Sick To Death > Chapter 2 > Working Out Patterns of Costs
What we know about the costs and use of health services near the end of life presents an interesting picture. The proportion of Medicare costs spent on the last year of life, or the last month, has stayed nearly constant for two decades. Other than for chronic renal failure, average costs for the last year decline sharply with age but are quite similar across all major diagnoses. End-stage renal failure has always been a costly way to live and to die, with average Medicare costs over $50,000 for all years, including the year of death (Hogan et al. 2000). All other major causes of death for people ages sixty-five to seventy-five cost about $26,000 (in 1997 dollars) in their last year. The fact that the major ways to die converge on a small range of average costs is quite curious. Perhaps the care system does not commonly implement more services than providers and caregivers consider decent and appropriate. It is an intriguing possibility that awaits more research.
Whereas being sick enough to die (and being covered by Medicare) trumps most of the biases that relate to wealth, race, income, and gender (Shugarman et al. 2004), variations in patients' age and in geography continue to correlate with major differences in costs. The decline in hospitalization and physician services as patients age may reflect fewer opportunities to affect frailty and dementia. Of course, the decline may also reflect age bias in offering care or older patients' and families' increasing preference for comfort. The differences in use by similar patients in different parts of the country might arise from differences in the availability of services, a possibility that has substantial support in the literature (Last Acts 2002; Dartmouth Atlas of Health Care 1999).
Perhaps the very different patterns between areas also reflect real differences in what communities feel is essential to good care and in how different parts of the care system are set to respond (Dartmouth Atlas of Health Care 1999). In Oregon, for instance, providers are proud of keeping very sick people well supported at home. Hospice is readily available, around the clock, almost everywhere. Adult foster-care homes aim to keep their residents in place. Patients, their families, and care teams see going into the hospital with an incurable condition as a loss. In New York, the same act of treating sick people where they live might lead to allegations of mistreatment by the nursing facility or home-care team. Hospice care at home is not common. In my experience, New Yorkers often feel that hospitals are best for treating sick people and that nursing homes and other settings for geriatric care are generally more risky settings for care. Thus, the variations in practice in different areas might arise from variations in the supply of services and possibilities for income, or they might depend on communities' expectations and capabilities. Most likely, factors such as these are mutually supportive and together create barriers to change. Nevertheless, the fact that geography has a greater effect on costs than diagnosis or wealth should help illuminate possibilities for change: sometimes to address inadequate services, often to challenge costly approaches on the basis of efficiency.