| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to Robert Newcomer, PhD, University of California, 3333 California Street, Suite 455, San Francisco, CA 94118. E-mail: robert.newcomer{at}ucsf.edu
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussion and ConclusionsReferences |
|---|
Key Words: Nursing home • Case management • Postacute care • Geriatrics
In this article we describe PACT outcomes from the phase of the program when a randomized control design was in place. We compare nursing home discharge rates and length of stay between those individuals in the intervention group and those in the usual care or control group.
Background
Two bodies of literature influenced the design of the PACT program. These include studies of caregivers prior to and subsequent to the nursing home admission of their spouse or family member, and research about problems affecting frail elders transitioning across the geriatric care continuum. The former work found caregivers at risk for depression, having a need for emotional support, and often with a limited understanding of the assistance needs and the problems they might have in the patient's transition to home (e.g., Gaugler, Leitsch, Zarit, & Pearlin, 2000; Lieberman & Fisher, 2001; Port et al., 2001). The PACT program recognized that the return of nursing home residents to the community might be facilitated if these caregiver needs were addressed.
Research on transitional problems arising with moves from the hospital to nursing homes or hospitals to home care comprised a second group of studies influencing the PACT design. Factors contributing to adverse events (e.g., emergency room use, rehospitalization) include fragmentation in treatment and communication between levels of care, conflicting care recommendations (including medication errors), and patient or caregiver distress (Bours, Ketelaars, Frederiks, Abu-Saad, & Wouters, 1998; Parry, Coleman, Smith, Frank, & Kramer, 2003; Richards & Coast, 2003). Care coordinators and interdisciplinary team interventions have been effective in improving transitions across the continuum of geriatric care (Parry et al.). These efforts have improved the flow of clinical information, improved client or caregiver satisfaction, and reduced rates of depression in caregivers; they have contributed to better clinical outcomes (e.g., reduced rates of readmission) and reduced Medicare expenditures.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussion and ConclusionsReferences |
|---|
PACT Program Intervention
The PACT program consisted of five main functions.
Case Finding and Recruitment
Contra Costa County's nursing home ombudsman's office allocated an ombudsman for 20 hr a week to be the primary recruiter of PACT participants. This was done through contact with nursing home discharge planners, meetings with patients and families, and the distribution of program flyers to patients and families. Recruitment targeted four nursing homes (35 participants), but patients in other facilities had access to program brochures and could self-refer (27 participants). Regardless of referral source, all interested families were required to submit PACT applications containing contact and eligibility information to the county's Information & Assistance program.
Patient Assessment
Care planning for the intervention group began with a PACT nurse (50% time). She met with the nursing home patient to conduct an assessment and medical chart review. The assessment identified the patient's health, function, and mobility issues that the caregiver would have to address for a successful transition. The PACT nurse presented her findings and recommendations at a PACT team care-planning meeting. Case management was then transferred to the program social worker. The nurse was available for ongoing consultation with the social worker. The nurse also monitored the status of the patients assigned to the comparison sample. She made referrals to the social worker if a discharge was pending or if the 60-day limit had expired.
Caregiver Assessment and Care Management
The second step in the intervention involved the PACT social worker (100% time). She conducted an assessment of the caregiver in his or her home setting. The assessment identified environmental, social, emotional, and financial issues that could pose a problem during the patient's transition from the nursing home. It also examined the caregiver's ability to provide care. The PACT social worker discussed the assessment and a care plan identifying the potential services and equipment needed to assist the caregiver at a PACT care-planning meeting. The social worker was the case manager. Case management involved facilitating service referrals and education about them, and providing emotional support. Postdischarge contacts were usually by telephone.
Assistive Devices and Environmental Assessments
Independent Living Resource, a nonprofit community provider specializing in assistive devices and housing for individuals with disabilities, provided staff at PACT care-planning conferences. The Independent Living Resource staff offered advice on assistive devices and environmental solutions for PACT clients, and they conducted in-home assessments, as necessary, to determine the need for assistive technology and assist with the purchase, set up, and training of caregivers in its use.
PACT Services and Financial Assistance
PACT-funded equipment, home modifications, supplies, and emergency services were not otherwise reimbursable under Medicare, Medicaid, or other insurance. Funds also were used by recipients waiting for payment from other sources and helped to avoid unreasonable delay in meeting an urgent need. Expenses were recovered from Medicare or other sources when possible. Table 1 lists the equipment and services reimbursed by the PACT program during the study period. Individuals in the intervention group were eligible for these funds with the start of case management. Those in the comparison group received this assistance only after being discharged into the community.
|
Names of those meeting the eligibility criteria were forwarded to the PACT program administrator for randomization into either the intervention group (i.e., immediate access to PACT case management) or the comparison group, which continued their usual medical and nursing home care. Approximately equal numbers were randomly assigned to the intervention and comparison care groups. Recruitment and informed consent procedures were approved (H945-21314-03) by the Committee for Human Research, University of California, San Francisco. The obtained sample size of 30 per group has a statistical power of.8 (single-tailed 
=.05), with the assumption that 85% of those in the intervention group would be discharged from the nursing home (the result found in the project's pilot work) versus about 60% of those in usual care (the countywide average of admissions discharged within 90 days).
Data Sources and Measures
We obtained data from nursing home patient charts and client records compiled by the PACT program. These contained information on date of nursing home admission and discharge, primary diagnosis, age and gender of the patient and caregiver, caregiver relationship, referral source, dates of nurse and social worker assessment, and postdischarge outcomes (e.g., readmission, emergency room use, hospital use, or death) and end of study status.
Analysis
We tested two hypotheses:
We conducted the analyses by assuming an intention to treat among those assigned to the PACT intervention. In other words, the outcome comparisons do not adjust for the promptness by which case management was implemented or the intensity of this assistance. We tested unadjusted differences between the study groups in the likelihood of nursing home discharge by using Fisher's exact test statistics. We used multivariate logistic regression with group assignment as the test variable to assess group assignment effects after we adjusted for patient primary diagnosis, age, gender, and ethnicity or race, and caregiver age and gender. We tested unadjusted differences in the length of the nursing home stay by using t tests. We used multiple regression to adjust for the attributes noted herein in testing the length of stay outcome.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussion and ConclusionsReferences |
|---|
|
For those in the intervention group (n = 33), the next program task was the conduction of the social worker's assessment of the caregiver. This was to be followed by care planning and case management. For 21 of those in the intervention group, the caregiver assessments were conducted in fewer than 7 days (median was 4 days) following the patient assessment. However, 6 individuals did not receive caregiver assessment for 3 weeks or more, and 6 caregivers declined these assessments. The delays were due to either changes in the patient's health status or staff vacations. Of those declining, only one was a family in which a determination had been made that the nursing home placement was permanent.
The time lags between the nurse and social worker assessments were potentially problematic for the predischarge case-management intervention, as this allowed little time for care-plan development and implementation prior to the discharge. For the caregivers receiving assessments prior to the care recipient's discharge (n = 19), the median time before discharge was 4 days, although 5 of these individuals received assessments with 2 days or fewer prior to discharge. Nine other PACT intervention patients were discharged without care plans. Four of these received the social worker assessment after they were home (median of 13 days after discharge); the balance had refused the assessments and declined further PACT involvement.
For those in the comparison group (n = 29), assessments were generally conducted only after the patient had been discharged home. The postdischarge period is outside the scope of this study, but these assessments were done as a courtesy to the comparison group in return for their study cooperation. The assessment led to a care plan and possible assistance with program services and case management. About one third (n = 8) of the comparison group caregivers declined the assessments after the nursing home discharge. Several others (n = 5) dropped out of the study earlier, when the nursing home placement was deemed as permanent.
Once discharged, all study participants had access to PACT service and equipment funds. Half of those in the intervention group (17 of 33) received such funding. Expenditures averaged $480 per group person, or $932 per recipient. The individuals in the comparison group became eligible for these services after they returned home following their nursing home discharge, with one third (10 of 29) receiving them. The group mean was $151; the mean expenditure per recipient was $439.
Did PACT Help Individuals Return Home?
We investigated the question of PACT's efficacy by testing whether the likelihood of a nursing home discharge was higher among those in the intervention group than among those in the control group, and whether the total number of days in the nursing home after program application was lower. Table 3 shows the unadjusted results. There is a trend suggesting a modest effect, reflected in both higher rates of discharge (84% vs 76%) and shorter median stays (42 vs 55 days), but these differences are not statistically significant. We also tested multivariate models adjusting for individual patient and caregiver characteristics. These models (not shown) also failed to show a statistically significant group-assignment effect after we adjusted for patient and caregiver characteristics. With the small outcome difference between the intervention and comparison groups, a sample almost 10 times larger than the one we used (i.e., 30 per group) would have been needed to show statistically significant differences.
|
|
Discussion and Conclusions |
|---|
|
TOPAbstractMethodsResultsDiscussion and ConclusionsReferences |
|---|
One problem for PACT was that patient recruitment was targeted on nursing homes expressing the most interest in the program. This reduced the margin for an efficacious effect. Within the study sample, just over 76% of those in the comparison sample were discharged from the nursing home. This rate is consistent with the historical average lengths of stay for recent admissions in most of the nursing homes referring patients to the PACT program (47 of 62 study participants came from such facilities). The comparison sample's discharge rate is almost 40% higher than the rate (62%) averaged among all nursing homes in the county (California Office of Statewide Health Planning and Development, 2003). Limiting patient recruitment to facilities with a historically lower proportion of patients discharged within 3 months may have provided more opportunity to make an improvement over usual care. There is, however, no assurance that the intervention effect of 85% being discharged would be achieved in such settings.
Another factor complicating matters was Medicare reimbursement. Nursing homes tended to accelerate their efforts to discharge the patient as the end date for Medicare reimbursement approached. This practice, coupled with the time lag involved in enrolling patients in the PACT program, reduced the time available for the PACT program to conduct caregiver assessments and to initiate care plans. For almost half of the intervention group, there were 2 days or fewer between the caregiver assessment and the nursing home discharge—a small margin for reducing the length of stay. In a steady-state program it might be possible to incorporate the nurse assessment into the facilities' admission Minimum Data Set assessment. This could reduce nursing expenses and the time lag in scheduling the caregiver assessment, and it could give the case manager more time to implement a care plan. Whether more time would have reduced the number of nursing home days in the current sample is unknown, but more time could likely be helpful in facilities having historically low patient-discharge rates.
Beyond the issue of time, there is the nursing home's interest in retaining patients who are receiving Medicare coverage and in discharging those without this source of payment. One approach for minimizing these incentives would be to work with a Medicare or Medicaid HMO, which would be paying for the skilled care–rehabilitation days and would have a financial incentive for reducing Medicare-paid days of care.
Another PACT program element that proved problematic was the patient selection criteria. Given that 75% of the comparison sample was discharged without PACT assistance, and that almost 20% of the PACT intervention group and 33% of the comparison group declined case management after discharge, there is ample evidence that not all caregivers "needed" this assistance—even though they had originally applied for it. Withdrawals from case management, coupled with the comparable postdischarge outcome status of those with and without PACT case management, suggest room for improvement in the identification of caregivers who require (as distinct from those who find it helpful but not necessary) the enhanced assistance of this program. Postparticipation exploratory interviews were conducted during PACT's development phase to ascertain consumer satisfaction with the program (Graham et al., 2005). Results from these qualitative interviews suggest that early withdrawal from PACT reflects judgments about not needing the assistance, rather than dissatisfaction with it. Program design could likely benefit from more research into the family and patient needs following a nursing home discharge.
Finally, we recognize that the development and persistence of nursing home transition programs is contingent on reducing operating expenses and the ability to offset these expenses with Medicaid or other savings. As implemented, the PACT program was not efficient. Recruitment and administrative costs were excessive and out of proportion to the program enrollment. Clinical staff were also underused and could have accommodated double or triple their case loads. The staffing and case loads require further investigation and adjustment. More important than operating expenses in sustaining programs is whether the nursing home discharge and transition functions can be integrated into programs that have a vested interest in reducing nursing home expenses. We earlier suggested integration into managed care programs as a means of achieving Medicare savings. Another approach would be to incorporate the PACT-like functions into programs such as Medicaid personal care or a home- and community-based waiver service. A third method would be to have nursing homes continue the financing of discharge planning, but to transfer this function from nursing home staff to a freestanding discharge-planning–continuity-of-care program.
|
Footnotes |
|---|
1 Department of Social and Behavioral Sciences, University of California, San Francisco.
2 Center for Community Wellness, University of California, Berkeley.
Decision Editor: Nancy Morrow-Howell, PhD
Received for publication July 11, 2005. Accepted for publication November 4, 2005.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussion and ConclusionsReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. E. Gaugler, A. M. Pot, and S. H. Zarit Long-Term Adaptation to Institutionalization in Dementia Caregivers Gerontologist, December 1, 2007; 47(6): 730 - 740. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
