Hospitalization of Nursing Home Residents: The Right Rate?

In this edition of the Journal,1 Fried and Mor describe the hospitalization experience of a group of long-term residents in a large for-profit nursing home chain. Very frail residents, especially if they had congestive heart failure or respiratory failure and did not have dementia, were most likely to be hospitalized. Most geriatricians will not find this surprising. However, given the paucity of research, perhaps no finding could have been surprising. Very few descriptions of the hospitalizations of nursing home residents have been reported, and they show substantial variation in rates2–4 and little reason for confidence that the hospital is particularly good for the well being of very frail persons.5,6 Hospitalization rates of older people vary across regions in ways that ordinarily tell more about the supply of services in the care system than about the care needs, personal situations, or preferences of the patients.7,8 What is the right rate? Who should be hospitalized? Surely some such decisions are easy. Virtually everyone would encourage hospitalization of a nursing home resident with severe arthritis who has an acute myocardial infarction with mild congestive heart failure and new onset of atrial fibrillation. Arrhythmias could be prevented or treated, and most likely the patient would be able to avoid serious iatraogenic illness. Another person may be near death, without symptoms, and receiving effective hospice-style care in the nursing home as she and her family wanted. She should not be moved for almost any complication. In these cases, decisions about hospitalization are easy, but most situations are likely to be more complicated. Physicians making decisions for acutely ill nursing home residents are confronted with a substantial array of idiosyncratically varying conditions. What is needed to make it possible to provide care in the nursing home for patients with a decline in health status? A list would include these conditions: nurses are available at all times, nurses can communicate with the physician and the families, advance care planning is comprehensive and reliable, symptom-relieving drugs are readily available and reliably administered, a competent physician or nurse practitioner is available on-site most days, routine laboratory tests and X-rays are available most days, dying well is seen as a worthy accomplishment by the care team, usual care is competent and caring, and financing allows some margin for graciousness. If opposite conditions apply, many nursing home residents are going to have to compromise comfort, competence, or confidence in order to remain in their nursing home settings. Optimal care of nursing home patients clearly requires substantial planning in advance of crisis. Only in this way can various considerations and perspectives be brought to bear so that the treatment plan can serve the patient's interests (as the patient does, or would have, defined them). The lack of serious advance planning in the series that Fried and Mor report is obvious. Only 32 people of 3782 for whom it might have been considered had a “do not hospitalize” order. Advance care planning requires knowledge not only of what the patient might want in some hypothetical setting, but of what really is likely to happen in the specific environments available to this patient at the time he or she needs it. Many geriatricians have been fortunate in that they have participated in nursing homes in which families, caregivers, and all concerned were proud of the assiduous and competent care that sick persons received where they lived, whether it consisted mainly of diagnosis and treatment aimed at restoring health or it focused on supportive care with the expectation of dying. In such settings, fewer than one in five persons dies in a hospital, and hospitalization rates can be less than those of the general Medicare population even though the nursing home patients are much more seriously ill. However, nursing homes have recently had sicker residents, more regulatory requirements, and reduced funding. In many areas, nursing homes are under more pressure to encourage hospitalization. With little continuity in nursing staff, uncertain Medicaid reimbursements, problems mobilizing services to the nursing home setting, and suspicion of the quality of care for relatively stable patients, responsible physicians often have to resort to unloading on the hospital the time-consuming problems presented by acutely ill nursing home residents. During this period of change, all concerned are frustrated and feel that they are being forced into an uncomfortable and unfortunate accommodation. However, people have an astonishing ability to accommodate. We run the risk of allowing these worsened conditions to come to be seen as acceptable, since they will seem to be the best that can be done. See also p 265 This setting of standards by happenstance, politics, and finances is not, however, the best we could do. Many of us will spend time in a nursing home, or our loved ones will. At present, Medicaid rates of pay strongly determine the environment of most long-term care institutions. Effectively, rates end up declining slowly over time as the distaste of the public for funding nursing homes is abetted by a broad array of more attractive competing priorities — schools, highways, or tax cuts, for example. Unless a nursing home has some protected source of income, such as endowment or charity, the rates finally get too low to ensure minimally decent care. This often engenders a round of scandal and reform, which produces, as a side effect, a rise in the Medicaid rates of pay. Researchers and practitioners should develop the procedures that would allow us to set patterns and standards pro-actively. Doing so would require modeling alternative patterns, measuring their outcomes comprehensively, and supporting serious quality improvement interventions. Such an agenda might end up defending substantially disparate rates of hospitalization. Areas with relative oversupply of hospital beds and underdevelopment of excellence in nursing home care, for example, might have to generate high rates of hospitalization in order to serve the frail and sick with decency, at least until nursing home care could be improved. Regions with better nursing home care and lower hospital bed supply would do well with much lower rates. Likewise, which nursing home patients utilize hospitals might depend on local conditions as much as the requirements of treatment. Thus, a region that is reacting to nursing home abuse might impose regulations or enforcement encouraging additional hospitalization of the frail, and a region that has an oversupply of nursing home beds might have younger and healthier residents who more often desire aggressive diagnosis and treatment. One can be grateful for the careful descriptions that Fried and Mor provide. Their findings do give some points of reference and comparison. However, one cannot know whether these rates were optimal for these patients or if any other setting should have higher rates, or lower, or just different for different residents. Assessing what one might hold to be benchmarks of good care requires knowing what could be accomplished in well described conditions and what the options are in the practical realities of each environment.