Patterns of Usual Care for Schizophrenia

Initial Results from the Schizophrenia Patient Outcomes Research Team (PORT) Client Survey


Recommendations for the treatment of schizophrenia based on existing scientific evidence were developed by the Schizophrenia Patient Outcomes Research Team (PORT). Based on exhaustive reviews of the treatment outcomes literature, the treatment recommendations focus on treatments for which there is substantial evidence of efficacy.

To examine conformance of current patterns of usual care for persons with schizophrenia with its treatment recommendations, the PORT surveyed a stratified random sample of 719 persons diagnosed with schizophrenia. Findings indicate that current usual treatment practices likely fall substantially short of what would be recommended based on the best evidence on treatment efficacy.


Adapted from Schizophrenia Bulletin 1998;24(1):11-20

By Anthony F. Lehman, Donald M. Steinwachs, and the Survey Co-Investigators of the PORT Project

Abstract / Background / Methods / Results / Treatment Recommendations /
Discussion / References / Author Information


Abstract

To examine the conformance of current patterns of usual care for persons with schizophrenia to the treatment recommendations of the Schizophrenia Patient Outcomes Research Team (PORT), the PORT surveyed a stratified random sample of 719 persons diagnosed with schizophrenia in two States. The types of treatment settings surveyed included acute inpatient programs and continuing outpatient programs in urban and rural locales. Using data from medical record reviews and patient interviews, the PORT assessed the conformance of current care with 12 of the treatment recommendations.

The rates at which patients' treatment conformed to the recommendations were modest at best, generally below 50 percent. Conformance rates were higher for pharmacological than for psychosocial treatments and in rural areas than in urban ones. Rates of treatment recommendation conformance for minority patients were lower than those for Caucasians, and patterns of care varied between the two States. The findings indicate that current usual treatment practices likely fall substantially short of what would be recommended based on the best evidence on treatment efficacy. This disparity underscores the need for greater efforts to ensure that treatment research results are translated into practice.

Key words: Mental health services, quality of care


Background

A principal aim of the Agency for Health Care Policy and Research/National Institute of Mental Health (AHCPR/NIMH) Schizophrenia Patient Outcomes Research Team (PORT) is a better understanding of variations in patterns of treatment for persons with schizophrenia in usual care and the implications of these variations in light of available scientific knowledge about treatment efficacy (Lehman et al., 1995b). A companion article (Lehman et al., 1998) presents the AHCPR/NIMH Schizophrenia PORT treatment recommendations, which attempt to summarize knowledge about treatment efficacy. This article provides initial estimates of the rates at which usual treatment conforms to the PORT recommendations based on the PORT Client Survey. By identifying aspects of usual treatment in need of further improvement, these rates of conformance add an important dimension to our assessment of implications of the PORT treatment recommendations.

Return to Contents

Methods

Overview

We administered the PORT Client Survey in face-to-face interviews with a random sample of 719 persons with a clinical diagnosis of schizophrenia who were under usual care in two States, one in the South and the other in the Midwest. All subjects provided written informed consent for the study and were paid $10 for their time. They completed a 90-minute face-to-face survey interview and permitted a review of their current medical record. We used medical record data, augmented by information from the interview, to assess whether current treatment conformed to the PORT treatment recommendations. Variations in rates of conformance were then determined and examined in relation to patient demographic characteristics and treatment settings.

Sampling

The PORT contract specified a random, though not necessarily epidemiologically representative, sample of persons currently under treatment for schizophrenia in usual care settings in two States. These settings included acute inpatient and outpatient programs in the public, private, and Department of Veterans Affairs (VA) sectors of care. The sampling strategy was conducted at four levels:

  1. States.
  2. Communities.
  3. Providers.
  4. Patients.

Criteria for selection of the two States emphasized three factors:

  1. A sufficient population base to make available for study an adequate number of persons with schizophrenia.
  2. The quality and thoroughness in the State's Medicaid claims data and the capacity to link these claims data to State hospital utilization data (a requirement for another phase of the PORT).
  3. Geographic and population diversity.

Based on these initial screening criteria, we identified eight candidate States. Additional information on mental health resources and utilization in each of these States and recommendations from the Schizophrenia PORT National Advisory Panel guided the selection of the two States, with final approval by AHCPR and NIMH. The States are not identified here to maintain confidentiality. One is a large Southern State (State A) and the other a large Midwestern State (State B).

Once these States were selected and agreed to participate, the next stage of sampling was conducted at the community level. The PORT required that treatment patterns in five "communities" be sampled across the two States and that in each State at least one rural community be included. To facilitate community selection, these sources of information were studied:

Again with the guidance of the two States and the AHCPR/NIMH, we then selected three urban communities (one in State A and two in State B) and two rural areas (one in each State).

The next level of sampling was at the treatment provider level within these five communities. The PORT design specified sampling of patients experiencing two different phases of treatment, either an acute inpatient episode or ongoing community-based care. An inventory of provider organizations located in each of the communities was compiled from the NIMH's Inventory of Mental Health Organizations and General Hospital Mental Health Services 1990 Survey. We further subclassified providers as follows:

To ensure that the provider organizations selected had adequate patient volumes, a criterion was established whereby an inpatient provider had to have discharged at least 100 patients with schizophrenia within the past year and an outpatient provider had to have at least 100 current outpatients with schizophrenia. Provider organizations were randomly selected from a list of those meeting these criteria and their participation solicited.

In State A, of the original 11 provider organizations that were selected randomly, 2 refused to participate. One provider did not want to participate in the study, and the other, selected as an outpatient site, reported that it did not actually provide outpatient services. In State B, of the original 14 provider organizations that were selected randomly, 3 refused to participate. Reasons given were insufficient volume of outpatients with schizophrenia seen, takeover by another corporation resulting in a state of transition during the survey period, and lack of an Institutional Review Board structure to review research efforts. Alternate providers were recruited successfully.

Finally within each provider setting, the PORT selected patients with a diagnosis of schizophrenia at random from treatment rosters. All subjects met the following criteria:

For the acute inpatient episode sampling, all admissions were screened for eligibility, and consent was sought from all eligible admissions. For the community-based sampling frame, each provider site compiled a roster of all current patients with schizophrenia. The number of patients sampled from each site's roster was determined by the practice setting stratum in which that site was categorized to achieve target sample sizes within each strata.

The original target sample comprised 750 community-based subjects and 450 acute episode subjects. Of a total of 663 inpatients screened as initially eligible for the survey, 69.1 percent (458) agreed to allow the treatment program to release their names to the study. Of these, 398 met a more detailed eligibility assessment, and 279 of them (70.1 percent) completed the survey. The reasons for these later ineligibility determinations included:

A total of 1,017 community-based patients met an initial eligibility screen, and 584 (57.4 percent) of them gave permission to their treatment program to release their names to the study. Subsequent eligibility assessments revealed that 550 were actually eligible for the survey, and 440 of them (80.0 percent) completed the survey. Reasons for these later ineligibility determinations included being clinically unable to be approached for an interview (n= 20), out of geographic area (n = 6), and deceased (n = 8).

Hence, two types of sample attrition were encountered, first at the point of patients' consent for their treatment providers to release their names to the study and second at the point of consent and completion of the actual survey. The former type of attrition was more substantial than the latter, especially among the community samples. There were no significant differences between completers and noncompleters by gender, race, or age. The final sample is presented in Table 1.

Table 1. PORT-Client Survey sample characteristics


                            State A                  State B                  Total
                       Acute      Community     Acute      Community     Acute      Community
                      inpatient                inpatient                inpatient
					  _______________________________________________________________________________________________________
Number of subjects      101          181         178          259         279          440
Percentage, male         65.4         60.2        65.2         62.9        65.2         61.8
Age (%)
 <35                     33.7         29.8        25.8         15.8        28.7         21.6
 35-44                   38.6         25.4        40.5         32.4        39.8         29.6
 45-64                   26.7         39.2        30.3	       42.1	   29.0         40.9
 65+                      1.0          5.5         3.4          9.7         2.5          8.0
Race (%)
 White                   18.8         52.5        58.8         64.6        44.2         59.6
 African-American        79.2         44.8        34.5         30.7        50.7         36.5
 Other                    2.0          2.8         6.8          4.7         5.0          3.9
Locale (%)
 Urban                   60.4         75.7        89.9         78.8        79.2         77.5
 Rural	                 39.6         24.3        10.1         21.2        20.8         22.5
Marital status (%)
 Never married           58.4         53.6        59.3         49.6        59.0         51.3
 Currently married        7.9         15.5         9.0         14.7         8.6         15.0
 Other                   33.7         30.9        31.7         35.7        32.4         33.7
Education (%)
 <High school            35.6         29.3        27.5         29.3        30.5         29.3
 High school graduate    40.6         28.2        34.3         46.3        36.6         38.9
 Some college            23.8         42.5        38.2         24.3        33.0         31.8
 

Note: PORT = Patient Outcomes Research Team.

Measures

The PORT created three data collection instruments: a client interview survey and chart abstracts for the inpatient and outpatient chart reviews. The PORT Client Survey has 10 sections:

Because the survey data were used in only a very limited way for the analyses presented here, the PORT Client Survey is not described in any detail. For these analyses, it provided information on patients' current level of symptoms, medication side effects and compliance, frequency of family contacts, and employment status. For the acute inpatient sample, the Client Survey was administered 3 months after the date of index hospital admission. For the continuing community care sample, it was conducted when the person consented to the survey.

Inpatient and outpatient chart abstract forms were designed to capture information on each patient's psychiatric and medical history, health services utilization, and to review the treatment plan, current medications, and family contacts and services. The inpatient record review focused on the treatment plan at the point the patient was discharged from the hospital.

Data Analysis

The primary sources of data to assess conformance with the PORT treatment recommendations were the PORT Outpatient Record Review and Inpatient Record Review forms, supplemented by some data from the PORT Client Survey. Even with these sources of data, it was not possible to assess conformance with all 30 of the PORT treatment recommendations. Some treatment recommendations require information about aspects of longitudinal course and response to treatment that simply was not available from the client interviews or the medical records.

However, using existing data from the Client Survey, we were able to estimate conformance with 12 recommendations to varying degrees. These 12 recommendations encompass all of the critical treatment modalities, including pharmacotherapies, psychological treatments, family interventions, vocational rehabilitation, and assertive community treatment (ACT)/assertive case management (ACM). For each recommendation, we identified and combined critical items from the Client Survey and the Inpatient and Outpatient Record Review forms to provide the best estimates of conformance for each recommendation.

For each patient, a dichotomous conformance rating (0 = nonconformance, 1 = conformance) was generated for each relevant recommendation. These conformance ratings were derived using electronic algorithms based on the data obtained from record reviews and client interviews. This procedure involved no judgment by the data abstractors about conformance.

The data abstracted from the medical records were quite concrete; that is, such items as names of medications, dosages, and the types of services specified in the formal treatment plan. Rates of conformance for each recommendation were then computed separately for the acute episode and community care samples. Conformance rates were also computed according to the following patient and treatment setting variables: gender (male/female), age (<35, 35-44, 45-64, 65+), race (Caucasian, minority), State (State A versus State B), and urban versus rural. We used chi-square analysis throughout the data analysis.

Return to Contents

Results

The results are presented by treatment recommendation. We first describe criteria for determining conformance and then summarize overall rates of conformance, as well as rates by patient and setting subgroups. The overall conformance rates are also summarized in Table 2.

Table 2. Rates of PORT Treatment Recommendation conformance for inpatients and outpatients


                                                              Conformance Rates (%)
Recommendation                                           Inpatients        Outpatients
_______________________________________________________________________________________________________
Acute neuroleptic                                           89.2                NA
Acute CPZ dose                                              62.4                NA
Maintenance neuroleptic                                      NA                92.3
Maintenance CPZ dose                                         NA                29.1
Anti-Parkinson                                              53.9               46.1
Depot                                                       50.0               35.0
Adjunctive depression medications                           32.2               45.7
Adjunctive anxiety medications                              33.3               41.3
Adjunctive psychosis medications                            22.9               14.4
Psychotherapy                                               96.5               45.0
Family                                                      31.6                9.6
Vocational rehabilitation                                   30.4               22.5
ACT/ACM                                                      8.6               10.1

Note: CPZ = chlorpromazine; ACT = assertive community treatment; ACM = assertive case management; PORT = Patient Outcomes Research Team; NA = not available.

Return to Contents

Treatment Recommendations

Pharmacotherapies

Recommendation 1. Antipsychotic medications, other than clozapine, should be used as the first-line treatment to reduce psychotic symptoms for persons experiencing an acute symptom episode of schizophrenia. (Acute neuroleptic)

Conformance Criterion. At the time of discharge from the index inpatient episode, the patient was prescribed an antipsychotic agent. This recommendation could not be assessed for outpatients because we did not have data on how recently any acute symptom episodes occurred for these patients.

Conformance Rates. Some 89.2 percent of inpatients were prescribed an antipsychotic agent, including 8.9 percent on clozapine, in their discharge treatment plans. There were no significant differences in rates of conformance by gender, race, or age; between the two States; or between urban and rural samples.

Recommendation 2. The dosage of antipsychotic medication for an acute symptom episode should be in the range of 300-1,000 chlorpromazine (CPZ) equivalents per day for a minimum of 6 weeks. Reasons for dosages outside this range should be justified. The minimum effective dose should be used. (Acute CPZ dose)

Conformance Criterion. The antipsychotic dosage at the time of inpatient discharge is within the 300-1,000 CPZ range. Data were not available to assess the length of the neuroleptic trial or reasons for dosages outside the 300-1,000 CPZ range. This recommendation could not be assessed for outpatients for the same reason given under Recommendation 1.

Conformance Rates. Among the inpatients prescribed an antipsychotic medication at discharge, 62.4 percent were receiving a dosage in the 300-1,000 CPZ range, 15 percent were prescribed a dosage less than 300 CPZ equivalents, and 22.5 percent were given a dosage above 1,000 CPZ equivalents. Rates of conformance did not differ between men and women or across the age groups, nor did State of residence or urban versus rural locale predict differences in Conformance Rates.

Race did have an effect on conformance rates: the percentages of Caucasian and minority patients whose dose of antipsychotic medication fell within the recommended range were equivalent (62.9 percent and 61.4 percent, respectively). However, among the patients whose dosage fell outside the recommended range, minority patients were much more likely to be on a high dose (>1,000 CPZ) than Caucasian patients (27.4 percent and 15.9 percent, respectively, x2 = 8.8, p = .01).

Recommendation 7. Prophylactic use of anti-Parkinson agents to reduce the incidence of extrapyramidal side effects (EPS) should be determined on a case-by-case basis, taking into account patients and physician preferences, prior individual history of EPS, and other risk factors for both EPS and anticholinergic side effects. The effectiveness of and continued need for anti-Parkinson agents should be assessed in an ongoing fashion. (Anti-Parkinson)

Conformance Criterion. The patient reported some EPS on the Client Survey, and an anti-Parkinson agent is prescribed in the medical record. These criteria can be assessed for both inpatient and outpatients. However, it was not possible to ascertain how the factors identified in the recommendation were related to the decision to prescribe an anti-Parkinson agent.

Conformance Rates. Of the inpatients who were receiving an antipsychotic medication at discharge, 74.2 percent reported at least one of the following possible EPS that they attributed to this medication:

Of these, 53.9 percent were prescribed an anti-Parkinson agent. Among the outpatient sample, 79.1 percent of those prescribed antipsychotic medication reported at least one of the possible EPS. Of these, 46.1 percent were receiving an anti-Parkinson agent. No differences in rates of conformance with this recommendation were found between inpatients and outpatients, urban and rural samples, or State of residence. Gender and age were also not related to recommendation conformance. Minority inpatients were more likely to meet this recommendation than were Caucasian inpatients (64.0 percent and 40.3 percent. respectively; x2 = 8.9, p = 0.003). This race effect was not found among outpatients.

Recommendation 8. Persons who experience acute symptom relief with an antipsychotic medication should continue to receive this medication for at least 1 year subsequent to symptom stabilization to reduce the risk of relapse or worsening of positive symptoms. (Maintenance neuroleptic)

Conformance Criterion. The current outpatient treatment plan includes prescription of an antipsychotic agent. Prior relief of an acute symptom episode is presumed, but cannot be confirmed, nor can the duration of the maintenance treatment be determined. This recommendation is not relevant to the inpatient sample.

Conformance Rates. Some 92.3 percent of outpatients were prescribed an antipsychotic medication. No effects of gender, age, race; State; or urban/rural locale were found.

Recommendation 9. The maintenance dosage of antipsychotic medication should be in the range of 300-600 CPZ equivalents (oral or depot) per day. If the initial dosage to relieve an acute symptom episode exceeds this range, efforts should be made to reduce the dosage gradually to this range, such as a 10 percent reduction in dosage every 6 weeks until either early signs of relapse begin or until the lower level of this recommended range is achieved (see Recommendation 2). The new maintenance dosage should be at the last level at which symptoms were well controlled. Dosages in excess of 600 CPZ equivalents per day should be avoided unless symptom control and patient comfort are clearly superior at these higher dosages. The lowest effective dose should be used. (Maintenance CPZ dose)

Conformance Criterion. The current dosage of outpatient antipsychotic medication is in the 300-600 CPZ range. It was not possible to assess attempts to lower dosages higher than 600 CPZ or reasons for doses outside the 300-600 CPZ range.

Conformance Rates. Among the outpatients, 29.1 percent were prescribed a dosage in the 300-600 CPZ range; 39.1 percent were prescribed dosages below this range and 31.9 percent above this range. Maintenance dosages among urban patients were less likely than those of rural patients to conform to the recommendation (26.7 percent and 36.3 percent, respectively) and were more likely to fall in the high range (>600 CPZ) (36.3 percent and 18.8 percent, respectively; x2 = 8.6, p = 0.01).

Recommendation 12. Depot antipsychotic maintenance therapy should be strongly considered for persons who have difficulty complying with oral medication or who prefer the depot regimen. Depot therapy may be used as a first-option maintenance strategy. (Depot)

Conformance Criteria. We used patient self-reports of noncompliance with medications and rates of prescription of depot therapy as the conformance criteria. The reasons for prescribing depot medication could not be determined. The Client Survey asked patients how often they take their medications as prescribed, and they could respond "always." "missed a few," "missed many," or "stopped medication." The question did not distinguish between missed pills and missed injections. Noncompliance was assessed using two thresholds to assess sensitivity. The low threshold was a response other than "always"; that is, the patient admitted to at least some noncompliance. The high threshold was a response of either "missed many" doses or had stopped the medication completely.

Conformance Rates. Of the inpatients, 48.3 percent met the low threshold, and 7.5 percent met the high threshold. Among those meeting the low threshold, 13.3 percent were receiving depot antipsychotic medication. Among those meeting the high threshold, 50 percent were receiving depot therapy. Similarly, 40 percent of outpatients met the low threshold for noncompliance, and 5.4 percent met the high threshold. Only 24.8 percent at the low threshold were receiving depot therapy, and 35 percent of those meeting the high threshold were prescribed depot medication. With the small numbers of patients on depot medication, comparisons across subgroups lacked power, and analyses revealed no significant relationships.

Return to Contents
Proceed to Next Section