Beyond the Water's Edge: Charting the Course of Managed Care for People with Disabilities - Conference Resource Book. Enrollment Duration, Service Use, and Costs of Care for Severely Mentally Ill Members of a Health Maintenance Organization

11/01/1996

Bentson H. McFarland, M.D., Ph.D.; Richard E. Johnson, Ph.D.; and Mark C. Hornbrook, Ph.D.
Archives of General Psychiatry 53:938-944 (October 1996)

Background: The rapid growth of prepaid health care and the increasing enrollment of Medicaid clients in health maintenance organizations (HMOs) raise concerns about the adequacy of services for persons with severe mental illness in capitated health plans. Uncontrolled studies have suggested that enrollment of HMO members with mental illness may be prematurely terminated.

Methods: We identified 250 adult Kaiser Permanente Northwest Region (Portland, OR) members who were enrolled during 1986 or 1987 and had chart diagnoses of schizophrenia or bipolar disorder. Severely mentally ill subjects were matched by age and sex with control HMO members with and without diabetes mellitus. Records of the HMO and the state mental health agency were reviewed to determine HMO enrollment duration, private and public service utilization, and HMO costs of care during the 4-year follow-up period.

Results: The severely mentally ill subjects had 42 months of HMO enrollment during the follow-up period compared with 37 months for the controls without diabetes mellitus and 47 months for the patients with diabetes mellitus (P<.001). When HMO enrollment prior to the study was taken into account, the severely mentally ill subjects and those with diabetes mellitus had similar membership duration. Among the severely mentally ill subjects, community mental health service use was related to longer duration of HMO enrollment (P<.05) but HMO costs of care per member per month were not related to retention. The severely mentally ill subjects were high users of mental health services but their use of general medical care was similar to that of the controls without diabetes mellitus.

Conclusions: This controlled study found no evidence for early termination of HMO members with costly mental illness. Use of community mental health care was associated with longer duration of HMO enrollment.

The dramatic growth in health maintenance organization (HMO) enrollment has heightened concern about the adequacy of treatment available for persons with severe mental illness in prepaid systems.1, 2 This topic is of particular interest to the dozens of states3, 4 that are now in the process of replacing fee-for-service with capitated health care systems for Medicaid clients, many of whom have severe mental disorders.5 Indeed, Mechanic6 has suggested that public mental health programs should be gradually integrated into the larger, prepaid health care system while bearing in mind the many challenges involved.7 Conversely, Scheffler et al.8 have recommended that programs for persons with severe mental illness remain "carved out" of general health care. Furthermore, editorial writers have claimed that traditional HMOs "disenroll individuals who develop serious mental disorders"9 and have stated that "HMOs have routinely excluded any coverage of chronic mental illness."10 On the other hand, HMOs have also been described in which persons with severe mental illness "receive relative priority."11 Inconveniently, there have been few empirical data with which to inform this debate.1, 2

One of the few pertinent studies is the 1987 Minnesota project, which included a comparison of health status for chronically mentally ill Medicaid clients who had been randomized to fee-for-service vs prepaid (HMO) health care.2, 12, 13, 14 Unfortunately, the project ended prematurely after only a year of operation. Few if any differences were found for chronically mentally ill persons, although the subset of this group with schizophrenia may have been adversely affected by assignment to the independent practice association model HMOs that participated in the project.2, 14

Somewhat related to this issue is the RAND observational Medical Outcomes Study,15, 16 which raised the possibility that during 1986 through 1988 psychiatric patients with major depressive disorder in prepaid health care may have switched insurance coverage (i.e., terminated HMO enrollment) sooner than their counterparts in the fee-for-service sector. It was suggested that the limited mental health services provided to these subjects (referred to as HMO "skimping") may have contributed to their departure from the HMO.16, 17 However, the Medical Outcomes Study lacked a control group within the HMO.

To address these issues, we conducted a multiyear longitudinal cohort study of HMO members with severe mental illnesses such as schizophrenia or bipolar disorder. Based on the existing literature,9, 15 we hypothesized that the severely mentally ill HMO members would disenroll earlier than their lower-cost counterparts.9, 10 We also wished to learn if there was a relation between HMO members' use of public mental health services and their duration of enrollment.

 

METHODS

Study Site

The study was conducted in the Northwest Region of Kaiser Permanente, a nonprofit, prepaid, group-practice HMO that currently serves some 385000 members in greater Portland, OR. The HMO has been in operation for over 50 years, provides comprehensive medical benefits, and includes a specialty mental health department presently consisting of about 20 psychiatrists and 80 other mental health professionals.18 The HMO's mental health and substance abuse benefits conform to those mandated by Oregon law. Since 1987, Oregon insurers have been required to cover up to $2000 of outpatient and up to $8500 of inpatient, residential, or day treatment mental health and/or substance abuse services every 24 months for adult beneficiaries. The HMO allows substitution of inpatient for outpatient benefits. However, the total adult mental health and/or substance abuse benefit is a maximum of $10500 per 24 months. In addition, the vast majority of HMO members have a pharmacy benefit. A 1995 survey of the membership showed that 84% of enrollees obtain all of their prescriptions and 12% obtain some or most of their prescriptions (including those written by non-HMO clinicians) at the HMO's pharmacies. The HMO maintains a membership information processing system that records eligibility for services based on monthly premium payments. Administrative personnel attempt to contact individuals whose premiums are unpaid. For purposes of this study, disenrollment was defined to have occurred at the beginning of a 90-day or longer period of ineligibility. The project was reviewed and approved by the Kaiser Permanente Northwest Region Committee for the Protection of Human Subjects on February 16, 1995.

Selection of Subjects

The years 1986 and 1987 were chosen as a baseline period so that this project could be compared with earlier work.2, 12, 13, 14, 15, 16 Because the HMO's outpatient charts were not computerized at the time, severely mentally ill subjects were selected from the 2334 persons who received an antipsychotic drug (excluding prochlorperazine and thiethylperazine, which are used in the HMO only for treatment of nausea and vomiting) or lithium from an HMO pharmacy during 1986 or 1987; individuals in the original group who also received anticancer drugs or drugs used in the treatment of acquired immunodeficiency syndrome were excluded. To minimize the numbers of subjects who might have conditions such as Alzheimer disease, the study focused on the 733 potential subjects who were between ages 10 and 46 years in 1986. Of this group, 526 had mental health department charts (indicating they had had at least 1 contact with an HMO mental health specialist at some time). Individuals were then randomly selected from these 526 persons for mental health chart abstraction. Subjects who carried chart diagnoses of bipolar disorder (including mania, manic-depression, and hypomania) or schizophrenia (including schizophreniform disorder and schizoaffective disorder) in the mental health record were retained in the study. Mental health chart abstraction proceeded through 440 charts until 250 subjects meeting these inclusion criteria were located. Persons who were excluded at this stage typically had diagnoses of substance abuse (primarily amphetamines, cocaine, and/or alcohol) or psychotic depression. Since this study was designed to be descriptive in nature, the sample size of 250 was selected so that the SEs of the mean of annualized utilization estimates (measured in office visits per person per year) would be less than 10% of the estimated mean value. As in other record review projects, diagnoses were assigned based on the majority of those found in the subjects' mental health charts.19, 20 This group of subjects was labeled "cohort 1."

Control Members With and Without Diabetes Mellitus

The severely mentally ill subjects in cohort 1 were then matched with other HMO members. The "pharmacy" controls were taken from the population of HMO members who used the system's outpatient pharmacies during 1986 or 1987 (and had not received an antipsychotic drug or lithium). This group included some two-thirds of the HMO's membership. The "membership" controls were selected from the HMO's enrolled population during 1986 and 1987, which averaged about 290000 persons on any given day at that time. Subjects were matched for sex, year of birth, and "coverage status" (i.e., whether the subject was a subscriber or a dependent of the subscriber). For studies of the enrollment duration of cohort 1, the subjects were also matched for sex and year of birth (within 5 years) with 234 people selected from the 2140 individuals who were HMO members during 1986 or 1987 and who had been discharged by a general medical-surgical unit (in 1986 or 1987) with a diagnosis of diabetes mellitus.

Service Utilization

Cohort 1 subjects' HMO service utilization data were obtained from 1986 through 1990 record reviews and computerized databases. In addition, the names and dates of birth for all subjects without diabetes mellitus in cohort 1 were matched against state mental health agency computerized utilization data. The state agency provided information about subjects' use of community mental health programs and the state mental hospitals from 1986 through 1990. While there were extensive data on state hospital usage (including dates of admission and discharge, diagnoses, and so forth) the community mental health data were limited to enrolled vs not enrolled during particular time periods. The Chronic Disease Score21, 22 (based on nonpsychotropic drug dispensing) was used to gauge the severity of physical illnesses in subjects from cohort 1 without diabetes mellitus.

Costs

The HMO's accounting data and Medicare cost reports became available in 1987 and were used to calculate cost coefficients for each unit of service (e.g., outpatient visit to a provider, day in a medical-surgical unit, and so on). The cost coefficients were then multiplied by units of service for cohort 1 to determine the cost (in 1990 dollars) for each type of care.23, 24 Billing records were used to determine the costs (in 1990 dollars) of services purchased by the HMO (e.g., general hospital inpatient psychiatric care) for members of cohort 1. Public sector costs were not available.

Secular Trends in Enrollment Duration

To address possible secular trends in enrollment duration, a second group of severely mentally ill subjects (labeled "cohort 2") was chosen from the HMO members discharged from a general hospital during 1990 through 1992 with diagnoses of schizophrenia or bipolar disorder. These people were matched for age and sex with HMO controls with and without diabetes mellitus as described for cohort 1. The enrollment duration study focused on the 165 female and 116 male (average age, 31 years in 1990) severely mentally ill subjects in cohort 2 who were between the ages of 10 and 46 years at the time of the index hospital discharge, of whom there were 139 with a diagnosis of schizophrenia and 5 who also had a diagnosis of diabetes mellitus. Follow-up for cohort 2 started with the index hospital discharge and ended December 31, 1995.

Statistical Analysis

Service utilization and cost data for cohort 1 are reported on a per member per month of enrollment basis. In 2-way analyses of variance (ANOVA), data were transformed as needed so that the residuals were roughly normally distributed. For example, the total cost per person per month was transformed by adding 1 to the cost numerator, dividing by the months of enrollment denominator, and then taking the logarithm of that ratio. To account for multiple comparisons, the studentized range test was used to compare the cohort 1 severely mentally ill subjects' utilization and costs with those of the controls.25 Enrollment duration comparisons used the log-rank test and the Cox proportional hazards model stratified to account for the matching.26 Changes in coverage status (subscriber vs dependent) were examined using the Miettinen method.27

Cox proportional hazards models were used to examine factors associated with retention of cohort 1 severely mentally ill subjects in the HMO during the follow-up period from 1986 through 1990.26 Blocks of potential predictors were planned for stepwise inclusion in the proportional hazards models. These potential predictor variables were demographics (age, sex, schizophrenia vs bipolar disorder); enrollment status at the start of the study (subscriber vs dependent, Medicare vs no Medicare, Medicaid vs no Medicaid, years of HMO enrollment prior to the start of the study); and utilization (state hospital admission at any time while an HMO member during the study period, use of community mental health services while an HMO member at any time during the study period, total HMO costs of care per member per month, and HMO mental health costs of care per member per month).

 

RESULTS

Demographics

In cohort 1 there were equal numbers of males and females. Subjects in cohort 1 were (on average) 32 years old in 1986, with an age range from 13 to 45 years and an SD of 8 years. The ethnicity distribution was 80% white, 5% African American, 2% Asian American, 2% Hispanic, and 12% unknown. There were no differences in the distributions of know ethnicity between the severely mentally ill subjects and the controls. Some 30% of the 250 severely mentally ill persons in cohort 1 (for whom data were available) had never married. In contrast only 12% of the controls (for whom data were available) had never married. At the beginning of the study about half (53%) of the 750 subjects without diabetes in cohort 1 were subscribers, while 44% were dependents and the remaining few were nonmembers. Only 12% of the cohort 1 severely mentally ill subjects changed coverage status from dependent to subscriber during the 4-year follow-up period, compared with 22% of the controls without diabetes mellitus (relative risk, 0.53; 95% confidence interval, 0.30-0.92). The severely mentally ill subjects in cohort 1 were much more likely to have Medicare coverage than the controls without diabetes mellitus (10% vs 0.2%, P<.001 by Fisher exact test). There was no difference in the prevalence of Medicaid coverage (5% vs 4%).

The 250 severely mentally ill subjects in cohort 1 had lengthy histories of mental illness. At the time of their first HMO mental health department contact, the vast majority (73%) reported having had at least 1 previous psychiatric hospitalization, with 40% of severely mentally ill subjects reporting 3 or more admissions. Many (41%) were known to have been admitted to a state mental hospital in Oregon. The majority (57%) had had contact with the HMO's emergency psychiatric service at some time during their enrollment.

Diagnoses

The diagnostic algorithm showed that 79 (32%) of the 250 severely mentally ill subjects in cohort 1 had chart diagnoses of schizophrenia, 98 (39%) had bipolar disorder, and the remaining subjects had multiple diagnoses. As expected, 92% of the 98 persons with bipolar disorder had received prescriptions for lithium, while 93% of the 79 subjects with schizophrenia had been dispensed antipsychotic drugs. Some 30% of the bipolar subjects had received antipsychotic drugs as well as lithium.

Service Use

Health care utilization data for the 1986 through 1990 study period are presented in Table 1. Not surprisingly, the severely mentally ill HMO members utilized greater amounts of services than did the controls without diabetes mellitus. As expected, the severely mentally ill subjects had greater per member per month use of mental health [F(2498)=120.1, P<.001] and substance abuse outpatient services [F(2498)=6.5, P<.003] as well as greater use of general hospital psychiatric inpatient services [F(2498)=86.3, P<.001]. During the study period, 88 (35%) of the severely mentally ill subjects in cohort 1 had 274 general hospital psychiatric admissions while the pharmacy controls had none and the membership controls had 1.

Interestingly, there were no statistically significant differences among the 3 groups without diabetes mellitus in the per member per month use of general medical outpatient services. However, differences were observed with respect to use of general medical-surgical inpatient care [F(2498)=4.5, P=.01]. The studentized range test indicated that the severely mentally ill subjects' use of general medical-surgical inpatient care was equivalent to that of the pharmacy controls.

The Chronic Disease Score (based on pharmacy data other than psychotropic drugs for the 665 subjects without diabetes mellitus enrolled in 1986) showed that the severely mentally ill subjects had the highest score (mean=0.48, SD=1.18), followed by the pharmacy controls (mean=0.31, SD=0.92), who were in turn followed by the membership controls (mean=0.15, SD=0.65). These differences are highly statistically significant (2-way ANOVA F[2498]=7.9, P<.001). Because the Chronic Disease Score was not normally distributed, we also examined the percentage of each group with a nonzero score (47 [20%] of 231 severely mentally ill subjects, 34 [15%] of 225 pharmacy controls, and 16 [8%] of 209 membership controls). These frequency differences were also highly statistically significant (X2=14.3, df=2, P<.001).

During their HMO enrollment in the follow-up period, 30 (12%) of the severely mentally ill subjects in cohort 1 were admitted to a state mental hospital compared with 3 (1%) of the pharmacy controls and 1 (0.4%) of the membership controls (X2=48.5, df=2, P<.001). Similarly, 101 (41%) of the severely mentally ill subjects in cohort 1 used community mental health services during the follow-up period compared with 12 each (5%) among the pharmacy and membership controls (X2=152.7, df=2, P<.001).

Costs

Table 2 and Table 3 show the HMO costs of care per member per month for the subjects without diabetes mellitus in cohort 1. The vast majority (98% of the severely mentally ill subjects, 95% of the pharmacy controls, and 87% of the membership controls) incurred HMO costs. The severely mentally ill subjects had substantially higher HMO costs of care per member per month of enrollment than did the controls without diabetes mellitus. The average cost for the subjects with severe mental illness was $380 (median of $203) per member per month vs an average of $149 (median of $33) for the pharmacy controls and $90 (median of $23) for the membership control subjects [2-way ANOVA on the transformed total cost F(2419)=81.4, P<.001]. The studentized range test showed that the 3 groups were all statistically significantly different from one another at the P=.05 level.

Looking in more detail at costs per member per month for the severely mentally ill subjects showed that the median combined mental health cost (inpatient, outpatient, and pharmaceutical) was $99 with the median, excluding psychotropic pharmaceuticals, at $74. Median outpatient mental health cost was $48. The 90th percentile figures were $798 per member per month for total costs, $544 for all mental health costs, $492 for mental health costs excepting psychotropic medications, $284 for inpatient care, and $214 for outpatient mental health costs.

Enrollment Duration

The Kaplan-Meier product-limit estimates in the Figure show the retention of cohort 1 subjects in the HMO from the start of time under observation (in 1986 and 1987) until disenrollment or the end of follow-up on December 31, 1990. Mean enrollment duration for cohort 1 is shown in Table 4. The enrollment durations of the cohort 1 groups differ significantly (log-rank X2=40.7, df=3, P<.001). The stratified Cox model showed that the most powerful predictor of enrollment duration was not being in the membership control group (X2=25.1, df=1, P<.001), with the next most powerful predictor being years of HMO enrollment prior to entering the study (X2=17.8, df=1, P<.001). Once prior years of HMO enrollment had been taken into account, there were no statistically significant enrollment differences among those subjects with diabetes mellitus, severe mental illness, and pharmacy controls in cohort 1. Table 4 also shows that the enrollment duration of cohort 2 subjects was similar to that of cohort 1.

Retention

Among the severely mentally ill cohort 1 subjects, stepwise Cox proportional hazards modeling showed that the factors related to longer duration of enrollment in the HMO were years of HMO enrollment prior to the start of the study period (X2=7.6, df=1, P<.006); age (X2=5.6, df=1, P<.02); and community mental health service use (X2=3.9, df=1, P<.05).

Costs of care (total costs per member per month or mental health costs per member per month) for cohort 1 were not significantly related to retention based on the Cox proportional hazards modeling. Other factors not significantly related to cohort 1 enrollment duration in the proportional hazards models included sex, schizophrenia vs bipolar disorder, subscriber status, or use of the state hospital.

 

COMMENT

Results from this study need to be interpreted in light of its design. The project was not a randomized trial nor did it include a comparison group of subjects outside the HMO. Since the study was designed to take advantage of existing data, subjects were not interviewed and clinical outcomes were not measured. Consequently, there could well have been important but unrecorded differences among the groups. Furthermore, subjects' reasons for disenrollment and for use of public mental health services were not available.

The project examined "prevalent" cases of people with severe mental illness who were using HMO services. At the time of this study it was not possible to identify HMO members with newly emerging (i.e., "incident") psychosis. Certainly, the "careers" of severely mentally ill persons who do not receive treatment may well be different from those of the subjects described here. For example, persons who become psychotic and refuse HMO mental health services might be unable to maintain enrollment and quickly leave the organization. Indeed, earlier work has shown that the treated prevalence of schizophrenia within this HMO is less than what would be expected from Epidemiologic Catchment Area data, although the treated prevalence of bipolar disorder is comparable to the national estimate.18 Very recent improvements in the HMO's automated data systems may provide an opportunity to conduct an incidence study focusing on people with newly emerging psychosis.

Another issue is the degree of severity of the subjects' mental disorders. For example, some 67% of the cohort 1 severely mentally ill subjects were self-reported to be employed at the time of their first HMO mental health clinic visit. Interestingly, the Epidemiological Catchment Area project found that 43% of the persons identified in that study as having schizophrenia were employed.28 The severely mentally ill HMO members may be that subset of persons with conditions like schizophrenia, who have a relatively good prognosis.29, 30, 31, 32, 33, 34, 35, 36, 37

Nonetheless, the frequent use of emergency and inpatient psychiatric services for this population suggests that many of these individuals were, indeed, severely disabled. Furthermore, the Chronic Disease Score indicated that the severely mentally ill persons appeared to have had physical as well as mental health problems. These individuals were also much more likely to have Medicare coverage than the controls without diabetes mellitus. Presumably, the severely mentally ill subjects became eligible for Medicare coverage by virtue of qualifying for Social Security Disability Insurance due to their mental illness.38 The relatively low rate of Medicaid participation by the severely mentally ill subjects in cohort 1 may well have been due to state policies at the time of the study, which, in effect, deemed persons receiving Social Security Disability Insurance to be "too wealthy" for Medicaid.

Relatively few of the severely mentally ill subjects in cohort 1 (compared with the controls without diabetes mellitus) changed coverage status from dependent to subscriber during the 4-year follow-up period. One explanation for these findings is that the severely mentally ill subjects who entered the study as dependents were not as likely as their matched controls to obtain competitive employment (and thereby become subscribers in their own right). Indeed, naturalistic follow-up studies of patients with mania suggest that significant disability would be expected for at least some of those severely mentally ill HMO subjects who had bipolar disorder.39, 40

Its limitations notwithstanding, this study showed that HMO members with severe mental illness had enrollment duration longer than that of controls without diabetes mellitus but somewhat shorter than that of members with diabetes mellitus. Furthermore, costs to the HMO were unrelated to duration of enrollment. To the authors' knowledge, cohort 1 has been followed up longer than any group of mainstream managed care beneficiaries with severe mental illness. This study is also one of the few that measured both private and public mental health service use.13 In contrast to the Medical Outcomes Study,15, 16, 17 this project involved a variety of HMO control subjects.

It is worthwhile examining the factors that did and did not explain the severely mentally ill subjects' retention within the HMO. There was no support for the contention that HMO members were "disenrolled" due to severe mental illness.9 Of course, as expected in a "prevalence" study, subjects with very brief periods of enrollment were unlikely to be included in the sample. Consequently, length of HMO eligibility prior to the study was a good predictor of enrollment duration during follow-up. Indeed, when length of enrollment before the study period was included in the Cox proportional hazards analysis, the severely mentally ill subjects had retention times longer than the membership controls but equivalent to that of the diabetic subjects and the pharmacy controls.

We were also unable to find evidence that this HMO "routinely excluded any coverage of chronic mental illness."10 Indeed, the severely mentally ill subjects in cohort 1 were provided amounts of service that generated costs to the HMO several times that of the membership controls. This cost difference was accounted for chiefly by mental health care. Based on the HMO's cost data, it appears that 36% of the severely mentally ill subjects in cohort 1 exceeded the state-mandated outpatient mental health benefit of $2000 per 24 months. Psychiatric inpatient costs were generally less than the state-mandated $8500 per 24 months, but 9% of severely mentally ill cohort 1 subjects did exceed the benefit limit. Looking at combined inpatient and outpatient mental health costs showed that 12% of severely mentally ill cohort 1 subjects exceeded the $10500 per 24 months limit. Of course, one could challenge the accuracy of the cost data. However, it should be noted that some of the costs (e.g., general hospital inpatient psychiatric services) represent payments from the HMO to its vendors. In any event, it seems clear that coverage was provided to HMO members who were severely mentally ill. Furthermore, HMO costs were not related to enrollment duration.

An important issue is the HMO's policies toward serving persons with severe mental illness. As with many HMOs, this organization's mental health services were theoretically limited to treatment of conditions that, in the judgment of the attending physician, were subject to significant improvement through relatively short-term therapy.41 In practice, as demonstrated by these results, mental health services were provided to persons with chronic conditions. Since this approach to persons with severe mental illness may not be found in other HMOs, these results may have limited generalizability.42, 43, 44

Indeed, the distinctions among HMOs44 may explain the apparent discrepancy between the retention data from this project and the implication from the Medical Outcomes Study15 that severely mentally ill subjects would have a shorter enrollment than healthier members. It should be noted that the Medical Outcomes Study was conducted in several prepaid settings (including a traditional staff model HMO), with the poorest outcomes for depressed psychiatric patients observed in independent practice associations.17 Differences between the independent practice association approach to severe mental illness and that provided by traditional HMOs could be responsible for the disparate outcomes observed in the 2 studies. As Judith L. Feldman, MD, remarked: "When you've seen one HMO you've seen one HMO" (oral communication, 1988).

The integrated service delivery system provided by traditional HMOs may be of particular value for severely mentally ill members who have physical as well as mental health problems, as suggested by our data. It is interesting to note that the costs of general medical-surgical care for severely mentally ill subjects were similar to those of the pharmacy controls even though the Chronic Disease Score suggested that the former had more physical illness than the latter. An integrated system might be more efficient than a mental health "carve-out" for people with both physical and severe mental health problems. On the other hand, while the data from cohort 2 suggest that this HMO is continuing to serve severely mentally ill members, the now fiercely competitive health care environment45 makes one wonder if any HMO will be able to provide the level of mental health service described here.

It should be pointed out that the HMO was by no means the sole provider of mental health care to these individuals. Nearly half of the severely mentally ill subjects in cohort 1 also used community mental health services. Furthermore, the use of community mental health care was associated with longer duration of HMO enrollment. While this observational study cannot determine causality, it is conceivable that the subjects who maintained their HMO membership were also to optimize use of both private and public services. One might imagine that the HMO's expertise in areas such as psychopharmacology, emergency psychiatric services, and inpatient psychiatric care could complement the public mental health sector's capabilities in fields such as rehabilitation and vocational training. Unfortunately, shrinkage of public sector mental health funds combined with private sector competition may leave persons with severe mental illness struggling to find appropriate care.46 Nonetheless, there may be considerable value in studying ways HMOs and community mental health agencies can work together to offer an efficiently integrated package of services that will benefit people with severe mental illness.6

 

NOTES

From the Kaiser Permanente Center for Health Research (Drs McFarland, Johnson, and Hornbrook) and Oregon Health Sciences University (Dr. McFarland), Portland.

Accepted for publication June 21, 1996.

Supported in part by grants P50 MH43458, R01 MH45015, and K02 MH01238 from the National Institute of Mental Health, Bethesda, MD.

Presented in part at the 147th Annual Meeting of the American Psychiatric Association, Philadelphia, PA, May 25, 1994, and at the 123rd Annual Meeting of the American Public Health Association, San Diego, CA, November 1, 1995.

Reprints: Bentson H. McFarland, MD, PhD, Center for Health Research, Kaiser Permanente, Northwest Region, 3800 North Kaiser Center Drive, Portland, OR 97227.

REFERENCES

  1. McFarland, B.H. "Health Maintenance Organizations and Persons with Severe Mental Illness." Community Mental Health Journal, 1994;30:221-224.
  2. Christianson, J.B. and F.C. Osher. "Health Maintenance Organizations, Health Care Reform, and Persons with Serious Mental Illness." Hosp Community Psychiatry, 1994; 45:898-905.
  3. General Accounting Office. Medicaid Expansions. Gaithersburg, MD: US General Accounting Office; 1993; Report #HRD-91-78.
  4. General Accounting Office. Medicaid Section 1115 Waivers. Washington, DC: US General Accounting Office; 1995; Report #HEHS-96-44.
  5. Taube, C.A., H.H. Goldman, and D. Salkever. "Medicaid Coverage for Mental Illness: Balancing Access and Costs." Health Affairs (Millwood), 1990; 9:5-18.
  6. Mechanic, D. "Integrating Mental Health Into a General Health Care System." Hosp Community Psychiatry, 1994; 45:893-897.
  7. Mechanic, D., M. Schlesinger, and D.D. McAlpine. "Management of Mental Health and Substance Abuse Services: State of the Art and Early Results." Millbank Quarterly, 1995; 73:19-55.
  8. Scheffler, R., C. Grogan, B. Cuffel, and S. Penner. "A Specialized Mental Health Plan for Persons with Severe Mental Illness Under Managed Competition." Hosp Community Psychiatry, 1993; 44:937-942.
  9. Sharfstein, S.S. "Capitation Versus Decapitation in Mental Health Care." Hosp Community Psychiatry, 1994; 45:1065.
  10. Schlesinger, M. "Perspectives: Ethical Issues in Policy Advocacy." Health Affairs (Millwood), 1995; 14:23-29.
  11. Sabin, J. "Perspectives: Organized Psychiatry and Managed Care: Quality Improvement or Holy War?" Health Affairs (Millwood), 1995; 14:32-33.
  12. Christianson, J.B., N. Lurie, M. Finch, and I. Moscovice. "Mainstreaming the Mentally Ill in HMOs." In: D. Mechanic and L.J. Aiken editors, Paying for Services: Promises and Pitfalls of Capitation, San Francisco, CA: Jossey-Bass, Inc.; 1989:19-28.
  13. Christianson, J.B., N. Lurie, M. Finch, I. Moscovice, and D. Hartley. "Use of Community-Based Mental Health Programs by HMOs: Evidence from a Medicaid Demonstration." American Journal of Public Health, 1992; 82:790-796.
  14. Lurie, N., I.S. Moscovice, M. Finch, J.B. Christianson, and M.K. Popkin. "Does Capitation Affect the Health of the Chronically Mentally Ill? Results from a Randomized Trial." Journal of the American Medical Association, 1992; 267:3300-3304.
  15. Sturm, R., E.A. McGlynn, L.S. Meredith, K.B. Wells, W.G. Manning, and W.H. Rogers. "Switches Between Prepaid and Fee-for-Service Health Systems Among Depressed Outpatients: Results from the Medical Outcomes Study." Medical Care, 1994; 32:917-929.
  16. Sturm, R. C.A. Jackson, L.S. Meredith, W. Yip, W.G. Manning, W.H. Rogers, and K.B. Wells. "Mental Health Care Utilization in Prepaid and Fee-for-Service Plans Among Depressed Patients in the Medical Outcomes Study." Health Services Research, 1995; 30:319-340.
  17. Rogers, W.H., K.B. Wells, L.S. Meredith, R. Sturm, and A. Burnam. "Outcomes for Adult Outpatients with Depression Under Prepaid or Fee-for-Service Financing." Archives of General Psychiatry, 1993; 50:517-525.
  18. Johnson, R.E., and B.H. McFarland. "Treated Prevalence Rates of Severe Mental Illness Among HMO Members." Hosp Community Psychiatry, 1994; 45:919-924.
  19. Lurie, N., M. Popkin, M. Dysken, I. Moscovice, and M. Finch. "Accuracy of Diagnoses of Schizophrenia in Medicaid Claims." Hosp Community Psychiatry, 1992; 43:69-71.
  20. Glazer, W.M., C.D. Pino, and D. Quinlan. "The Reassessment of Chronic Patients Previously Diagnosed as Schizophrenic." Journal of Clinical Psychiatry, 1987; 48:430-434.
  21. VonKorff, M., E.H. Wagner, and K. Saunders. "A Chronic Disease Score from Automated Pharmacy Data." Journal of Clinical Epidemiology, 1992; 45:197-203.
  22. Johnson, R.E., M.C. Hornbrook, and G.A. Nichols. "Replicating the Chronic Disease Score (CDS) from Automated Pharmacy Data." Journal of Clinical Epidemiology, 1994; 47:1191-1199.
  23. Mullooly, J.P., M.D. Bennett, M.C. Hornbrook, W.H. Barker, W.W. Williams, P.A. Patriarca, and P.H. Rhodes. "Influenza Vaccination Program for Elderly Persons: Cost-Effectiveness in a Health Maintenance Organization." Ann Intern Med, 1994; 121:947-952.
  24. Hornbrook, M.C., and M.J. Goodman. "Assessing Relative Health Plan Risk with the RAND-36 Health Survey." Inquiry, 1995; 32:56-74.
  25. Armitage, P. and G. Berry. Statistical Methods in Medical Research, second edition. Oxford, England: Blackwell; 1987.
  26. Collet, D. Modelling Survival Data in Medical Research. London, England: Chapman and Hall; 1994.
  27. Miettinen, O. "Estimation of the Relative Risk from Individually Matched Series." Biometrics, 1970; 26:75-86.
  28. Keith, S.J., D.A. Regier, and D.S. Rae. "Schizophrenic Disorders." In: L.N. Robins and D.A. Regier, editorsPsychiatric Disorders in America: The Epidemiologic Catchment Area Project. New York, NY: Free Press; 1991:33-52.
  29. Breier, A., J.L. Schreiber, J. Dyer, and D. Pickar. "National Institute of Mental Health Longitudinal Study of Chronic Schizophrenia." Archives of General Psychiatry, 1991; 48:239-246.
  30. Breier, A., J.L. Schreiber, D. Pickar, and J. Dyer. "Long-Term Outcome in Chronic Schizophrenia." Archives of General Psychiatry, 1992; 49:503.
  31. Harding, C.M., and J.S. Strauss; Editorial. "How Serious is Schizophrenia? Comments on Prognosis." Biol Psychiatry, 1984; 19:1597-1600.
  32. Harding, C.M., J. Zubin, and J.S. Strauss. "Chronicity in Schizophrenia: Fact, Partial Fact, or Artifact?" Hosp Community Psychiatry, 1987; 38:477-486.
  33. Harding, C.M., G.W. Brooks, T. Ashikaga, J.S. Strauss, and A. Breier. "The Vermont Longitudinal Study of Persons with Severe Mental Illness, I: Methodology, Study Sample, and Overall Status 32 Years Later." American Journal of Psychiatry, 1987; 144:718-726.
  34. Harding, C.M., G.W. Brooks, T. Ashikaga, J.S. Strauss, and A. Breier. "The Vermont Longitudinal Study of Persons with Severe Mental Illness, II: Long-Term Outcome of Subjects Who Retrospectively Met DSM-III Criteria for Schizophrenia." American Journal of Psychiatry, 1987; 144:727-735.
  35. Marengo, J., M. Harrow, J. Sands, and C. Galloway. "European Versus US Data on the Course of Schizophrenia." American Journal of Psychiatry, 1991; 148:606-611.
  36. McGlashan, T.H. "A Selective Review of Recent North American Long-Term Follow-up Studies of Schizophrenia." Schizophr Bull, 1988; 14:515-542.
  37. Schwartz, F., K.G. Terkelsen, and T.E. Smith. "Long-Term Outcome in Chronic Schizophrenia." Archives of General Psychiatry, 1992; 49:502.
  38. McFarland, B.H.; Commentary. In: P. Backlar, editor The Family Face of Schizophrenia. New York, NY: Tarcher/Putman; 1994.
  39. Harrow, M., J.F. Goldberg, L.S. Grossman, and H.Y. Meltzer. "Outcome in Manic Disorders: A Naturalistic Follow-up Study." Archives of General Psychiatry, 1990; 47:665-671.
  40. Maj, M., R. Pirozzi, and D. Kemali. "Long-Term Outcome of Lithium Prophylaxis in Bipolar Patients." Archives of General Psychiatry, 1991; 48:772.
  41. Lubotsky-Levin, B., and J.H. Glasser. "Comparing Mental Health Benefits, Patterns, and Costs." In: J.L. Feldman and R.J. Fitzpatrick, editors Managed Mental Health Care. Washington, DC: American Psychiatric Press; 1992.
  42. Fink, P.J., and W.R. Dubin. "No Free Lunch: Limitations on Psychiatric Care in HMOs." Hosp Community Psychiatry, 1991; 42:363-365.
  43. Westermeyer, J. "Problems with Managed Psychiatric Care Without a Psychiatrist-Manager." Hosp Community Psychiatry, 1991; 42:1221-1224.
  44. Moldawsky, R.J. "In Defense of HMOs." Hosp Community Psychiatry, 1992; 43:81-82.
  45. Jellinek, M.S., and B. Nurcombe. "Two Wrongs Don't Make a Right: Managed Care, Mental Health, and the Marketplace." Journal of the American Medical Association, 1993; 270:1737-1739.
  46. McFarland, B.H. "Ending the Millennium: Commentary on HMOs and the Seriously Mentally Ill: A View from the Trenches." Community Mental Health Journal, in press.
TABLE 1. Service Utilization*
Number (Percent)
Service Severely Mentally Ill Subjects
(n=250)
Pharmacy Controls
(n=250)
Membership Controls
(n=250)
Inpatient Admissions
- Medical-surgical
- Psychiatry
- State hospital
 
11 (27)
31 (66)
6.1 (32)
 
15 (48)
0
0.88 (11)
 
5.9 (18)
0.096 (1.5)
0.068 (1.1)
Outpatient Visits
- General medical
- Mental health
- Substance abuse
 
770 (1010)
460 (560)
27 (110)
 
770 (1130)
8.7 (46)
7.3 (64)
 
610 (920)
18 (140)
4.9 (32)
Used Community Mental Health Program 101 (40) 12 (5) 12 (5)
* Services per 1000 member-months of health maintenance organization enrollment. All data are presented as mean (SD) unless otherwise indicated. Data were collected from 1986 through 1990.

 

TABLE 2. Costs of Care, 1987 Through 1990
Number (Percent)
Service Mean (SD) Cost Per Member Per Month, 1990 $
Severely Mentally Ill Subjects
(n=225)
Pharmacy Controls
(n=223)
Membership Controls
(n=218)
Inpatient
- Medical-surgical
- Psychiatry
- Substance abuse
 
59 (182)
118 (317)
2 (12)
 
77 (531)
0
0.1 (1)
 
35 (147)
0.4 (6)
0
Outpatient
- General medical
- Mental health
- Substance abuse
 
59 (69)
94 (138)
7 (32)
 
55 (83)
4 (24)
2 (16)
 
43 (67)
3 (19)
1 (8)
Pharmaceutical
- General medical
- Psychiatric
- Substance abuse
 
14 (25)
28 (39)
0.0004 (0.005)
 
9 (19)
0.4 (2)
0.03 (0.4)
 
6 (11)
1 (5)
0
Total 380 (473) 149 (592) 90 (194)

 

TABLE 3. Subjects With Nonzero Costs*
Number (Percent)
Service Severely Mentally Ill Subjects
(n=225)
Pharmacy Controls
(n=223)
Membership Controls
(n=218)
Inpatient
- Medical-surgical
- Psychiatry
- Substance abuse
 
76 (34)
88 (39)
4 (2)
 
49 (22)
0
2 (1)
 
34 (16)
1 (1)
0
Outpatient
- General medical
- Mental health
- Substance abuse
 
212 (94)
192 (85)
36 (16)
 
208 (93)
13 (6)
7 (3)
 
183 (84)
17 (8)
8 (4)
Pharmaceutical
- General medical
- Psychiatric
- Substance abuse
 
205 (91)
205 (91)
1 (1)
 
198 (89)
44 (20)
1 (1)
 
159 (73)
35 (16)
0
Total 220 (98) 212 (95) 189 (87)

 

TABLE 4. Duration of Enrollment in Days*
  Cohort 1
(1986-1990)
Cohort 2
(1990-1995)
Subjects with diabetes mellitus
Severely mentally ill subjects
Pharmacy controls
Membership controls
1424 (39)
1263 (45)
1236 (45)
1023 (47)
1437 (40)
1298 (48)
---
---
* Data from Kaplan-Meier survival distribution function. Data are given as mean (SE); ellipses indicate not applicable.

 

 

View full report

Preview
Download

"96cfpk01.pdf" (pdf, 147.17Kb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®

View full report

Preview
Download

"96cfpk02.pdf" (pdf, 45.07Kb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®

View full report

Preview
Download

"96cfpk04.pdf" (pdf, 50.53Kb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®

View full report

Preview
Download

"96cfpk05.pdf" (pdf, 52.03Kb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®

View full report

Preview
Download

"96cfpk06.pdf" (pdf, 1.02Mb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®