MODELS OF GERIATRICS PRACTICE

Series Editor: David B. Reuben, M D

Post-Hospital Sub-Acute Care: An Example of a Managed

Care Model
T. von Sternberg, MD,* K. Hepburn, PhD,t P. Cibuzar, RNC, GNP, L. Convery, RNC, GNP,
B. Dokken, RNC, GNP, J. Haefemeyer, MD, S . Rettke, RNC, GNP, J. Ripley, MBA,
V . Vosenau, MD, P. Rothe, MD, D. Schurle, MD, and R. Won-Savage, MD

physician satisfaction with the units was high. The program

OBJECTIVE: This article describes Transitional Care Cen- provided economic benefit to both partners. The health
ters (TCC), an innovative sub-acute care program developed plan’s negotiated rate for the TCC units was 38%) less than
by a large managed care organization (HealthPartners in that paid in noncontractual facilities.
Minnesota) in partnership with five area nursing homes. The
purpose of the TCC is to promote continuity of care for frail CONCLUSION: The TCC partnership provides rehabilita-
older members covered under a TEFRA risk contract. tive and geriatric evaluation services in settings more condu-
cive to and less costly than such care usually, and yields
DESIGN: This is a retrospective study of the experiences and improvements in care and utilization outcomes. J Am Geriatr
outcomes of enrollees who received TCC compared with a SOC45~87-91,1997.
like group of enrollees who received customary continuity
care through contract services.
SETTING: The TCCs are established contractually in five
area nursing homes; these facilities keep at least 15 beds
I ncreasing economic pressures are driving fee-for-service
and managed care organizations to develop new systems to
provide their geriatrics clients with better care, in particular
available to the health plan for round-the-clock, 7 days per care that promotes timely discharge from the hospital.’ Al-
week admissions for sub-acute care. Designated staff from ternative delivery methods have been instituted to provide a
these facilities and designated geriatric nurse practitioners place for older people to receive post-acute therapy and
and geriatricians from HealthPartners follow established tar- nursing care in lieu of continued hospitalization. Hospital-
geting, admissions, assessment, care planning, and discharge based rehabilitation models include units on separate hospi-
planning procedures to provide team care for these patients at tal floors, acute rehabilitation units, and geriatrics rehab
the facilities. h ~ s p i t a l s . l -Long-term
~ care- based, sub-acute or transitional
PARTICIPANTS: The TCC program is targeted to patients care units have also emerged to address these needs. As such
requiring rehabilitation therapy (post-cardiovascular acci- models evolve, a number of dimensions need to be examined
dent, post-fractureheplacement) who are deconditioned, or to determine their impact. Among the most important of
those with uncomplicated infections (urinary tract infection, these are the use of principles of comprehensive geriatrics
pneumonia). A total of 1 144 patients participated in the TCC evaluation and team management, attainment of therapeutic
program in the 1-year program under report, and 253 were goals, reduction in cost and length of stay, and the effective-
surveyed in regard to their experience. One hundred Health- ness of discharge arrangements to promote recovery and
Partners physicians were surveyed about the program. prevent compIication~.~-’
RESULTS: Post-acute length of stay in the TCC was substan- HealthPartners, a staff model nonprofit H M O based in
tially lower than that in customary care settings in contract Bloomington, Minnesota, currently enrolls 21,000 geriatrics
nursing homes (14.3 versus 20.5 days). Rehospitalization patients in a TEFRA risk contract to provide a wide array of
rates from these units were comparable to or better than preventive, acute, rehabilitative, and ancillary services to
those from other sub-acute units. Patient and primary care older adults.8 T o manage the needs of this population, Geri-
atrics Programs of HealthPartners has developed and imple-
mented a number of care programs to improve the quality
This article is onc in a series on Models of Geriatrics Practice. and cost-effectiveness of care. This paper reports on one of
From the ‘HealthPartners Geriatrics Program and the tuniversity of Minnesota the most successful of these programs, the Transitional Care
Medical School Department of Family Practice and Community Health, Minne- Centers (TCCs). The TCCs are a partnership between
apolis, Minnesota.
Address correspondence to T. voii Sternbcrg, MD, Chair, Geriatric Division, HealthPartners clinicians and a network of sub-acute geriat-
Healthl’artners, 8100 34th Ave. South, PO Box 1309, Minneapolis, MN 55440- rics rehabilitation and assessment units in five separate nurs-
1309. ing home locations across the service area.

JAGS 4.5:87-91, 1997

0 1997 by thc American Geriatrics Society 0002-86 14/97/$3..50
88 __ ET AI..
V O N STERNBERG ~ _____ IANUARY 1997-VOL. 45. NO. 1 IACS

PROGRAM STRUCTURE TCC medical management and rehabilitation. The

The long-term care institutions in the TCC program TCC unit is also a logical place to complete a compre-
were chosen because of interest and reputation and because hensive geriatrics assessment (CGA).9 The hospital
of a staffingmix that allows for service to patients with higher setting is a difficult venue for CGA because of pressures
acuity level care needs. The TCC units are staffed at a higher to limit length of stay;“’.” however, staffing, staff
nurse-patient intensity than are other units in these facilities skills, unit rehabilitation orientation, and lower per
(approximately 1 to 9 versus 1 to 15). Each facility guaran- diem costs make the TCC units more conducive set-
tees that at least 15 beds in a unit will be available to admit tings for evaluation. Patients seen in a primary care
Healthpartners patients on a 7 days a week, 24 hours a day clinic, in urgent care, or in the emergency room and
basis. Patients may be transferred from acute care or may who are at risk of decline and eventual hospitalization,
bypass hospitalization completely and be placed directly on deconditioned, have a caregiver in crisis, have poor
the unit after an assessment in a primary care clinic, an urgent nutrition, or a probable diagnosis of dementia or de-
care clinic, or an emergency room. The units offer physical pression are appropriate for CGA.
0 At any given time, approximately 80% of TCC pa-
and occupational therapy twice a day, 6 days per week, and
they can provide IV and TPN therapy and administer contin- tients are hospital transfers and 20% are direct admits.
uous heparin and IV hydration when indicated. The facilities Of the hospital transfers, about one-third are orthope-
are paid on a contracted per diem rate, which is inclusive of dic patients, one-third deconditioned and Dostsurgical
room and board costs, nursing services, therapy services, and patients, and one-third post-CVA patients.*All patrents
discharge planning. upon admission require assistance with at least one
A Healthpartners board certified geriatrician and a GNP activity of daily living (ADL).
0 Care in the TCCs is data driven. On the day a patient
work in conjunction with the facility nursing, therapy, and
social worker staff to form a stable clinical team at each TCC. transfers from acute care to a TCC unit, the hospital
Eight physician members of Geriatrics Programs take part in physician dictates a discharge summary. This sum-
the teams, each taking a 4-month rotation at one of the five mary, copies of the progress notes, laboratory data
TCC sites. Five full-time Healthpartners nurse practitioners from the patient’s last 3 hospital days, copies of the
remain assigned continuously to specific TCC facilities. This history and physical, and consultants’ dictations are
arrangement assures communication and collaboration sent with the patient. On admission to the TCC, initial
among the clinical partners and promotes continuity of pa- assessments, including functional and self-care status,
tient care. mental status, and depression are performed by TCC
team members from the nursing homes’ rehabilitation
PROCESS OF CARE staffs and the geriatrician or nurse practitioner. TCC
team members from Healthpartners determine ad-
Admission to a TCC unit occurs either directly or vance directive status at this time. This information
through a transfer from an acute care setting. A HealthPart- and the transfer data contribute to development of an
ners hospital physician working with a discharge planner and interim plan of care to be coordinated and carried out
in communication with the patient’s primary care physician by the multidisciplinary care team made up of the
at a community clinic may decide to transfer a patient from HealthPartners GNP and geriatrician and staff from
the hospital to a TCC unit. The TCCs offer greater flexibility the TCC. The plan charts the care team’s activities for
in discharge planning, and they provide a viable transfer rehabilitation or evaluation, as appropriate.
alternative for older patients with a number of conditions 0 Each week components of the assessment are repeated,
including: including review of:
0 orthopedic patients (hip/knee replacements, hip frac- 0 patient’s medication use

tures) requiring more than 5 to 7 days of postsurgical a gait, fall risk, and progress in reconditioning therapy
therapy (typically, these patients transferred on 0 mood and affect (geriatrics depression scale)I2

post-op Day 3 or 4) 0 mental status’” (performance based test of judgment,

0 musculoskeletal injuries and stable fractures (pelvis safety, and problem solving)
and lumbar spine) a nutritional status and skin care
0 post-cardiovascular accident (CVA) patients 0 advanced directives and Do Not Resuscitate status (if

0 deconditioned congestive heart failure (CHF) and patient status changes)

chronic obstructive pulmonary disease (COPD) pa- 0 comprehensive discharge plan

tients 0 This information is reviewed at the team rounds, and

a postsurgical patients who have wound tissues or are decisions are made about next steps.
deconditioned 0 The care team, led by the geriatrician, meets weekly for

0 patients with uncomplicated infections (e.g., urinary chart rounds and to discuss each patient, reviewing
tract infection (UTI))or who are deconditioned by an care plan, progress, and the status of discharge plans.
acute illness (e.g., pneumonia) The meetings are attended by the geriatrician, GNP,
0 The Medicare risk arrangement makes it possible to unit head nurse, occupational, physical, and speech
admit directly to the TCC units patients with new therapists, and social workers. Each member is ex-
medical conditions that threaten functional ability but pected to contribute suggestions to overcome limita-
who are stable enough not to require an acute stay. tions of patients’ progress or barriers to discharge
Patients injured in falls (without fractures), with uri- home and to assist in targeting the date of discharge.
nary tract infections with mild dehydration, or with The geriatrician and GNP emphasize review of phar-
pelvic fractures typify those who might benefit from macy and medical issues. Family difficulties or home/
JAGS
_______ JANUARY 1997-VOI.. 45,NO. 1 __ POST-HOSPITAL SUB-ACUTE CAKE 89

environment obstacles are also discussed. Should the linked through clinician participation or negotiated contract.
team decide returning home will not be appropriate, During this year, patients discharged to nursing homes for
the patient and family are assisted with identifying a rehabilitation could choose where they would get their post-
more appropriate living arrangement. hospital care. Although patients were not randomly assigned
0 After the weekly meetings, patients are seen by the to the two conditions of post-hospital care, our review of
geriatrician and the GNP. Family members are in- their diagnoses indicates the groups are comparable.
formed of the time of the visits and are encouraged to Length of stay was monitored for 12 months, comparing
be present with any questions they may have for the the number of patients sent to TCCs with the number of
geriatrics team. The geriatrician visits the facility a patients sent to non-contracted, non-managed venues (in-
second time later in the week to review any new cluding community long-term care facilities and rehabilita-
admissions and see current patients who may be in tion units attached to hospitals). The average length of stay
need of reevaluation. The GNP ensures an ongoing (ALOS) for the TCC facilities was 14.3 days. ALOS for
clinical monitoring of the patient’s status by visiting non-contract facilities was 20.5 days. This also compares
two or three additional times in the course of the week. favorably with results of a survey of the other sub-acute care
The GNP or the geriatrician update the patient’s pri- units in the metro area, which showed that typical length of
mary care physician on a weekly basis by telephone. stay was in the 20 to 22 day range (personal communication).
The geriatrician and/or GNP also ask for the primary These data indicate there may be benefits to tighter clinical
physician’s input on any difficult clinical or social monitoring and case management.
issues. Post-discharge status was monitored through a series of
0 Discharge from the TCC occurs when the patient no phone calls by a HealthPartners social worker at points 3 and
longer requires or benefits from the sub-acute level of 6 months after discharge. Depending on the TCC unit, 76 to
care. This may happen because the therapeutic goals 80% of the patients discharged to home from sub-acute care
have been met or the geriatrics evaluation is com- on the TCCs remained at home. When compared with ADL
pleted. It may also occur because the team determines status on the discharge summary, all patients were at the
the patient will no longer benefit from intensive ther- same or higher ADL functional levels at 3 and 6 months from
apy. The unit social worker, who has managed the discharge.
discharge planning process from a point before admis- Retention rates in the TCC are also favorable. The
sion, works with the patient and family throughout the rehospitalization rate during stays in the sub-acute units is the
stay to keep them involved in and informed about the same or lower than at non-contracted facilities. The rate of
process and to make the arrangements necessary to admissions to the hospital of patients from long-term care is
maintain continuity of care following discharge, re- monitored on a monthly basis. Overall, the rehospitalization
gardless of discharge location (i.e., to home or another rate for TCC units and non-contract homes was approxi-
care facility). mately six patients per 100. However, TCC hospital readmis-
sion rate for patients with diagnoses of pneumonia, UTI, and
OUTCOMES sepsis was lower ( 3 patients per 100 admissions) than the
The greater clinical presence of the geriatrics team and non-contract rate despite the high level of acuity, which
the emphasis on appropriate advanced directives allow care would lead to an expectation of higher rehospitalization
to be delivered to very frail, functionally disabled patients rates. (See Table 1)
who would normally require an acute hospital setting. Primary care physician satisfaction with the TCC pro-
Within the sub-acute units, these patients receive appropriate gram is high regarding the care given to their patients and the
care in an appropriate venue. To HealthPartners this also communication they have with the geriatrics team. We sur-
means that the overall health care cost has decreased for the veyed 100 primary care internists and family physicians
individual patient while medical and rehabilitative care has about their experience with patients sent to the transitional
been maintained. care units. The survey sought information on TCC admission
The TCC program has been functioning for 2 years and practices, satisfaction with care, care continuity, and dis-
has shown encouraging results. During its first year, the charge planning, and with communication with the team.
structure of the program allowed a natural comparison be- Sixty percent of those surveyed responded, and all indicated
tween outcomes on the TCC units and those on other nursing that the availability of the TCCs was helpful in facilitating the
home units available to Healthpartners members but not patient’s discharge from the hospital. Eighty-nine percent

Table 1. Summary Table Comparing TCC Units with Non-contract Facilities Accepting Patients for Post-hospital Care and
Rehabilitation

Transitional Care Non-contract

FaciIities Facilities

Average length of stay 14.3 20.5

Per diem costs (including therapies) $1 85 $280-$300
Re-hospitalization rate per 100 admits 0.06 0.05
Re-hospitalization rate for infections 0.03 0.06
(urosepsis, pneumonia) per 100 admits
YO VON STERNBERC ET AL. ~~
JANUARY 1997-VOL.. 45, NO. 1
~~
JACh

responded their patients received appropriate care; 62% in- post-hospital sub-acute care for its frail geriatrics population.
dicated that they had admitted patients to the TCCs, in lieu of The program is built on partnerships with high quality com-
the hospital, directly from the emergency room or clinic. Of munity long-term care facilities with an interest in geriatrics
those responding, 82% indicated strong or very strong satis- rehabilitation and a commitment to development of a transi-
faction with care delivery in the TCCs, and 71% indicated tional care system. The program is centered around an ag-
these levels of satisfaction with communication with the TCC gressive clinical management team. The TCC model switches
geriatrician and GNP. the emphasis for comprehensive care of older people away
Patient satisfaction was also examined. A survey was from the acute care hospital facility, where it is often over-
mailed to all 253 TCC patients surviving at discharge and shadowed by invasive high-tech care, to a more appropriate
sent home over a period of 6 months; 92% responded. Using setting that focuses on geriatrics principles and rehabilita-
a 5-point scale (1 = strongly disagree to 5 = strongly agree), tion. l 4 At the sub-acute unit, geriatrics rehabilitation can be
patients provided a positive evaluation of the care they re- emphasized fully and the recommendations of the geriatrics
ceived at the TCCs. The item regarding satisfaction with the team can be implemented successfully.
frequency with which they were seen by the geriatrics staff The presence of Healthpartners clinical staff on a daily
received a mean score of 3.68 (SD 0.97).The item regarding basis has been one of the main reasons that patients with high
satisfaction with how appropriately questions about their levels of acuity can be managed successfully. Although the
condition and progress were answered received a mean score multidisciplinary team approach was received enthusiasti-
of 3.79 (SD 0.86). The mean score for patients’ sense that cally by the staff at the TCCs, nursing staff on the units werc
care met their needs was 3.92 (SD 0.90). not always prepared to manage these higher levels. The social
The program has a substantial impact on the costs of workers also needed to be prepared for shorter lengths of stay
care. Because competition for patients to fill sub-acute care and more intense discharge planning. We addressed these
beds is intense, facilities are open to establishing lower per challenges, in part, by taking a more active role in medical
diem rates, and we were able to negotiate TCC contract rates education of the staff (in services and a structured lecture
that reduced per diem costs by 30 to 40%. The presence of series).
Healthpartners’ geriatricians and GNPs assured facilities Patient and family satisfaction with the program is high;
they could manage this frail population and provided added however, convincing them while in the hospital of the need
incentive in terms of potential for staff development. The for transfer to the sub-acute unit has been difficult at times. It
TCC units’ per diem includes therapies, while the non- has taken consistent communication from all providers in the
contract facilities’ therapy departments bill separately from hospital to ensure patient acceptance. Once patients arc
the room and board rate. When we compared per diem rates admitted to the units and begin their rehabilitation, they are
at the TCC units with those rates charged to Healthpartners pleased with their care. Families, alerted to what days the
by non-contract facilities, we determined that the TCC per geriatrician will be at the unit, are pleased to find they can
diem cost is 38% less than care provided in non-contracted have questions answered directly. Enhanced communication
facilities. Combined with the decreased length of stay in the between the patients and families and the geriatricians and
TCCs, we estimate reduced service costs at approximately $1 GNPs has led to franker and more detailed and informative
to 1.5 million dollars during the first year of implementation. conversations about prognosis. Details of prospects for resus-
This figure does not include the savings from decreased citation are given more clearly. As a result, there is a very low
hospital length of stay, a potential beneficiary of the program, incidence of hospital readmissions for patients with end-stage
especially for orthopedic and stroke-related diagnoses. disease (i.e., COPD, CHF), and approximately 90% of the
Healthpartners inpatient length of stay trends are favorable patients have do not resuscitate orders.
(comparing before and after introduction of the TCC pro-
The program faces a number of ongoing challcnges.
gram); however, detailed analysis has not been completed.
Foremost among these is the need to assure that participating
Active participation and leadership by HealthPartners’
geriatricians are relieved of hospital and clinic responsibilities
geriatricians and GNPs come at a cost. Each site occupies
so they can fully support the TCC units. This is an issue of
approximately 0.75 full time equivalent (FTE) of a GNP and
cost and value. HealthPartners has rccognized that the TCC
0.2 FTE of a geriatrician. So far, these costs have been offset
program addresses its commitment to its geriatrics popula-
by the savings that have resulted. Once the TCC is adopted as
tion and its interest in developing innovative models of care
the system-wide standard of practice, TCC length of stay
delivery, There is also an understanding of the cost savings
figures will become the norm and will no longer represent
savings to the system. Incremental gains will be made as the potential. There is continued need to emphasize to the orga-
system becomes even more aggressive in discharge planning nization, but also to clinician colleagues, the important con-
from the hospital where appropriate because of comfort with tribution physicians make in the sub-acute units so that these
the level of clinical care that can be provided in the TCCs. are valued equally to the efforts required in the clinic and
The negotiated rates with the TCCs proved advanta- hospital settings. In addition, information transferred from
geous for the nursing homes as well because, although the per the hospital to the units needs ongoing monitoring; hospital
diem, including therapies, was less, they benefited from a chart information can still be difficult to obtain.
guaranteed high volume o f patients. In addition, their cash The TCC program is not static, and new elaborations are
flow improved because payment from Healthpartners occurs already under consideration. The next steps in the develop-
within 60 days com[pared with many months from Medicare ment of this model for Healthpartners include: earlier trans-
and most other third-party payers. fer from the hospital for appropriate patients, the identifica-
tion of clinical pathways for diagnosis such as hip fracture
DISCUSSION and stroke rehabilitation, and development of a similar geri-
The Geriatrics Programs of Healthpartners has devel- atrics unit for frail patients who still require care in the
oped a succcssfuI, cost-effective, comprehensive model of hospital.
JAGS JANUARY 1997-VOL. 45. NO. I POST-HOSPITAL SUB-ACUTE CARE 91

REFERENCES 8. Zarahozo C:, LeMasurier J. Medicare and managed care. The Managed
Health Care Handbook 1992;321-344.
1. Schwartz WB, Mendelson DN. Hospital cost containment in the 1980s; les-
sons learned and prospects f o r the 1990s. N Engl J Med 1991;324:1037- 9. Applegate WB, Miller ST,Graney MJ et 81. A randomized controlled trial of
1042. a geriatric assessment unit in a community rehabilitation hospital. N Engl
2. Fitzgerald JF, Moore PS, Dittus, RS. The care of elderly patients with hip J Med 1990;322:1572-1S7X.
fracture. Changes since implementation o t prospective payment system. 10. Winograd CH, Stearns C. Inpatient geriatric consultation, Challenges and
N EnglJ Med 1988;319:1392-1397. benefits. J Am Geriatr Soc 1990;.38:926-932.
3 . I.andefeld CS, Palmer RM, Kresevic DM et al. A randomized trial of care in 11. Fretwell MD, Raymond PM, McGarvey ST. The senior care study: A con-
a hospital medical unit especially designed to improve the functional out- trolled trial o f a consultativc/unit based geriatric assessment program i n
comes o f acutely older patients. N Engl J Med 1995;332:1338-1344. acute care. J Am Geriatr Soc 1990;38:1073-1081.
4. Palmer RM, Landefeld CS, Krescvic D,Kowal J. A medical unit for the acute 12. Ycsavage JA, Brink TL.,Rose T et al. L~evelopmentand validation of a geri-
care of the elderly. J Am Geriatr Soc 1994;42:545-552. atric depression screening scale: A preliminary report. J I’sychiatr Res
5 . Lafevre F, Feinglass J, Ports S et al. Iatrogenic complications in high risk el-
1983;17:.34-49.
derly patients. Arch Intern Med 1992; 152:2074-2080.
6. Creditor MC;. Hazards of hospitalizatitrns of the elderly. Ann Intern Med
13. Folstcin MF, Folstein SE, McHugh I’R. Mini-mental state. A practical
1993;l 18:219-223. method for grading the cognitive state of patients f o r the clinician. J Psychi-
7. Naylor M, Brooten D, Jones R et al. Comprehensive discharge planning for atr Res 1975;12:189-198.
the hospitalized elderly, A randomized clinical trial. Ann Intern Med 14. Brunimel-Smith K. Geriatric rehabilitation. Clin Gcriatr Med 1993;9:689-
1994;120:999-1006. 871.