The prospective payment system (PPS) was implemented in 1983 to limit the cost of hospitalizations under Medicare. At the same time, other social and economic changes occurred, including community options programs and the limitation of Medicaid reimbursement in nursing homes to persons needing higher levels of care. As expected, the PPS has shortened the average length of stay (ALOS) in acute care facilities.
Sager et al studied nursing home patients in Wisconsin's Medicaid program before and after the implementation of the PPS system.1 The ALOS dropped from 10.8 to 8 days. During the same period, there was a 73% increase hi the rate of hospitalization. A study by the Rand Corporation found that implementation of the PPS led to a greater decrease in hospital ALOS in patients discharged to nursing homes than in the overall Medicare population.2
The decrease in ALOS in acute care has had far-reaching consequences. Fitzgerald studied the outcome of elderly hip fracture victims in a Midwestern urban teaching hospital before and after the implementation of the PPS.3 The hospital ALOS decreased from 21.9 to 12.6 days. At the same time, average distance walked prior to discharge decreased from 27 m to 11m. The proportion of patients discharged to nursing homes increased from 38% to 60%. Of even greater concern was that the proportion of patients remaining in nursing homes 1 year after hospitalization increased from 9% to 33%. This study not only found a shift of burden from hospitals to nursing homes under the PPS, but it also raised questions about quality of care.
A similar study confirmed the decrease in ALOS, but failed to find an increase in discharges to nursing homes or differences in ambulation before and after the implementation of PPS.4 After the implementation of a PPS in a Wisconsin Veterans Administration teaching nursing home, Tresch et al reported nursing home admissions with greater severity of illness requiring more frequent hospital readmissions.5 Many of the hospital readmissions resulted from problems that were not adequately resolved or even identified during the previous admission. Harron and Schaeffer reported that patients discharged from hospitals required higher numbers of nursing procedures within 3 days of nursing home admission since implementation of the PPS.6 Kayser-Jones et al speculated that 48% of hospital admissions could have been prevented by adequately trained staff and additional services in the nursing home, such as intravenous therapy and intravenous antibiotics.7
The new payment system has also been associated with changes in location of death. Sager et al observed a shift in the location of death in Wisconsin from acute care facilities to nursing homes.1 Between 1982 and 1985, mortality rates in nursing homes increased 26.2% while hospital deaths declined by 9.5%. Lyles also reported a 20.5% increase in the number of deaths in nursing homes in Portland, Oregon, under the PPS system.8
Hie combination of shorter lengths of stay in acute care facilities, increasing rates of acute care rehospitalizations, the shift in location of death from the hospital to the nursing home, and higher numbers of required nursing procedures suggests that more seriously ill patients are being cared for in nursing homes. We explored resident acuity in our nursing home under the PPS by reviewing acute care transfers for 12 months. In this article, we present data on recidivism, mortality, location of death, and length of stay (LOS). We found that a subgroup of 10% of our residents accounted for 54% of the admissions to acute care facilities.
The Wisconsin Veterans Home is a long-term care facility with an average daily census of 68 1 (500 men and 1 80 women), a mean resident age of 77 years, and ALOS of 5.33 years. On-site primary medical care is provided by four physicians who make daily "sick" rounds. Laboratory, x-ray services, and limited intravenous therapy are available within the facility, as are pharmacy, dietary, and rehabilitation services. Registered nursing care is provided in 13 nursing units.
1986 ACUTE CARE ADMISSIONS
The discharge summaries of all acute care transfers between January 1, 1986 and December 31, 1986 were analyzed by a gerontological nurse practitioner (GNP) and a geriatrician. Data included name of hospital, primary diagnosis precipitating admission, and LOS. Thirtyday recidivism was recorded through February I1 1987. Variables generated from these data included length of time between admission and readmission, the relationship between the initial admission diagnosis and readmission diagnosis, and location of death. The location of all deaths from 1981 to 1986 was provided by the medical records department.
A GNP and a geriatrician categorized the readmissions into three groups according to the relationship of the initial and subsequent admitting diagnoses: same, related, and unrelated. The "same" diagnosis group was defined as those readmissions with the same pathological process in the same organ or a direct complication of treatment initiated in the previous hospitalization. The "related" group included three situations in which the initial admitting pathological process potentiated a problem in the same organ system (eg, an admission for myocardial inunction and a subsequent admission for congestive heart failure); directly caused a problem in another organ system (eg, an admission for hip fracture followed by an admission for lower extremity phlebitis); or was a normal follow-up procedure, hi the "unrelated" group, the second readmission had no obvious relationship to the prior admission.
Length of Stay
The ALOS in the community hospitals (81% of admissions) was 6.34 days (Sager et al reported an ALOS of 8 days'). In the Veterans Affairs hospitals (19% of admissionsX ALOS was 2 !days (Tablet).
Rates of Admission
During the study year, 847 individuals resided in the facility either all or part of the year, with an average daily census of 681. TWo hundred sixty residents (31%), average age 75.9 years, were transferred to acute care for a total of 378 admissions. One hundred seventy-four residents had single admissions, and 86 residents were admitted a total of 204 times with between two and five admissions per individual (Table 1). Therefore, 10% of the resident population (86 of 847) accounted for 54% of the total hospitalizations.
Readmission Within 30 Days
Forty-three individuals (half of the residents with readmissions) were readmitted 50 times within 30 days of the previous discharge. The average age of this group was 75.3 years. The ALOS for the initial admission was 6.3 days versus 7.4 for subsequent admissions (Table 1).
The mortality in the 43 residents readmitted within 30 days was 40% (17 residents) during 1986. Nine of these residents died hi the hospital, whereas six died in the nursing home within 15 days of discharge (Table 2).
Twenty-seven (54%) of these readmissions were in the "same" diagnostic category, 12 (24%) were in the "related" category, and 1 1 (22%) were unrelated to their previous admission. The reasons for readmission hi the "same" and "related" diagnostic groups are noted in Table 2. Cardiopulmonary conditions caused 32% (16 of 50) of these readmissions and were associated with 47% of the deaths (7 of 15) in the hospital or within 15 days of hospital discharge. Included in these deaths were six residents who were readmitted three to five times during that year.
On review of the medical records and discharge summaries, no dominant nursing or medical management problem was identified as a cause of the 30-day readmissions. Multiple disease states and the complex interaction between them were the major factor identified underlying the readmissions.
Mortality and Location of Death
In 1986, 14% (121 of 847) of the Wisconsin Veterans Home residents expired. Twenty-three percent (59 of 260) of the residents who were admitted to acute care died during that year; 21 died while in the hospital. Of the 38 who died after hospital discharge, 20 died within 15 days of discharge. Sixty-two decedents had not been admitted to acute care during the year.
From 1980 to 1983, when more residents were referred to Veterans Affairs hospitals, 56% of deaths occurred within the nursing home (Table 3). After the institution of the PPS, the percentage of deaths within the facility increased significantly to 80% (χp 2 = 48.604, P <-001).
This study of readmissions in a nursing home has supported previous findings showing increased resident acuity in nursing homes after the implementation of a PPS. The increased acuity is evidenced by complex disease interactions, frequent hospital readmissions, and death within 15 days of discharge from the hospital. Our finding of complex disease interactions as the primary readmission factor emphasizes the need for sophisticated nursing care in nursing homes which are frequently understaffed with undertrained nurses and are accustomed to providing primarily maintenance care.6,9
The increased acuity in the nursing home requires nurses to provide highly skilled physical, functional, and mental status assessments. These assessments are required to appreciate subtle changes in condition before major problems develop. They must be followed with adjustment of therapy, especially upon return from acute care but also during limited acute illnesses of residents who are not hospitalized. These assessments must be accompanied by complete documentation of the alterations in and effectiveness of therapy.
Nurses must also direct multidisciplinary planning in which the small percentage of residents responsible for a majority of hospital admissions are identified. Predictors of rehospitalization, such as disease severity index, diagnosis, and previous hospitalization,10,11 need to be considered during this care planning process. Critical care planning must be done for all nursing home residents to determine the resident's or surrogate's directives for hospitalization, medication interventions, and medical care options, such as life support.
One approach to providing the more sophisticated nursing care needed in the nursing homes of the 1990s is the use of GNPs who are prepared to provide advanced nursing care. The sensitive monitoring of a resident's status provided by a nurse practitioner could prevent the hospitalization of some residents with acute illnesses and chronic disease exacerbations as well as the rehospitalization of others. Kane et al reported that GNPs employed by nursing homes decreased the number of hospital days required by newly admitted nursing home residents.12 The acute care cost savings resulting from the implementation of a PPS could be extended even further through the use of GNPs. Residents could receive some acute care services in the nursing home without the disruption of a transfer to an unfamiliar environment. The GNP could also function as an educator to assist the nursing staff in improving nursing assessment and planning skills.
GROUP READMITTED WITHIN 30 DAYS: HOSPITAL ADMISSIONS AND MORTALITY BY DIAGNOSTIC GROUP*
A major obstacle to improving overall nursing home care is the current reimbursement process for skilled nursing facilities. Regulated staffing time minimums for a given level of nursing care have not changed significantly for Wisconsin nursing homes in the past 15 years. Nursing homes are paid on a prospective basis according to levels of care, ie, skilled or intermediate, using projected payment maximums based on industry norms.
LOCATION OF DEATH 1981 to 1986*
This payment approach has two major problems. First, payment adjustments lag behind actual costs. When the needs of residents at the highest level of care increase, when costs of supplies and salaries increase, or when a nursing home wishes to upgrade its services, payment adjustments may not occur for 2 or more years. Second, the current methods for determining a resident's level of care and need for healthcare resources are generally recognized to be imprecise. Delayed payment adjustments combined with imprecise measures of acuity cause inadequate compensation for nursing homes that respond to a challenge. Consequently, efforts to provide better services to benefit the debilitated resident seriously strain the nursing home's own personnel and fiscal resources.
Nursing home residents readmitted to acute care use personal financial resources as well as those of the nursing home and acute care facility. Initial hospital admission is subject to a deductible of $592 under Medicare. If readmissìon occurs within 60 days, another deductible of $592 is required. Although Medicare will pay 80% of the ambulance transportation if it is medically necessary, hospitalizations are also costly because the remainder of the transportation expenses are borne by individual residents, the operating budgets of nursing homes, and third-parly payers.
A major dilemma and opportunity in cost containment is to better care for residents and prevent rehospitalization. Peer review organizations review Medicare admissions targeting specific admission objectives, including appropriateness of admissions, review of certain Diagnosis Related Group assignments, and réadmissions within 7 days of discharge. Quality objectives, including hospital réadmissions resulting from substandard care, provisions of adequate services, and reduction of deaths or unnecessary procedures, are also reviewed.13 Ru bens te i n et al demonstrated this review system to be a valid and reliable method.14 However, during this study, it was also noted that more patients were discharged from hospitals too soon and hi unstable conditions since the advent of the PPS. Because peer review organizations choose the specific areas or Diagnosis Related Groups to target for review, the peer review organizations could focus their review process on this area to decrease the number of early hospital discharges and potentially decrease the numbers of readmissions.
The use of GNPs to fill the gap between standard nursing home care and acute care created by the PPS would require specific adjustments in the current reimbursement system for nursing homes. GNPs can only provide services in nursing homes if adequate reimbursement is available. OBRA legislation allows third-party reimbursement for GNPs employed by physicians to alternate with the physician in providing visits to the nursing home. However, the system also needs to provide reimbursement for GNPs to provide care as employees of the nursing home. Ouslander has suggested establishing a reimbursement schedule that would cover acute disease exacerbations in the nursing home.15 Another option would be to increase the per them rate to allow nursing homes to upgrade their professional staffing. We need to encourage legislators to consider funding programs that provide better staffing and GNP services in nursing homes to meet the challenges created by the PPS in acute care.
- 1. Sager MA, Leventhal EA, Easterling DV. The impact of Medicare's Prospective Payment System on Wisconsin nursing homes. JAMA, 1987; 257:1762-1766.
- 2. Neu CR, Hanison SC. Posthospital Care Before and After the Medicare Prospective Payment System. Santa Monica, Ca: RAND/ UCLA Center for Health Care Financing Policy Research; 1988. R-3590-HCFA.
- 3. Fitzgerald JF, Moore PS, Dittus RS-The care of elderly patients with hip fracture- N Engl J Med. 1988;319:1392-1397.
- 4. Palmer RM, Saywell RM, Zollinger TW, Emer BK, LaBov AD, Freund DA, et al. The impact of the Prospective Payment System on the treatment of hip fractures in the elderly. Arch Intern Med. 1989; 149:2237-2241.
- 5. Tresch DD, Duthie EH, Newton M, Bodin B. Coping with diagnosis related groups. The changing role of the nursing home. Arch Intern Med. 1988; 148:1393-1396.
- 6. HaiFon J, Schaeffer J. DRGs and the intensity of skilled nursing. Geriatr Nurs. 1986; 7:31-33.
- 7. Kayser-Jones JS, Wiener CL, Barbacela JC. Factors contributing to the hospitalisation of nursing home residents. Gerontologist. 1989; 29:502-510.
- 8. Lyles YM. Impact of Medicare diagnosis related groups (DRGs) on nursing homes in the Portland, Oregon metropolitan area. J Am Geriatr Soc. 1986:34:573-578.
- 9. Nearly 90 of nursing homes arc short staffed. GeriatrNurs. 1990; 11:266.
- 10. Anderson GF, Steinberg EP. Predicting hospital readmissions in the Medicare population. Inquiry. 1985; 22:251-258.
- 11. Fetnke CC, Smith IM, Johnson N. "Risk" factors affecting readmission of the elderly into the health care system. Med Care. 1986; 24:429-437.
- 12. Kane RL, Gamut) J, Skay CL. Effects of a geriatric nurse practitioner on process and outcome of nursing home care. Am J Public Health. 1989; 79:1271-1277.
- 13. Dans PE, Weiner JP, Otter SE. Peer review organizations: Promises and pitfalls. N Eng! J Med. 1985;313:1131-1137.
- 14. Rubenstein LV, Kahn KL, Reinisch EJ, Sherwood MJ, Rogers WH, Kamberg C, et al. Changes in quality of care for five diseases measured by implicit review, 1981 to 1986. JAMA. 1990; 264:1974-1979.
- 15. Ouslander JG. Reducing the hospitalizaron of nursing home residents. J Am Geriatr Soc. 1988:36:171-173.
1986 ACUTE CARE ADMISSIONS
GROUP READMITTED WITHIN 30 DAYS: HOSPITAL ADMISSIONS AND MORTALITY BY DIAGNOSTIC GROUP*
LOCATION OF DEATH 1981 to 1986*