Advance Directives for Nursing Home Residents: Title and subTitle BreakAchieving Compassionate, Competent, Cost-effective Care

Having an advance directive, a living will, and a durable power of attorney¹ has been proposed as means for a competent person to legally assert his or her autonomy before he or she becomes unable to make decisions. Whether promoting the completion of an advance directive saves resources otherwise squandered on care not wanted by the dying patient remains a tantalizing question.^{2 - 3} Medicine and society are now faced with an aging population that is slowly dying of chronic ailments and the question of whether too much is spent on the dying. Such concerns have been fueled by the statistic that 28% of Medicare program expenditures were accounted for by 6% of persons who died during that year.⁴

The key, policy-relevant hypothesis behind advance directives proposes that if an intervention enhances a person's right to choose, dying persons will not opt for technological (and costly) medical care, physicians will honor that choice, alternative palliative care will be less costly, and an ethically defensible saving of squandered resources will result. Three nonexperimental studies have substantiated this hypothesis,^{5 - 7} while data from 2 randomized trials did not find reductions in medical costs.^{8 - 9}

In this issue of THE JOURNAL, Molloy and colleagues¹⁰ report the results of their well-conducted study, in which nursing homes matched in pairs on several key characteristics were randomly assigned so that one of the nursing homes in the matched pair received systematic implementation of an advance directive program and the other adhered to existing policies. The intervention consisted of the Let Me Decide directive, which included instructional and proxy components that provided a range of choices for levels of care for a serious illness, resuscitation, and nutritional support. The intervention also included education programs for nursing home staff that included video tapes and in-service training, use of a "health care facilitator," the requirement that physicians must review and sign the directive, and steps to ensure that the directive was prominently placed in the medical record and transferred with the patient to the acute care hospital. The intervention targeted nursing homes with advance directive completion rates of less than 25%. All these components most likely were critical to the impressive reduction in the rate of hospitalizations (0.27 per control resident compared with 0.48 per intervention nursing home resident) and the documented reduction in resource use (saving approximately Can $1700, or about US $1200, per patient). These features of the intervention combined—none of them in isolation—likely contributed to this dramatic reduction in hospitalization for nursing home residents.

High-quality end-of-life care includes much more than advance directives for nursing home residents. In addition to advance directives, the residents in the intervention group should have been provided with appropriate palliative services intended to reduce pain and other physical discomfort, improve functioning, and help resolve issues of life closure such as spirituality and saying good-bye. Ultimately, medical decisions ought to reflect nursing home residents' informed preferences and values. Although the findings of Molloy et al suggest progress in these areas, 2 concerns should be noted.

First, the study did not collect information on palliation of symptoms or quality of life. This is an important limitation. Bernabei et al¹¹ have shown that 1 in 4 patients with cancer in a nursing home did not receive any analgesia for daily pain. Advance directives are not an end in themselves; rather, they must be seen as part of the process that includes ongoing communication about the goals of care and the development of contingency plans that will ensure both that preferences will be honored and appropriate palliation provided.¹²

For example, consider Mrs M, an 82-year-old woman who is dying of chronic obstructive pulmonary disease. It is important for her physicians to understand Mrs M's thoughts and perceptions in living with her disease. Has she reached a turning point, such that treatment should focus on comfort even if it shortens her life? Based on conversations with her about that issue, goals of care can be clarified and decisions concerning intubation, resuscitation, and hospitalization documented in an advance directive and in the medical record. Arriving at and documenting those preferences are the components of the first of 2 critical steps. The second step is development of contingency plans that will ensure that her preferences will be honored. If the physician does not anticipate the problems that Mrs M probably will encounter and provide appropriate plans for palliation, Mrs M might die gasping for breath or be transferred to an emergency department where her preferences could well be ignored. Future interventions must focus both on the completion of advance directives and the development of contingency plans that ensure appropriate palliation.

A second concern is the degree to which preferences stated in the advance directive intervention used in the study by Molloy et al was critical for achieving the reduction in hospitalizations. The directive, as described in the article, collects information on whether persons want palliative care, resuscitation, intravenous feedings, intensive care, and types of feeding. The question remains how often preferences were stated in advance regarding the circumstances under which that patient wanted or did not want to be hospitalized. Previous research in the United States has found that orders such as "do not hospitalize" were rare¹³ and that "do not resuscitate" orders were associated with striking reductions in hospitalizations.¹⁴ The finding of Molloy et al that hospitalization rates were reduced by nearly one half is ethically defensible only if that result reflects the informed preferences of a competent nursing home patient or an appropriate surrogate decision maker and if the decision not to hospitalize did not reduce the quality of end-of-life care.

The authors examined nursing home resident and surrogate satisfaction with health care received and involvement in decision making at the beginning of the study and at 6, 12, and 18 months after the intervention. They found no statistically significant change in reported satisfaction in either the cognitively intact nursing home residents or the surrogates of cognitively impaired patients. This is reassuring. However, future studies should examine whether medical care was consistent with the written preferences contained in the advance directive and whether family members believed that medical care was consistent with the patient's preferences.

Nursing homes will play an increasingly important role in care of the dying. Of all deaths in the United States in 1993, 20% occurred in US nursing homes, and the proportion of deaths occurring in nursing homes is expected to increase to 40% by 2020.¹⁵ Additional research, such as that reported by Molloy and colleagues, that focuses on improving the quality of end-of-life care for patients in nursing homes is urgently needed. Future interventions should consider explicitly linking the process of negotiating the goals of care, completing advance directives, and formulating plans that ensure appropriate palliation. Physicians and all health care professionals involved in caring for patients at the end of life must strive to provide holistic medical care that is competent, cost-effective, and compassionate.