Korbin Liu

Predicting Nursing-Home Admission and Length of Stay: A Duration Analysis

Information on who enters nursing homes and how much nursing-home care people use continues to be important for the consideration of various long-term care policies. Research on how characteristics of noninstitutionalized disabled elderly persons are related to both risk of admission and length of stay is presented here. Hazard models were developed using data from the 1982-84 National Long-Term Care Survey. Personal characteristics of elderly persons and state policy variables had varying effects on nursing-home use in a 2-year period.

Determinants of Transitory and Permanent Nursing Home Admissions

This article describes an analysis of data from the 1982-84 National Long-Term Care Demonstration Project to estimate the risks of any nursing home admission, a temporary or transitory admission, and a permanent admission. Using a multinomial logit model, the relative predictive power of several individual characteristics on nursing home use and admission type were evaluated. It was found that the cognitively impaired subgroup was at the greatest risk of entering a nursing home, especially on a permanent basis. The results also demonstrated that the combination of cognitive impairment and functional impairment further increased the risk of a nursing home admission, particularly a permanent one. Other subgroups that had high probabilities of experiencing a nursing home admission were whites, nonhomeowners, those living alone, and those with prior nursing home stays. The findings identified several aged subgroups that were at no greater risk of nursing home admission regardless of admission type: older persons who were unmarried, had a low income, had no assets, and those on Medicaid.

Risk of entering nursing homes for long versus short stays.

Research on the risk of nursing home use has generally not distinguished between risk for long versus short stays. This paper presents an analysis of data from the 1982–84 National Long-Term Care Surveys to identify characteristics of disabled persons that predicted one or the other type of stay. Measures that are recognized to be strong predictors of nursing home use in general, such as dependencies in activities of daily living and cognitive impairment, were not significantly associated with admission for short stays. Results from the multivariate analysis were also used to make projections of lifetime risk of long and short stays. Approximately one-third of lifetime nursing home risk applies to stays of 90 days or less.

Disability and Medicare Costs of Elderly Persons

Rapidly growing Medicare expenditures, which are already expected to exceed revenues by the year 2002, are likely to accelerate as a result of the increasing numbers of very old beneficiaries (85+ years), who are particularly at risk of being disabled. The concern for Medicare is that elderly persons with disabilities incur higher than average health care costs because of the underlying causes of their disability, which are usually chronic medical conditions (Manton and Stallard 1992). Such conditions often require long-term care, but they also tend to induce illnesses requiring acute care. This study analyzed data from the Medicare Current Beneficiary Survey to examine the relationship between disability and Medicare costs. It examined relations between levels of disability and costs, interactions between disability and other beneficiary characteristics, and changes in the level of disability and costs. (The Milbank Quarterly 1997 December; 75(4): 461-493).

The Length-of-Stay Pattern of Nursing Home Admissions

The length of stay (LOS) pattern of nursing home admissions has not been directly measured because the long duration of stay of some patients (e.g., 25% of persons admitted reside in homes for longer than 1 year) makes it impractical to follow specific admission cohorts until the members have been discharged. By applying life-table methods to data on current and discharged residents from the 1977 National Nursing Home Survey, the authors were able to generate estimates of this distribution. They estimated that the typical nursing home admission initially has an expected LOS of 456 days, but, more interesting, the remaining LOS is expected to increase, in the intermediate term, with the amount of time spent in the facility. They also examined LOS distributions of subsets of the national nursing home population, and illustrated potential applications for this new information for policy formulation.

Hospital-Based and Freestanding Skilled Nursing Facilities: Any Cause for Differential Medicare Payments?

Differential Medicare payments for hospital-based and freestanding skilled nursing facilities (SNFs) were eliminated by the SNF prospective payment system initiated in 1998. Closures and high negative margins of hospital-based facilities have prompted consideration of the need to revisit payment adjustments for this group of SNFs. We examine case mix-related and other factors behind the cost differences between hospital-based and freestanding SNFs. Some payment adjustment, notably for nontherapy ancillary services, may be reasonable for the short term.

Morbidity, Disability, and Long-Term Care of the Elderly: Implications for Insurance Financing

A scarcity of empirical information to specify appropriate provisions and base rates for coverage has hindered the development of long-term-care (LTC) insurance. Data from three nationally representative surveys on the prevalence of morbidity and functional limitations among the elderly population suggest that health status among older Americans is highly dynamic, especially at higher disability levels. The bioactuarial data may help insurers define potential markets of purchasers of policies, and identify the numbers of persons with disabilities severe enough to trigger use of benefits. If the accuracy of individual service predictions could be increased further, reserve requirements and overall costs to LTC insurance carriers might be reduced.

Projecting Chronic Disease Prevalence

Health and long-term care planning for an aging population is an important and necessary function for both the public and private sector. Unfortunately, efforts at planning have often been limited by difficulties in making estimates and forecasts of chronic disease prevalence in the population. These difficulties are a direct result of the natural history of chronic diseases that normally have long presymptomatic stages. Because of this characteristic, predictions of the magnitude of chronic disease prevalence in the United States often fail to reflect an important dimension of the health state of the population: that there is a sizeable proportion of chronic disease prevalence and risk existing in the population in a preclinical phase. The authors present a strategy for obtaining more complete estimates of chronic disease prevalence. This approach entails the creation of an illness–death model representing the natural history of individual chronic diseases and the application of the model to infer morbidity incidence and prevalence patterns from national mortality statistics. They illustrate the approach with an example of lung cancer and discuss the applications of the outputs of the modeling strategy for health care resource planning.

Case-mix differences between hospital-based and freestanding skilled nursing facilities. A review of the evidence.

Many health care services are available in both hospital and freestanding settings. Hospital-based providers are usually much more expensive than freestanding providers. According to Medicare data, costs in hospital-based skilled nursing facilities (SNFs) are twice those of freestanding facilities. While critics charge that this is the result of inefficiency, hospitals counter that higher costs are caused by treatment of sicker patients and provision of higher-quality care. This paper analyzes the research on the case-mix differences between hospitalbased and freestanding SNFs. On the basis of this analysis, it appears that hospital-based facilities tend to serve more severely ill patients (i.e., have a more difficult case mix) than do freestanding facilities. Case-mix differences, however, appear to explain less than half of the cost differential between the two types of facilities.

How Do They Manage? Disabled Elderly Persons in the Community Who Are Not Receiving Medicaid Long-Term Care Services

ABSTRACT Objective: To expand our understanding of how low-income functionally impaired elderly persons are able to remain in the community. Data Sources and Study Setting: In-person and telephone interviews with 25 elderly individuals who applied for but did not enroll in Connecticuts Home Care Program for Elders (CHCPE). All met the states nursing home level-of-care criteria. Study Design: In-depth discussions with a small, purposefully selected sample of functionally impaired elderly persons in the community. Principle Findings: Many sample members with very high levels of impairment and multiple chronic health conditions remained in the community without CHCPE services because of Medicare home health services combined with extensive levels of informal care. Some sample members, particularly those with more limited informal care networks, did not receive the level of care that they needed. Virtually all were at high risk for medical complications, hospitalizations for acute illnesses, falls, and further loss of functioning. Further, in many cases, informal care networks were overextended, stressed and vulnerable to breakdown. All but a few of those we interviewed were not receiving services through the waiver program for financial reasons. Most met Medicaids income criteria but had assets that exceeded Medicaids