Careful delineation of types of functional decline may help tailor expectations, care plans, and organizational approaches to the care of patients at the end of life. Recognizing that functional decline before death differs by age and medical condition, Lunney and colleagues developed a study to understand the influence of different diseases on patterns of dependency before dying.

Using data from the Established Populations for Epidemiologic Studies of the Elderly (EPESE), the authors identified 4,190 decedents who happened to be interviewed within one year of dying. These participants had been asked about their ability to perform activities of daily living, other physical tasks, and health issues. Participants were divided into 12 cohorts, based on the number of months between their final interview and death. The groups were also divided into four categories related to the cause of death: cancer, organ failure, frailty, and sudden death, with methodologic measures taken to account for overlap. Patients who could not be categorized were included as the “other” group.

There were 649 patients (15 percent) in the sudden death group, who were younger than the mean age; 897 patients (21 percent) in the cancer group, constituting the youngest group; 817 (20 percent) organ failure patients, 837 patients (20 percent) who were frail, the oldest group in this study; and 990 patients (24 percent) in the other group.

While the overall level of dependency was greater with increasing age in the last year of life, the trajectory of decline in activities of daily living exhibited internally by each group was similar, irrespective of age. Those in the sudden death group remained independent and retained function until death. Cancer patients remained functional until about three months before death, at which point they became markedly more disabled. Organ failure and frail patients also exhibited substantial decline in the last three months, but organ failure patients had an erratic decline in the last year overall, and frail patients were more disabled relative to the other groups for the entire year before dying. The other group showed a modest, gradual decline in the last year, most closely matching patients with ischemic heart disease, when these patients were analyzed as a separate group. A higher proportion of the other group had ischemic heart disease than the full decedent pool. Overall, the frail group was eight times more likely to be dependent than those who died suddenly, even after controlling for age and other factors.

These findings suggest that cancer patients are the only ones who fit into the hospice model of expected death. Given that only 23 percent of the U.S. population dies of cancer, end-of-life care should focus more on patients who are likely to die unpredictably, including patients with chronic illnesses and those who become increasingly frail, even without a diagnosis of life-threatening illness.