A proposal for care of the frail elderly in assisted living homes in 1995.

This proposal was submitted for approval to the administration of UCI Medical Center in 1995. The faculty medical group was enthusiastic and the chair of the Family Medicine program, Joe Scherger, was equally enthusiastic. This came up because, for whatever reason, I was seen at that time as a potential ally of assisted living administrators who were having trouble with HMOs. I was interested in health policy and they were trying to deal with HMOs that would not provide on-site care for residents who were often of limited mobility. The license of these assisted living facilities required that they provide transportation to medical appointments for residents within 8 miles. The HMOs would not provide such simple on-site treatment as podiatry care. They were looking for help.

I came up with the following proposal, which was approved by the faculty medical group and the Department of Family Medicine. It was disapproved by the administrator of UCIMC. His reasoning,as I understood it, was that frail elderly were more expensive to care for. The proposal tries to refute that myth.


The basic premise of Health Maintenance Organizations, as envisioned by the original legislation in 1974, is the preventive maintenance of good health in a basically healthy population. The recent application of this model to Medicare beneficiaries was intended to bring under control the escalating costs of the Medicare program and to “manage” the health of an older population with a higher incidence of chronic disease. In this instance, “managed care” had less to do with “health maintenance” and more to do with cost control.

A strong incentive for seniors to join Medicare HMOs has been the existence of pharmacy benefits which pay for a large formulary of expensive medications not covered by Medicare Part B. There has been some evidence of selective marketing to healthy populations by Medicare HMOs but the pharmacy benefit provision, which would tend to attract patients with chronic conditions requiring medication, serves to counter this trend. It may be possible to manage the care of patients with chronic conditions to minimize hospitalizations by preventing exacerbations of chronic diseases and this would qualify as improved care as well as cost effectiveness. Studies of the cost of care for the frail elderly have shown no difference between HMOs and fee-for-service systems in the last year of life in either decedents or survivors. Medicare-Medicaid beneficiaries have much higher costs in the last year of life than either fee-for-service or HMO members. This suggests a difference between the Medicare-Medicaid group and the rest of the Medicare population based on socio-economic and disability status but also suggests the possibility of potential cost reduction. The Medical Outcomes Study found significant differences in health status trends over a four year period based upon the type of health delivery system. Elderly and poor chronically ill patients, whose SF-36 scores were equal at the beginning of the study, showed decline in health status when health care was obtained in a prepaid HMO system as compared to fee-for-service care. These two groups are responsible for a disproportionate share of medical expenditures and therefore at increased risk for cost containment measures by HMOs. The authors believed that the bureaucratic barriers to obtaining care are less easily traversed by these patients than by healthy middle-class subscribers, the group which seems to do well in HMOs.

The high cost Medicare beneficiary

In the fee for service sector of Medicare only 21% of beneficiaries used Part A services in 1995, although 84% of beneficiaries had Part B claims, indicating some covered medical service use. Median total expenditures per beneficiary were $884 per year in 1995, a modest amount indicating the small size of most claims. High expenditures are concentrated in a small percentile of the population. The top 2% of beneficiaries spend as much as the bottom 90% and the top 1% spend 19% of all Medicare expenditures. It is unusual for beneficiaries to remain in the very high expenditure group for more than one year. Patients in this category either die of an illness or recover after therapy and return to the low expenditure category. 47% of beneficiaries will be in the top quintile of expenditures for one year in a nine year study. 30% will be in the top decile for one year but only 10% are in the top decile for two or more years and only 1.2 % are in the top decile for four or more years. Mortality is high in this group and, for the beneficiaries who are in the top decile for two or more years, it exceeds 50% and is higher than that of the very high cost one year group. Individuals who have persistently high costs, while a very small group, account for a disproportionate share of expenditures. It is possible that multi-year high expenditures represent cases in which improved care could save substantial amounts of money. In addition, it is apparent that high cost Medicare beneficiaries, even in the un-managed fee for service sector, are a very small proportion of the total.

The Frail Elderly

The frail elderly, usually described as those over age 85 and those with chronic illness, are estimated to comprise 10% of all Medicare beneficiaries. If all of the multi-year high cost beneficiaries were in this group it would still leave 88% of this group with costs close to the Average Annual Per Capita Cost over a multi-year span. It is unlikely that costs for this group exceed the average for other age groups and studies of end-of-life costs show a decline in costs after age 80 for decedents and after age 90 for survivors.

The existence of pharmacy benefits has attracted a population of chronically ill elderly patients to Medicare HMOs. This population, with very limited functional status, is having difficulty accessing the primary care and rehabilitation services which are needed for their “health maintenance”. Many of these patients are home bound or are living in residential care facilities. In some cases the administrators of residential care facilities are providing medical care and paying for health care services to comply with their own license requirements when the services cannot be obtained from HMO plans. A group of these administrators has held meetings with HMO representatives without any improvement in the situation and has been seeking assistance from others in the community, for example the Health Care Council, in dealing with this problem.

There have been efforts in the past to contract the health care of nursing home patients with a physician group who would employ nurse practitioners and provide physician backup. In several cases the nursing homes were reluctant to enter into such a contract because many referrals to the home come from physicians who might be excluded from the contract and might then refer their elderly patients elsewhere for residential care. These proposed contracts were for fee-for-service care and there might be less resistance on the part of potential referring physicians when HMO patients are involved. There have also been studies on the feasibility of organizing a Social HMO in Orange County. SCAN, a Social HMO, is based in Long Beach and has a small number of members in Orange County.

The PACE Program

A program named On Lokbegan in San Francisco 25 years ago to care for frail elderly with low incomes, initially among the Chinese-American population of that city, and to keep them in their homes and out of nursing homes. It now operates five community service organizations in the city and has been the impetus for a national demonstration project called PACE (Program of All-inclusive Care for the Elderly). In 1983 On Lok became a dual capitation program funded by Medicare and Medicaid as an alternative to long term nursing home care which is covered, for Medicaid eligible individuals, by Medicaid. At present 15 PACE programs exist nationally as Medicare providers, 10 of which were mandated by OBRA in 1986 as a demonstration project (plus the five On Lok sites in San Francisco), and 40 sites are planned to be in operation next year. The program includes a four year demonstration project of up to 10 for-profit entities which, if successful, could qualify as permanent PACE providers. The PACE model includes a day care center which provides nursing care, case management, social activities, a nutrition program and respite for family care givers. Eligibility for the PACE program is income limited with a maximum income of $14,000 for individuals and $17,000 for married couples. It does not include pharmacy benefits but a new enhancement, called PACENET (Pharmaceutical Assistance Contract for the Elderly Needs Enhancement Tier) will cover drugs after a deductible ($500) and copayment ($8 for generic/ $15 for brands).

An independent evaluation by Abt Associates in 1995 was critical of the PACE model in several respects and suggested that the replication sites may be less successful than On Lok for a number of reasons. The catchment area of On Lok was only 2.5 square miles at initial development and is still less than 10 square miles after expansion. The replication sites have, in general, larger catchment areas and are located in communities with less population density than San Francisco. Another problem has been the slow growth of the replication sites’ client volume. Some of the reasons for slow growth include resistance to four or five day per week visits, the need to apply for Medicaid and the loss of choice of physician and other providers. Family support is necessary for this model to succeed and this varies among the replication sites. The report emphasizes that all sites exclude some eligible patients in spite of a need for expansion and this suggests that selection is occurring. This selection process is described as “niche marketing” or “skimming” in the report and the conclusion is that enthusiasm for the model among providers is greater than among clients.

The CARE Program

A similar program, at the University of Pennsylvania, called CARE (Collaborative Assessment and Rehabilitation for Elders), has been established since 1993. It is certified as a CORF (Comprehensive Outpatient Rehabilitation Facility) and functions as a day hospital similar to the British day hospital. This program found that nursing, mental health, social work and care management services are not well reimbursed in a fee for service environment. Certification as a CORF improved reimbursement. Other limitations included relations with other providers who might feel competition, transportation, limitations on the services reimbursed by the CORF structure, cost of the day care facility and time limits of the CORF model. They are dependent on referral from primary care physicians and have no managed care patients. Patients are seen by Advanced Practice Nurses at intake and needs assessment is completed. Typical problems include poly-pharmacy, depression and cognition. Patients are seen in a “day hospital” for four hours per day, two to three times per week for an average of six weeks. Then the patients are returned to primary care physician follow-up and the CARE program has no long term follow-up. They perceive their function as providing a patient “tune-up”. They use a patient information system called The Omaha System which was developed for home care. It is patient centered and includes four modules: teaching, case management, surveillance and treatment. They are pursuing the PACE model to expand the range of services for their clients. Long term nursing care is very limited in Philadelphia and long waiting lists are common for health services for seniors. They charge for present services on a fee for service system billing at $60 per 15 minutes. They have not seen any reduction in cost or hospital admissions.

The GEM Program

The Veterans Administration studied the effectiveness and efficiency of an outpatient Geriatric Evaluation and Management (GEM) program with a randomized trial comparing GEM patients and the usual primary care (UPC) population. This program was based on studies of inpatient geriatric units which had shown improved care. Other outpatient GEM studies had shown little evidence of efficacy of the concept, as measured by consumption of medical services. Tulloch and Moore, in Britain, showed increased number of admissions in GEM patients but reduced bed days and an increased use of social and health services with no difference in prevalence of health problems. The VA study showed significant improvement of health status and quality, as measured by the SF 20, and a significant improvement in survival. No significant differences in costs were reported in the original publication of the study although the GEM patients were seen twice as often.

In a long term follow-up of the study a potential “investment effect” was identified in that, during the first 8 months of the program, GEM patients incurred 35% more health care cost than the UPC group but in the later stages, after one year, the ratio reversed with GEM patients’ costs being 38% less than the UPC group. This suggests the possibility that initial evaluation identified deferred health needs which, once corrected in the early stage of the study, resulted in less health care expenditures later. If this trend continues it may indicate that early correction of health problems results in less cost in the long term and may support the cost effectiveness of the GEM concept.. Both the GEM group and the control group showed improved health perception and functional status. This effect on the control group was attributed to the “study effect” first noted in the 1920s in efficiency measurement. Control patients responded to the increased attention of the study by improved status. The VA GEM study with long term follow-up has not been replicated and their patients were all male.

UCLA Study

A similar study of geriatric assessment but this time performed at home was reported in 1995 by a group at UCLA. The patients were selected from voter rolls of the city of Santa Monica for age over 75 and 37% of those contacted by telephone agreed to participate. They were randomized to an intervention group and a Usual Care group. After a three year study period the intervention group were more able to live independently, less likely to require assistance and less likely to be in nursing homes long term. There was no difference in acute care hospital admissions and the intervention group had more hospital days per year and more physician visits per year than the control group. The principle benefit of the program seemed to be in the reduction of long term nursing home admissions.

Medical Groups

Medicare HMOs provide health care through medical groups and hospitals which contract with the parent HMO and are paid on a capitation basis. The parent HMO, with the exception of staff model HMOs like Kaiser Permanente and Group Health of Puget Sound acts as a broker selling HMO insurance to employers and, in the case of Medicare, to individuals. It then sells contracts to provide care (lives) to medical groups or Independent Practice Associations. The medical group receives a fixed sum per member per month which is then administered by the group to pay for outpatient and hospital care, as well as for other services including home care.

Some services, such as mental health and rehabilitation, are often “carve-out” services contracted to separate providers and not provided by the medical group on a per capita basis. A recent study of a mental health “carve-out” program in a non-Medicare setting raised questions of adequate care. Mental Health utilization had been modest at baseline but, after the carve-out program began, inpatient use fell and outpatient use, surprisingly, fell substantially. The usual expectation would be a shift from inpatient to outpatient services which was not seen here. Overall mental health care spending fell substantially and further investigation was suggested to see what happened to the subscribers at risk.

The HMOs have had little direct supervision of medical groups with respect to medical outcomes. Financial controls have been the principle feedback mechanisms between HMOs and medical groups. The IPAs, themselves, are usually oriented toward utilization control and have little in the way of outcome measurement except for measures of hospital days and procedures. Specifically few measure quality of life or activities of daily living functioning. Recently many large HMO companies have reported financial losses and in many cases attributed these problems to information system problems and the inability to track the cost of care for subscribers. If they cannot control costs, what are they doing about quality of care ? Nothing.

The HAPS Program

UCI has had a program called Health Assessment Program for Seniors in the past but that program is quite small at this time. The HAPS program involved health assessment with planning based upon the assessment. Continuing care by the HAPS group was not part of the program and patients were returned to their own physicians after evaluation. This approach has been validated in several studies, described above, which have shown improved health status but little or no cost reduction. The Medical Center, the Faculty Medical Group and the teaching program, have been interested in obtaining additional managed care contracts and in increasing contacts with the Medicare population. The existence of this unmet need in the frail elderly seems to offer an opportunity to provide care to an underserved group while adding a significant number of managed care lives to the Medical Center patient population.

There is also an opportunity to study the effect of a program developed specifically for this frail elderly population. The key to cost-effectiveness, based on the VA GEM experience, seems to be the continued care of the population by the evaluation group. When this occurs, the early high costs may be balanced by later cost reduction as chronic illness is kept under control without high expenditures. The VA study suggests that two years of care should balance the costs of the early intervention program and that, after two years, costs are lower in the GEM study group. This experience should be replicated in another population which includes both sexes. The problem of poor outcomes in managed care could be addressed by a demonstration project in which long term care with continuing assessment similar to the GEM program is maintained in an HMO capitated environment. The early initial costs could be negotiated with the parent HMO as part of a project and grant support of some of these costs may also be available as part of a demonstration project.

The PACE program includes a day care facility. While a PACE program might work in Orange County, the proposed program would be designed for residential care and home care without the added cost of a new facility. Effort would, of course, be made to integrate existing day care facilities into the program who wish to participate. It may be that a number of referrals for such a program would come from existing day care facilities whose clients have unmet needs in either fee for service or HMO health care plans. In some of the larger residential care facilities it might be cost-effective to establish a permanent site for the case manager/ nurse practitioner. The existing clinics would be included in the program for those patients who are not as limited as the target population but who might be expected to benefit from the multidiscipline approach.

House Calls in Modern Medicine

Recent studies have commented on the decline in the provision of medical care to patients in their homes. Home health care services have grown substantially except for physician visits and house calls by physicians were made to only 0.88% of Medicare beneficiaries over age 65 in 1993. This is estimated at 27 house calls per 1000 patients compared to 299 per 1000 patients in England. The physicians in England make 3009 house calls per 1000 patients over age 85 per year. Some of this is undoubtedly related to reimbursement which has been, until recently, shamefully low compared to non-physician providers (MDs $47 vs RN $97 and OT $102 per visit). The physician reimbursement has recently increased but is still only $87 per visit. High technology support is becoming available in the home and may make the physician house call more cost-effective in the near future.

Grant Support Availability

Grant money is available to support efforts to improve care to underserved and chronically ill populations. Such organizations as The California Endowment and the California Fund have funds available to improve the health of underserved populations and the frail elderly qualifies as one such group. The is in fact a program to study the effects of managed care on the frail elderly and the California Fund has an RFP (which is attached) for such a project. The Robert Wood Johnson Foundation has announced a similar program and information has been requested.

The Orange County Managed Care Task Force has plans to apply for a grant in conjunction with the Orange County Health Care Council. A new 501©(3) corporation is planned to administer this program and will form a partnership with UCIMC depending on the level of interest at UCI. The grant source is concerned with community based programs and with funding services to underserved populations so funding for research alone or for the cost of medical care itself is unlikely. It should be possible to use some of the grant funds for development of a program to correct health care problems due to the inadequate existing system. Other needs of this frail elderly population including transportation to obtain health care, nutrition and caregiver respite will be included in the grant proposal.


I propose that a program be developed to provide care to the frail elderly Medicare population by UCIMC and the UCI faculty. This would include home care and RCF visits by primary care physicians and nurse practitioners. The program would emphasize strong case management by nurse practitioners with geriatrician backup. Support services for these visits would include Occupational and Physical Therapy, Podiatry, Social Workers, Psychiatric Nurses or Psychiatric Social Workers, laboratory and xray services and home nursing care. There have been suggestions that a pharmacist should be included in such a program. Needs assessment would occur at intake and on an annual (or more frequent ) basis. The initial health evaluation would be along the lines of the HAPS program but would be performed in a cost effective manner consistent with financial constraints of managed care. At least part of the initial evaluation process would be supported by a grant which would assess the results of such a program by measuring patient outcomes. Many of these patients are victims of stroke and Alzheimer’s disease and psychological evaluation would be part of the initial and subsequent assessments for those with such problems. Part of the initial evaluation will identify unrecognized psychological problems such as depression and further psychological workup will be included as needed. Since many of these patients are already in residential care the assessment would be performed on site unless a visit to the clinic facility could be made with little difficulty. Transportation for diagnostic studies and specialty clinics would be part of the program and funding would be included with the grant application.

Evaluation of Need for a Day Hospital

Consultation with existing clinics and day care centers should be done to assess the need for a “day hospital” along the lines of the CARE study and the PACE program. This could be set up on the UCIMC campus and/or at the satellite clinics operated by UCIMC. The initial emphasis would be on the RCF population while investigation of the need for on site services is conducted. Since there is already a study of a PACE program in existence which is a cooperative venture of St. Joseph’s and CalOPTIMA this option will not be pursued at this time. Communication will be opened with the organizations conducting this study. At least some of this consultation can be carried out through the Health Care Council of Orange County. The CARE program in Philadelphia has been contacted for further information on their experience. It might not be necessary to build a day care center to deliver on site care to many of these patients. If a facility is large enough such a site could be established within the RCF if a large proportion of the residents join the program.

The Health Care Council

The grant application will be submitted in partnership with or under the administration of the Health Care Council. The role of UCI in administration of this grant proposal will be evaluated as the level of interest on the part of UCI is determined. Grant funds will be used to establish the program, support the initial assessments and monitor the demonstration project. Results will include evaluation of health care and social outcomes, quality of life evaluation for client/patients and family members, cost of care and savings if observed and the feasibility of operating such a program within the managed care system. The advantage of partnership with the Health Care Council includes concerns of the grant sources that community issues be addressed and communication with stakeholders who have had little contact with the University and who may not appreciate the concerns that UCI has for the residents of the county. The University has, in the past, avoided contact with community agencies and this should be corrected. The project described here is an excellent opportunity for the University to become involved with the community.

Hospice and Advance Directives

Most of these patients are not candidates for expensive high technology services and family conferences and advance directives would be part of the initial assessment. A hospice-like program would be an integral part of the plan but would only be implemented with patient and family consultation. Hospice Benefits under Medicare would be obtained for those patients who qualify under Medicare criteria. Hospital care would be part of the plan and would be provided at UCIMC.

UCIMC Contracts with HMOs

UCIMC has a contract with Secure Horizons and several other HMOs already. As a contracting medical group it should be possible to enroll as many of these Medicare beneficiaries in the new plan as are interested. There are several RCFs who are willing to promote the new program to residents when it is established. Marketing in retirement communities may require consultation with Pacificare with respect to their relations with their own medical groups in areas near large retirement communities such as Leisure World. There are geographic restrictions included in contracts between HMOs and provider groups which restrict the ability of other medical groups to compete for HMO patients within the same region. There are, however, thousands of retired persons living in retirement communities and residences within a 10 mile radius of UCIMC. Many of these patients, when active and ambulatory, may prefer to remain with their present providers. Eventually, a reputation of excellence may attract the more healthy population as well. Some of the provisions of the proposal will have to be approved by the HMO as they may not correspond to language in the usual medical group contract. An effort should be made to include mental health services in the contract as these will be an integral part of a program serving this population. Many of these patients are Alzheimer’s victims and mental health providers should have a major role in the care team.

Program Design

The design of such a program would be along the following lines.

1. Each patient would have a case manager who would be a geriatric nurse practitioner.

2. The initial intake interview would include history and physical examination, assessment of cognitive function and psychological status, function with Activities of Daily Living and nutritional status.

3. When appropriate, based upon history, the patient would be seen by physical and/or occupational therapists.

4. Social service agency evaluation would also be part of intake evaluation.

5. A primary physician would be assigned and will be the backup for the nurse practitioner. The initial evaluation will involve the physician who will develop the care plan in consultation with the other members of the team. New geriatric physicians and nurse practitioners will be recruited beyond those currently at UCI when case loads increase. Once established this should be a good source of teaching experience for primary care training and nurse practitioner training.

6. Psychiatric social workers would be part of the initial evaluation and the HMO would be requested to include psychiatric services in the plan rather than using a “carve out” program.

7. Consultations would be obtained within the UCI medical group and all services would be provided by UCI providers. Rehabilitation services are also “carve-out” programs in many areas and these will also be requested as part of the contract.

Case management will be the primary method of coordination of care and the case manager will present each patient to the team after the intake interviews. Diagnostic studies and necessary therapeutic interventions will be completed in the first month or two after intake. The CARE program uses a 6 to 12 week period to complete their “tune-up” of patients and this should be the goal of this program. Weekly staff meetings will be used to coordinate treatment plans until stability is accomplished. The plan is based on long term participation by patients in the program. The increased early expenses incurred by correcting neglected health problems will be offset by later cost savings as acute care admissions and ER visits are reduced. A short term GEM program for fee for service patients will be kept available.

Case Management in HMOs

In the current managed care system two levels of case manager are encountered. At the medical group level there is a high turnover among case managers and contacts are difficult to maintain between the care givers and case managers whose identity (and skill level) constantly changes. Pacificare has a competent case manager staff but they act only in an advisory role to the medical group. Rehabilitation services, for example, are contracted out to the lowest cost bidder and the benefits often do not fulfill legal minimums. Physicians also have a high turnover rate in the contracted services which UCI has provided to nursing homes. They are often unfamiliar with Title 22 requirements and OBRA requirements of nursing homes. Use of case managers with experience in this setting will solve many of these problems. Discussions will be held with Pacificare medical directors to determine the level of cooperation which can be expected from the HMO.

Electronic Medical Record

Grant support will be sought to develop a medical information system which will allow an electronic medical record to be accessible from the treatment site by laptop computer. This will be used for data entry as well as coordination of care. All progress notes will be typed into the system and all lab and other diagnostic studies, as well as the notes of other members of the treatment team, will be accessible on site. HL7 interfaces with legacy systems will be used to download all lab and other UCIMC data to the system. An attempt will be made to include this system in a new outpatient records system which will establish an electronic medical record at UCIMC for outpatient services. Included in this electronic record will be decision support systems using rules based decision trees developed by the medical staff to monitor and supervise the medical care delivered on site by caregivers. This system is already in use at the University of Washington outpatient clinics and uses a browser intranet interface for ease of use. The CARE program uses an existing system from Epsilon which was developed for home care programs and which is accessible from remote locations by laptop computer.


There is an opportunity for UCIMC to develop a partnership with community groups which are asking for help with managed care. The rapid increase in the Medicare HMO market share has not been addressed by the other models of frail elderly care. This is an opportunity to develop a new model, possibly in partnership with HMOs as well as community groups. Strong case management skills and the use of physician extenders should be of help in controlling costs and keeping the program within the budget set by the capitated contract. A demonstration project to improve the care of frail elderly in managed care may attract funding from HMOs who are interested in improving care or deflecting criticism of their performance with this frail population. One aspect of the problem is the timely availability of medical records for on-site care. UCIMC clinics have been plagued by medical record problems, some of which have resulted in legal liability. This is also an opportunity to develop an electronic patient record for the entire outpatient department of UCIMC and grant funds are available for such a project. Disease management software for the geriatric practice might be developed through such a project.

Michael T Kennedy MD FACS
Department of Surgery
UCI College of Medicine

The above proposal was approved by the faculty medical group and by the Department of Family Medicine but was vetoed by the UCIMC administrator. The proposal is posted here as an indication of what might have been.

One Response to “A proposal for care of the frail elderly in assisted living homes in 1995.”

  1. […] A proposal for care of the frail elderly in assisted … – This proposal was submitted for approval to the administration of UCI Medical Center in 1995. The faculty medical group was enthusiastic and the chair of the Family … […]

Leave a Reply