Case Management of the Elderly in a Health Maintenance Organization: The Implications for Program Administration Under Managed Care

EXECUTIVE SUMMARY Kaiser Permanente initiated a two-year demonstration ambulatory case management program in its Ohio region to evaluate five outcomes: perceived health status, functional status, and satisfaction with care, service use, and service costs. Expected results were not consistently obtained for the five outcome measures. Treatment group members did not, however, experience the functional status impairments or decline in health status perceptions reported by the control group during the study period. The unexpected finding that costs were not affected may be attributed to the type of case management intervention used in the demonstration program. This study is broadly applicable to managed care settings facing the challenge of developing programs to minimize the risk for bearing the costs of the Medicare beneficiaries' overall health when all services are not covered. Managed care administrators should be favorably disposed to implementing a case management model with the potential for affecting functional status, the most significant predictor of expensive continuing care for this cohort of Medicare beneficiaries, while working to develop more effective protocols and resource control strategies. INTRODUCTION Health maintenance organizations (HMO) have used case management—a care model whose precise definition depends on an organization's perspective and goals—for many years as a cost-containment strategy for elderly, high-risk enrollees. Even though HMOs are the most common form of managed care used by Medicare enrollees (U.S. GAO 1995, 1996), few have undertaken formal, rigorous analyses of these programs (Kramer, Fox, and Morgenstern 1992; Pacala et al. 1995). Evidence of case management's effectiveness in containing costs or achieving other targeted outcomes remains largely anecdotal (Pacala et al. 1995; DeBusk et al. 1994; Capitman 1988). As a result, the findings from previous research are limited in their capacity to provide a basis for further inquiry or justification for case management as part of a benefit package for the elderly or use as a utilization management strategy (Capitman 1988; Zawadski and Eng 1988; Pacala et al. 1995). Previous studies of case management for elderly enrollees in HMOs range from targeting a single diagnosis in an inpatient setting, such as myocardial infarction (DeBusk et al. 1994), to developing, coordinating, and implementing a care plan for a comprehensive array of covered acute and long-term care services as in the social health maintenance organization (S/HMO) (Yordi 1988). Studies evaluating the effect of case management within an HMO usually focus on the S/HMO (Hicks et al. 1981; Leutz et al. 1988). S/HMO demonstration projects have shown that case management, when combined with a variety of prepaid long-term care alternatives, can reduce costs attributable to hospitalization and nursing home placement but result in increased aggregate costs (Yordi 1988; Zawadski and Eng 1988). These findings are, however, not applicable to the majority of managed care systems because of the absence of a long-term care benefit. This may well be the first investigation to adopt a comprehensive approach to case management that included a cost analysis in an HMO setting without a long-term care benefit (Pacala et al. 1995; Riley 1995). Thus, the study presented here has significant operations implications for the vast majority of managed care programs providing care to Medicare beneficiaries. In this article we describe and evaluate a case management model designed for individuals 75 years of age and older who are enrolled in the Ohio region of Kaiser Permanente, the nation's largest federally qualified HMO. METHODS Background/Research Design/Setting/ Subject Selection Aware that the unique needs of its elderly membership required a different approach to care, in 1991 the Kaiser Permanente Medical Care Program's Interregional Committee on Aging issued a request for proposals to implement demonstration programs designed to improve service delivery for enrollees 65 years of age and older. All demonstration programs were required to develop approaches to test the Interregional Committee's newly developed Model of Care, a framework for organizing and providing services to its older members (Larson 1991). The Model of Care was designed to respond to the problems that physicians and managers in eight Kaiser Permanente regions identified as barriers to meeting the service requirements of elderly enrollees and emphasized three core components: member screening, assessment, and care coordination. Seven demonstration programs in six Kaiser Permanente regions were funded through the Garfield Memorial Fund. In addition to implementing one or more of the core components of the Model of Care, each of the projects shared another common required focus, implementation in an actual delivery setting with the potential for continuing use in that region or other regions. The Ohio region's two-year funded demonstration project comprised screening, assessment, and care coordination in an ambulatory case management program that encompassed both inpatient and outpatient settings as well as the coordination of covered medical and noncovered nonmedical services, but did not include a prepaid long-term care benefit. This comprehensive model was seen as the first step in designing and implementing a region-wide approach to care for the elderly. The coordinated care model shown in Figure 1 provided a framework for the intervention by emphasizing the development of a care plan to bring together all aspects and components of an enrollee's care. The goal of this demonstration project, a randomized trial, was the elimination of fragmented care, inappropriate utilization, unnecessary costs, and the enrollee confusion frequently associated with coordinating care for chronic ailments. Five outcomes were evaluated: perceived health status, functional status, and satisfaction with care, service use, and service costs. For those enrollees receiving case management, improvements in perceived health status, functional status, and satisfaction with care were expected. A substitution of less expensive outpatient visits for hospital admissions and emergency department visits was also expected. A priori expectations for all other service use and costs were equivocal.FIGURE 1: The Coordinated Care ModelThe medical office providing care to the largest number of enrollees in the 75+ category was selected as the study site. This cohort was selected because, from 1988 through 1990, membership in the over-74 age category grew from 4,522 to 6,363, an increase of 40.7 percent, the greatest rate of growth of any age cohort, as a percentage and as a proportion of total membership. In 1990, the 75+ cohort constituted only 2.8 percent of the total membership. They were, however, responsible for 5.3 percent of provider visits, 10.7 percent of hospital admissions, and 13.8 percent of hospital days, with an average length of stay 1.6 days longer than the regional average for all enrollees. Initially 2,297 Kaiser members aged 75 and over receiving care at the medical offices were mailed the Ohio senior health survey, an abridged version of the Medicare Plus assessment instrument used by Kaiser Permanente's Oregon region in the S/HMO demonstration project. Of this target group, 218 were deceased, permanent residents of nursing homes, had moved out of the service area, or were no longer enrolled in Kaiser Permanente. Two mailings yielded a total of 1,901 responses for a 91 percent response rate from the remaining 2,079 eligible enrollees. Surveys completed by a surrogate were accepted after telephone verification. Using the responses from the returned surveys and utilization data, enrollees were identified as being at high risk for poor outcomes if they met the following eligibility criteria: (1) two or more activities of daily living (ADL) impairments; (2) three instrumental activity of daily living (IADL) impairments; (3) two IADL impairments and one ADL impairment; (4) confined to home; or (5) confined to bed. High utilization was defined by two hospital admissions within a 6-month period or two emergency department (ED) visits within the same month. Length of time as a Kaiser member was not a criterion for eligibility, as previous studies within Kaiser had shown this was not a relevant factor in utilization concerns (Mullooly and Freeborn 1979; Freeborn et al. 1990). A computer-generated table of random numbers was used to assign the study participants (227 who met the severity criteria and 90 who met the hospital admissions or emergency department criteria) to treatment and control groups. Two hundred ninety individuals were enrolled in the study; 140 in the treatment group and 150 in the control group. The difference in numbers for each group is a result of the larger number of refusals to participate by the treatment group (19) as opposed to the control group (8). Individuals who refused to participate were followed through utilization data only. Between the baseline assessment and final (Year 2) reassessment, completed in December of 1994, there were 107 surviving treatment group members and 101 surviving control group members. The surviving treatment group members were transferred into the region's newly established ambulatory case management program based in part on the findings from this study. In addition to the 27 enrollees who refused to participate, 89 other participants did not complete the study. Thirty-four treatment group members died; none withdrew from the study but five left the health plan. Forty-three control group members died and seven left the health plan. Those individuals who refused to participate comprised 7.6 percent of the total number of enrollees randomized initially. Bias is unlikely, however, because the proportionate dropout rate is equivalent. In this article, we present all available data for the 317 selected participants for the two-year study period. The Intervention Two case managers from different disciplines—nursing and social work—with prior geriatric case management experience had individual caseloads, evenly divided. A case manager was assigned to each of the treatment group members. Protocols defining the case management process were developed by the co-principal investigators and a geriatrician who functioned as the physician advisor for the study, and were approved by the physicians at the medical offices. These protocols, defining the type of services the case managers could provide and the scope of their authority, were reviewed and adapted as needed during weekly meetings of the research team with the geriatrician. To facilitate an assessment of this process, case managers identified and kept records of the provider (i.e., physician or nurse practitioner), frequency of contact, and type of service (e.g., education, medical visit) they had either recommended or ordered for each treatment group member during the study period. Initially, the case manager made a home visit to each treatment group member to introduce herself and explain the purpose of the study to the participant, to the family if and to the if this was obtained and a to assessment to not on the survey, such as and was this visit the functional status obtained from the surveys was also and to be for all of the study this first a care plan was developed for each treatment group This plan was approved by the physician and if prior to care physicians were kept of each through individual or weekly geriatric team For more such as participants who from functional and for more than one treatment developed by the case manager were reviewed by an team of during weekly geriatric team The core team included the the case a geriatric nurse a in and a as these geriatric team meetings were for each treatment group treatment plan had reviewed at by the of the study. participants with problems from two or more treatment group medical and nonmedical needs and medical the development of the treatment plan with the implementation the of the case For while the case managers made at one home visit six to most study weekly visits for treatment group members were not on their medical and nonmedical each treatment group member was through or nursing home family and telephone managers medical participants to and met with other to care and treatment as by the medical and treatment plan. managers also for nonmedical services such as care, nursing home family eligibility, and to and from the In to the intervention control group members the services by their without the benefit of a case manager to their care. control group member was by telephone of or in the study, mailed a and provided with a telephone number to use if either the or or family had the study. treatment and control were the same version of the baseline instrument at and after the For the and reported the response were and percent, for the control group. The case manager in or by all treatment group members. group members were by and through made by a nursing when was needed in the or when they to respond to the to evaluate functional status as by and perceived health status, and satisfaction were from the baseline surveys and the and ADL was by five and on a from 1 to with 1 and needed all of the IADL was as a or with was needed and To only and were used in the Health status was on a of 1 to defined as or with care was on a of 1 to with 1 and systems comprised the data for service use and costs. All for each study for the study period were The following services were hospital admissions and length of stay outpatient and and visits nursing home home health visits and medical and care costs were for the and the Kaiser Permanente Ohio Kaiser outpatient and visit costs were based on an average cost visit by medical and provider type or costs were from the Kaiser Permanente Medicare and included and for and costs. and costs for and were by the Ohio region's overall cost of percent, and percent, Kaiser Permanente's average cost for Medicare risk enrollees was obtained from the region's the average study member as the basis for this All costs are in outpatient visits for inpatient was a for management and the outcome To a increase in outpatient provider visits and a in hospital admissions and visits for the treatment group, we that enrollees were needed of a was for outpatient visits, the use of data to for inpatient utilization was not because of in the hospital data for the total regional 75 years of age and older. Thus, for inpatient use, only may be was for the treatment and control only for of the number of these two are in to their baseline of the treatment and control were at baseline the for and the test for A of the the two for each of the outcome at baseline was an analysis of at baseline the were for a potential effect on the outcome during the study period a analysis of A was used to the status of the five outcome within each group over the two-year period. 1 and present all available data for the 317 enrollees to part in the study. These data are presented by the enrollee's study status, as a treatment or control group and to participate as a treatment or control group 1: of the by Health with Care, and by at and treatment and control have baseline with to group, and education, as shown in The study was with less than a high education, and living in during baseline are, however, significant from this group members are more to and are, on older. treatment group members were more to the study a single eligibility while control group members were more to eligibility of these significant at baseline living and eligibility criteria are for potential effect on the five outcome measures. to evaluate the five outcome status, health status, satisfaction with care, service use, and service the baseline period are provided in the of use and costs, no significant were the treatment and control group members at for the five outcome are over the two-year study period for the treatment and control group members. the 1 only the perceived health status outcome is for the treatment group. group members were more with their care, an unexpected A increase in visits and costs for the treatment group members was also the final reassessment, treatment group members were more than the control group, as shown by their ADL and IADL the treatment and control group members at baseline had an effect on outcome measures. the study age and living affected IADL while use and costs were to eligibility criteria a the that case management was not in this inappropriate service use over the two-year study within the treatment and control for the five outcome were also but are not shown in These results the when the two were the control group treatment group members did not experience significant in functional status during the study period or in their perceptions of health Using the that all enrollees in the study are Medicare the total cost of implementing this case management model was as shown in In the study is of both Medicare risk percent, control and Medicare cost percent, control enrollees. costs are defined here as the and benefit of the case managers and the research costs attributable to the program were not available because of the Ohio region's for Treatment group members in of the for and in of the control group members for both should be that intervention costs are more member because participants who died or left were not by other eligible enrollees during the A of the of Care for the Treatment and the two the participants in the to they were all of those who refused to participate in the treatment group in their the of the control IADL the two were their at status, health status, and satisfaction with care are not for those enrollees who refused to participate because data are not These however, their with in use and costs during the first of the study and in hospital admissions and costs during the of the study. Kaiser Permanente initiated this in its Ohio region to evaluate the effectiveness of a comprehensive ambulatory case management program for elderly enrollees 75 years of age and older who were identified as being at high risk for poor outcomes because of functional status impairments or hospital or for the five outcome were over the two-year study period. Expected in satisfaction with care, the substitution of outpatient for inpatient care, and in the cost of care did not for the treatment group to the control group. the treatment and control were a significant decline in the control functional status to the of significant for the treatment group and the treatment perceptions of their health status were the only To further these when each group is treatment group members did not experience the decline in ADL and IADL or in their perceptions of health status for the control group members. should be that a analysis of all inpatient costs that treatment group costs for the two years of the study were than those of the control group because of the of The costs for and were for two treatment group one who died during the study. The cost for was also for a treatment group these inpatient costs for the treatment group are to for and for 1994, and a total cost of care member that did not the for The unexpected finding that case management did not costs may be attributed to the type of case management intervention used in the demonstration program. In addition to the of care the of the intervention the case managers in service and managers made for the for and the to these In other the case managers were for care. is not to with this of are more to be made and is also not to that with this of the intervention into a form of long-term care for those treatment group participants who may have to a nursing home in the absence of the case management program. a of selected a difference in the number of enrollees in each group for to a nursing home during 1994, the the study as many treatment group members reported they were to nursing as control group members To the potential of case management as a cost-containment and service coordination two from the study The implementation of this or another comprehensive case management model in a different type of managed care setting could different management protocols designed to cost or nursing to care within the model here could also different Thus, the setting or the protocols within the setting could different findings from those reported the of research in this area, we not to on the and effectiveness of case management we this case management model can provide a basis for further inquiry and for two this may be the first randomized of case management for in a Medicare service the challenge of developing programs that minimize the risk for bearing the costs of the overall health when all services are not covered. the focus, this may be the first broadly applicable to the vast majority of managed care settings Medicare beneficiaries. the process, and outcome can as a for the and of other case management programs for the elderly in managed care The continuing effect of living and eligibility criteria on IADL use and costs, and on overall cost participant, the of comprehensive of elderly enrollees for medical use and that may utilization and costs The model in Figure 1 can as a to the of a case management program. to baseline and eligibility criteria developed for this study could be used to those enrollees who case geriatric care emphasizing a team approach to care may be used to and evaluate care and to develop a more approach to the case management process to utilization more Two case management programs have used in the Ohio The coordinated care model with its components of screening, assessment, and service coordination was to a more case management based on eligibility criteria such as or the of chronic was made available to enrollees of all are by in the outpatient or at the is case management is comprehensive case management to the model here was only made available for These programs were not to be or As a result, utilization and cost data are not available for to the program This study has three that can also as for administrators to and evaluate the effectiveness of case management though the medical use and cost for the elderly are in Kaiser Permanente an study and single study did not the effect of ambulatory case management for a of elderly Medicare enrolled in As a result, the to the study findings can be used for and in HMOs and other managed care settings can only be through the implementation and of other case management A concerns that not be in a more but frequently be made when demonstration programs are in actual delivery The that the process may have for the treatment and control group functional status and health status from such a a case manager the to treatment group participants was as an part of the process in a setting because concerns the study. administrators may to have outcome by other than the case manager to a more rigorous assessment of a The on the of the results can also be in actual delivery settings because the same are for treatment and control group members. to be effective for the treatment group may have used for the control group and there is to this type of because of the geriatric care should be that in may be to have this effect and the of as a study because of the to and improve managers a more rigorous assessment of the of a case management program may to implement such a program on a management for the elderly aged 75 and should be a management because this is the within the 65 and over cohort in the and is frequently by more functional and high use of medical and services et al. 1990). Thus, this cohort is as to be at risk for chronic and more to in the coordination of care for these and the increase in Medicare risk and a percent increase was GAO continuing in and the response GAO 1996), case management of the elderly on a This is by the number of enrollees with chronic in in Managed care administrators should be favorably disposed to implementing a case management model with the potential for affecting functional the most significant predictor of continuing care for this cohort of Medicare beneficiaries, while working to develop more effective protocols and resource control strategies.