazepine or propranolol has merit for persistent anxiety. An antidepressant trial should be considered for persistent depression. Adjunctive therapy with lithium, a benzodiazepine, or carbamazepine should be considered for persistent hostility or manic-like symptoms. The reasons for the absence of such trials for appropriate patients should be documented. Certain adjunctive medications should be avoided in patients currently receiving clozapine to avoid synergistic side effects, for example, respiratory depression with benzodiazepines and bone marrow suppression with carbamazepine. (Adjunctive depression medications; adjunctive anxiety medications)
Conformance Criteria. Patients who score within the upper quartile on the Symptom Checklist (SCL-90; Derogatis et al., 1973) depression subscale or who report having a current diagnosis of depression on the Client Survey or who have a current chart diagnosis of comorbid depression are prescribed an antidepressant. Patients who report a current diagnosis of anxiety disorder on the Client Survey or who have a current chart diagnosis of comorbid anxiety disorder are prescribed an antianxiety agent. It was not possible to assess hostility from the data.
Conformance Rates (Adjunctive depression medications). Some 48.3 percent of inpatients met at least one of the comorbid depression criteria; and of them, 33.8 percent were prescribed an antidepressant. Among the outpatients, 42.8 percent met the depression criteria, and 45.7 percent of them were prescribed an antidepressant. Among patients meeting the criterion of need for antidepressant therapy, outpatients were more likely than inpatients to conform to the treatment recommendation (45.7 percent and 32.2 percent, respectively; x2 = 4.8, p = 0.03). Depression treatment for outpatients in rural areas was more likely to meet the recommendation criteria than treatment for urban outpatients (66.7 percent and 39.6 percent, respectively; x2 = 7.9. p = 0.005). Overall, treatment of depressed minority patients was less likely to meet the conformance criterion than treatment for depressed Caucasian patients (29.6 percent and 47.3 percent, respectively; x2 = 8.6, p = 0.003).
Conformance Rates (Adjunctive anxiety medications). Some 17.8 percent of inpatients met the current anxiety disorder criterion, and of these, 33.3 percent were prescribed an anxiolytic. Among the outpatients, 22.8 percent met the current anxiety disorder criterion, and 41.3 percent of them were prescribed an anxiolytic. There was a significant interaction between gender and treatment setting. For the inpatient sample, men with a comorbid report of anxiety were more likely than women to be prescribed an antianxiety agent at discharge (48.4 percent and 5.9 percent, respectively; x2 = 7.1, p = 0.008). However, the opposite was observed among outpatients, among whom 56.8 percent of women with anxiety were prescribed antianxiety agents compared with 30.9 percent of men (x2 = 5.1, p = 0.02). This interaction suggests that antianxiety agents are used for different purposes among inpatients and outpatients. The pattern is consistent with the use of anxiolytics to calm agitated men in the hospital and to reduce anxiety among women in the community.
Recommendation 18. Persons who experience persistent and clinically significant positive symptoms despite adequate antipsychotic therapy, including trials with the newer antipsychotics. should receive a trial of adjunctive pharmacotherapy as described in Recommendation 17. (Adjunctive psychosis medications)
Conformance Criteria. Patients who score in the upper quartile on the SCL psychosis subscale and who are currently on antipsychotic medication are prescribed either lithium or an anticonvulsant. It was not possible to ascertain whether prior trials with the newer antipsychotic medications had been conducted.
Conformance Rates. Of the inpatients on antipsychotic medications who fell into the upper quartile on persistent psychotic symptoms. 22.9 percent were prescribed either lithium or an anticonvulsant. Among the outpatients, 14.4 percent of those meeting the persistent psychotic symptoms criterion were receiving either lithium or an anticonvulsant. No significant relationships were found between conformance with this recommendation and the various patient and treatment setting variables.
Recommendation 23. Individual and group therapies employing well-specified combinations of support, education, and behavioral and cognitive skills training approaches designed to address the specific deficits of persons with schizophrenia should be offered over time to improve functioning and enhance other target problems, such as medication noncompliance. (Psychotherapy)
Conformance Criteria. It was difficult to assess this recommendation adequately from the available data. All that could be determined was whether the treatment plan included individual or group psychotherapy and whether the patient reported that a treatment provided had addressed any of a series of psychological and practical life problems.
Conformance Rates. Among inpatients, treatment plans indicated that 96.5 percent were prescribed either individual or group therapy at the time of discharge. In the inpatient group, 64.7 percent indicated at the followup interview that they were receiving help with at least one life problem. Most received assistance from a psychiatrist or a case manager. Among the outpatients, the treatment plan indicated that 45.0 percent were receiving individual or group therapy; 76.7 percent stated that they were receiving help for at least one life problem.
Several significant relationships were found between the prescription of psychotherapy and patient and setting characteristics.
Across the entire sample, psychotherapy was more likely to be prescribed to inpatients at the time of discharge than to outpatients (96.6 percent and 45.3 percent, respectively; x2 = 2.24, p = 0.001), to younger patients (80.5 percent for age under 35, 72.0 percent for 35-44, and 52.9 percent for age 45 or older: x2 = 39.7, p = 0.001), to rural patients (74.8 percent rural and 63.4 percent urban; x2 = 6.1, p = 0.01), to patients in State A than in State B (84.2 percent and 53.8 percent, respectively; x2 = 65.3, p = 0.001), and to minority patients (70.6 percent and 61.9 percent, respectively; x2 = 5.l, p = 0.02).
Recommendation 24. Patients who have ongoing contact with their families should he offered a family psychosocial intervention that spans at least 9 months and that provides a combination of education about the illness, family support. crisis intervention, and problem-solving skills training Such interventions should also be offered to nonfamily members. (Family)
Conformance Criteria. Available data do not provide a very adequate basis for assessing this recommendation. However, the following assessment is possible. For patients who report some ongoing contact with family in the local area, they report that some family member(s) has received information about schizophrenia and/or has attended an educational or support program. An alternative criterion used is that "family therapy or support" is prescribed in the treatment plan for patients with available family.
Conformance Rates. Most inpatients (84.5 percent) had some ongoing contact with their families during the past year. Based on review of their records or on the Client Survey interviews, a total of 40.8 percent inpatients were offered or received a family service. As with the inpatients, most outpatients (77.2 percent) had some ongoing contact with their families, and 37.2 percent were offered or received a family service. State A patients were more likely to be prescribed a family intervention than were State B patients (26.8 percent and 13.9 percent, respectively; x2 = 13.1, p = 0.001), and rural outpatients were more likely than urban outpatients to be prescribed a family intervention (23.4 percent and 5.1 percent, respectively; x2 = 20.2, p = 0.001).
Recommendation 27. Persons with schizophrenia who have any of the following characteristics should be offered vocational services: The person:
- Identifies competitive employment as a personal goal.
- Has a history of prior competitive employment.
- Has a minimal history of psychiatric hospitalization.
- Is judged on the basis of a formal vocational assessment to have good work skills.
Conformance Criteria. Currently unemployed patients who have a prior work history or who are currently looking for work either report participating in a program to help them find a job, or vocational rehabilitation is prescribed in the treatment plan. Currently employed patients report receiving assistance from a job coach or other employment specialist.
Conformance Rates. For the unemployed inpatients, 30.7 percent either had vocational rehabilitation prescribed in their discharge plans and/or reported participating in a vocational program. Among the employed inpatients, 25 percent had a job coach. Combining employed and unemployed inpatients, 30.4 percent met the conformance criteria. Among the unemployed outpatients, 22.6 percent either had vocational rehabilitation prescribed in their treatment plans and/or reported participating in a vocational program. Of the outpatients who were employed, 21.7 percent had a job coach. Combining employed and unemployed outpatients, 22.5 percent met the conformance criterion. For the total sample, the vocational rehabilitation recommendation was more likely to be met among younger patients (35.1 percent of age under 35, 29.2 percent of age 35-44, and 17.7 percent of age 45 and older; x2 = 17.3, p = 0.001). The treatment of men was somewhat more likely than treatment of women to conform to this recommendation (28.4 percent and 21.1 percent, respectively; x2 = 4.2, p = 0.04). Inpatients at the time of discharge were more likely to meet the criteria than were outpatients (30.4 percent and 22.5 percent, respectively; x2 = 4.6, p = 0.03), as were State B inpatients compared with State A inpatients (30.4 percent and 22.5 percent, respectively; x2 = 4.6, p = 0.03).
Recommendation 29. Systems of care serving persons with schizophrenia who are high users should include ACT and ACM programs. (ACT/ACM)
Conformance Criteria. Whether a patient was participating in a formal ACT or ACM program that meets formal standards for these program models could not be determined from the data. However, a series of ACT/ACM service characteristics could be determined from the patients' responses to questions about case management services. These criteria included:
- Receipt of case management services.
- Having at least some visits by the case management team outside the office.
- Seeing the case management team at least weekly.
- Receiving help from the case management team for at least four of seven life problem areas listed.
Two criteria levels were set. For the first, narrow level, a patient has to report receiving services that met all four criteria. For the second, broad level, a patient has to report services meeting three of these criteria.
Conformance Rates. Some 1.9 percent of inpatients met the narrow criterion at followup, and 8.6 percent the broad criterion. For outpatients, 2.2 percent met the narrow criterion, and 10.1 percent the broad criterion. It is not known what percentage of patients with schizophrenia should receive ACT or ACM. The only variable significantly associated with the criteria for ACT/ACM was urban/rural setting: Rural patients were more likely to be receiving services consistent with these criteria (17.0 percent and 10.1 percent, respectively; x2 = 4.6, p = 0.03).
This study is one of the first attempts to examine variations in patterns of care for persons with schizophrenia under usual treatment conditions in relationship to scientifically based treatment standards. As such, the findings should provoke both further study and action to try to improve care in the community.
Several strengths of the study provide confidence that its findings can form the basis for actions to improve care:
- First, the sampling frame was broad and the sample large, lending confidence that the findings are reasonably representative of current practices.
- Second, the sampling procedure sought to reduce avoidable selection bias by randomly sampling treatment settings and patients within these settings.
- Third, standardized procedures were employed for abstracting medical record data and for interviewing subjects, and the response rate for this type of survey was substantial.
- Finally, the PORT procedures employed in establishing the treatment recommendations were quite rigorous, and therefore the standards of care applied have substantial scientific validity.
At the same time, this study is an initial attempt to answer complex questions about how usual care measures up to scientifically derived standards of care, and it has a variety of limitations. The process of reducing the available data on treatment to dichotomous conformance/nonconformance ratings undoubtedly results in varying degrees of imprecision and reductionism. The data and criteria for prescription and dosages of antipsychotic medications are most precise. Far less precise are the data and criteria for judging the receipt of psychosocial interventions that meet the PORT recommendation criteria. For example, the Conformance Criteria for the recommendation on psychotherapy relied simply on whether treatment in this category was prescribed in the treatment plan. Most likely a substantial proportion of the services actually delivered in these psychosocial categories did not meet all criteria specified in the recommendations. Therefore, these estimates of conformance are probably inflated.
The following example illustrates the complexity of estimating treatment conformance and the methodological issues in this type of research that need further study. In assessing the conformance of treatment to Recommendation 17 on the use of adjunctive antidepressant medications, we used medical record and client interview items to ascertain the presence of depression (a chart diagnosis or high level of self-rated depression symptoms). Even assuming the accuracy of these data for ascertaining depression. it is not clear that all patients who meet this criterion need antidepressant therapy. The conformance rating provides the percentage of patients who meet the depression treatment recommendation criterion, but lacking is an empirical standard for the percentage of schizophrenia patients with comorbid depression who should receive an antidepressant. The efficacy data are clear that such adjunctive pharmacotherapy is helpful to many of these depressed patients, but because of individual variations in response and the course of depression and contraindications to the use of antidepressants, it is unlikely that 100 percent should receive the medication.
This said, it still seems likely that our finding that only 32.2 percent of inpatients and 45.7 percent of outpatients with comorbid depression received adjunctive antidepressants does point to undertreatment of this problem. Furthermore, it is difficult to reconcile the observed racial differences in the use of antidepressants on methodological grounds. It seems clear that there is a relative undertreatment of depression among minority patients with schizophrenia.
There is also the assumption that conformance to treatment standards reflects quality of care and produces better outcomes. We did not attempt to address this assumption in this study, although we may be able to examine some relationships between patterns of care and outcome in future analyses of the PORT data. Stated somewhat differently, it may not be assumed that conclusions about treatment efficacy derived from randomized clinical trials translate directly into effectiveness in usual practice (Lehman et al., 1995a). There are a variety of reasons for this so-called efficacy-effectiveness gap, including greater patient heterogeneity (for example, presence of comorbid conditions), greater practice heterogeneity (for example, different dosing practices), and greater patient noncompliance under usual practice conditions than in well-controlled clinical trials. The PORT treatment recommendations are best estimates based on the scientific data available, but their validity as quality-of-care indicators for usual care remains unconfirmed. In addition, it is not assumed that all patients should be treated in conformance with the recommendations. The rates of appropriate deviations from the treatment recommendations that represent optimal individual variations in care are not known.
There is a variety of strategies available for converting information on treatment into ratings of quality of care (Wells and Brook, 1989). How sensitive our findings are to alternative methods for rating treatment recommendation conformity is not known. We employed a highly structured approach to estimating conformance. Trained, nonclinical record abstractors extracted concrete medical record information, such as type and dose of medications and treatments specified in the formal treatment plan. They did not provide conformance judgments.
The conformance ratings were derived with electronic algorithms using the abstracted data and Client Survey items. The algorithms reflected the PORT investigators' best judgments about how conformance could be rated using the available data. The advantage of this procedure is that it avoids the problem of interrater variance in judgments when reviewing a medical record. The disadvantage is that it limits the types of data available for estimating conformance and precludes the type of sophisticated judgments about the quality of care that can be made when highly trained clinician assessors read and integrate information from the entire record. Future research should examine conformance estimates using this latter, more standard quality of care methodology and should compare findings from other studies with those presented here.
Finally, the methods used were primarily cross-sectional. We focused on point-in-time treatment plans and did not attempt to look at changes in treatment over time. For example, we asked patients if they were depressed and evaluated their use of antidepressants based on the presence of depression. This analysis does not capture prior use of antidepressants among persons who were treated appropriately and responded, nor does it reflect failure to treat prior depressions, which subsequently resolved after considerable suffering.
With these caveats in mind, it is nonetheless worthwhile to note some of the major trends observed in the survey.
- For nearly all recommendations, the level of conformance is modest at best, with the exception of the rates of prescription of antipsychotic medications (acute neuroleptic and maintenance neuroleptic). For most recommendations, less than half the patients were receiving treatment that met the recommendation criteria.
- Overall, rates of conformance are lower for the psychosocial treatment recommendations than for the pharmacological recommendations. Data were not available to determine whether this relative lack of access to psychosocial treatments was due to the absence of these treatments in certain locales or to inadequate use of existing psychosocial services.
- Few consistent relationships were found between conformance with the recommendations and patient demographics. Younger patients were more likely to be offered psychotherapy and vocational rehabilitation. Of concern is the finding that minority patients were more likely to be on higher doses of antipsychotic medications and less likely to be prescribed an antidepressant when depressed.
- The discharge treatment plans of the inpatient sample received higher conformance ratings with the psychosocial treatment recommendations (psychotherapy, family, vocational rehabilitation) than did the treatment plans of outpatients. This finding suggests a tendency for psychosocial treatments to be prescribed at the point of discharge, but also suggests a low rate of follow through. Failure to consider these important treatments for more stable outpatients may be a serious problem in ongoing community-based care.
- Patterns of recommendation conformance varied by location for the psychosocial treatments, but much less so for the pharmacological recommendations. Patients in State B were more likely than those in State A to be prescribed a vocational intervention and less likely to be prescribed a family intervention or psychotherapy. The appropriate use of psychosocial interventions may be more vulnerable to local idiosyncrasies than are pharmacological treatments.
- The treatment of patients in rural areas was more consistent with the maintenance CPZ dose. adjunctive depression medications, psychotherapy, family, and ACT/ACM recommendations than was treatment of urban patients.
The findings of this survey need to be replicated in other samples and settings using variations on our methodology to evaluate their generalizability and robustness. We consider this study an early step in the development of quality-of-care research and standards in schizophrenia. As such, it moves the field forward. At the very least, it should stimulate more quality-of-care research in schizophrenia and also provoke concern about the quality of care currently afforded persons with schizophrenia in the United States.
Lehman AF, Carpenter WT, Jr, Goldman HH, and Steinwachs DM. Treatment outcomes in schizophrenia: Implications for practice, policy, and research. Schizophrenia Bulletin 1995a;21(4):669-75.
Lehman AF, Steinwachs DM, the CoInvestigators of the PORT project. Translating research into practice: The Schizophrenia Patient Outcomes Research Team (PORT) Treatment Recommendations. Schizophrenia Bulletin 1998; 24(1):1-10.
Lehman AF, Thompson JW, Dixon LB, Scott JE. Schizophrenia: Treatment outcomes research—Editors' introduction. Schizophrenia Bulletin 1995b; 21(4):561-6.
Wells KM, Brook R. The quality of mental health services: Past, present, and future. In: Taube CA, Mechanic D, Hohmann AA, editors. The Future of Mental Health Services Research. Washington, DC: U.S. Government Printing Office, HHS Publication No. (ADM)89-1600, 1989. pp. 203-24.
Anthony F. Lehman, M.D., Principal Investigator of the Schizophrenia Patient Outcomes Research Team (PORT), is Professor of Psychiatry, University of Maryland School of Medicine and Co-Director of the Center for Mental Health Services Research, University of Maryland, Baltimore, MD.
Donald M. Steinwachs, Ph.D., Co-Principal Investigator, is Professor of Health Policy and Management and Chair of the Department of Health Policy and Management, The Johns Hopkins University School of Hygiene and Public Health, Baltimore, MD.
PORT Co-Investigators at the Center for Mental Health Services Research, University of Maryland School of Medicine include Lisa B. Dixon, M.D., M.P.H.; Leticia Postrado, Ph.D.; and Jack E. Scott, Sc.D. Co-Investigators from the Department of Health Policy and Management, The Johns Hopkins University School of Hygiene and Public Health include Maureen Fahey, M.A.; Pamela Fischer, Ph.D.; Jeffrey Hoch, M.A.; Judith Ann Kasper, Ph.D.; Alan Lyles, Sc.D.; Andrew Shore. Ph.D.; and Elizabeth Ann Skinner, M.S.W.
Adapted from the article published in Schizophrenia Bulletin 1998:24(1)11-20.