Institutional Care at the End of Life

Abstract

this issue of Medical Care, Dy and colleagues1 present an analysis of linked National Long-Term Care Survey (NLTCS) and Medicare data to examine use of Medicaresupported institutional services at the end of life. They examine 3 types of institutional services covered by the Medicare program: hospice, inpatient hospital and skilled nursing facility (SNF). There is no shortage of advocates for less aggressive care at the end of life. Many note that patients and providers alike report preferring a palliative approach to care of the dying. Why then are palliative services so difficult to integrate into care? We see in this report that many of our measures of care intensity are relatively crude: whether patients are dying in an institutional setting that is designed to emphasize palliative care (ie, hospice) or not; whether they are hospitalized in the last year of life; and expenditures by the Medicare program in the last year of life. These measures, unfortunately, reflect the level of resolution available for many of our population-based studies. Although we have few measures of nonhospice palliative care use, we have even fewer of palliative care need. Most of our studies use a look-back approach-start at death and look back 1 year. These studies face challenges that are well-recognized (see Emanuel and Emanuel,2 for example), but few other options for studying end-of-life care are presently available. This particular example asks about use of institutional care at or near the time of death. This article points to 2 major aspects of institutional care at the end of life: whether there is a link between rates of in-hospice and in-hospital deaths; and what the use of dual hospice/nursing home care tells us about the care provided in these other institutions. Comparing samples from 1989 and 1999, Dy and colleagues1 report important changes. First, hospice use increased for elderly without chronic ADL impairments, for those with chronic ADL impairments, and for elderly who were institutionalized. This pattern of increasing hospice use over time is consistent with multiple other reports showing an increase in the use of hospice before death for elderly. In contrast to these encouraging increases in hospice care at the end of life, the percentage of elderly who died in-hospital declined variably depending by group. For those without ADL impairment, there was a considerable decrease; for those with chronic ADL impairments and for the institutionalized, in-hospital death rates declined only slightly. For all 3 beneficiary groups we see an overall increase in the percentage of deaths occurring while receiving Medicare-funded institutional services. Others examining urban/rural differences in location of death have shown a strong trend in hospice use at death with lower use in rural areas than urban ones. This trend is in contrast to the percentage of beneficiaries dying in-hospital, which is stable across the urban/rural continuum.3 These different trends in in-hospice and in-hospital deaths have been attributed to the differential availability of inpatient beds and hospice services. However, combined with the current study, the evidence points to different drivers of hospital use at the time of death and hospice use at the time of death. Ideally, we would see this shift from in-hospital to in-hospice deaths and be able to conclude that through increasing hospice availability institutional care at the end of life is becoming less aggressive. However, the current study does not point to as strong a link between increases in hospice use and declines in the use of other Medicare-funded institutional care as we might hope. If, as many studying end-of-life care seem to believe,