LOS ANGELES — End-of-life spending for Medicare beneficiaries varies widely based on geographic location, but individual patient preferences do not drive these regional variations, Dr. Amber E. Barnato reported at the annual meeting of the Society of General Internal Medicine.
For example, average per capita costs in the last 6 months of life among beneficiaries in Portland, Oregon, total $9,600, compared with $2,400 in Los Angeles, said Dr. Barnato, of the University of Pittsburgh.
A previous national survey with structured vignettes asking about patient preferences for treatment found that doctors in high-intensity regions are more likely to recommend tests and refer to specialists, and are less likely to recommend hospice care (Ann. Intern. Med. 2003;138:288–98).
The goal of her study was to determine if these different end-of-life practice patterns could be explained by patient preferences. The investigators surveyed a national probability sample of fee-for-service Medicare beneficiaries aged 65 years or older. Potential subjects were identified from the Medicare beneficiaries database for the entire United States; a sample of 3,845 people were asked to participate in the study, and 2,515 (65%) completed the survey.
The researchers asked respondents to imagine that they had less than 1 year to live because of a serious illness. Given this scenario, participants responded to questions about their preferences for active treatment, ventilator use, and palliative care.
The mean age of respondents was 75.6 years, 42% of the respondents were male, and 82% were non-Hispanic white, 7% black, and 5% Hispanic.
Overall, 44% of participants said they worried about getting too little treatment in the last year of their lives, whereas 49% worried about getting too much treatment. A total of 16% would prefer potentially life-prolonging drugs that made them feel worse all the time, while 75% would prefer palliative drugs, even if they might shorten life.
Thirteen percent of respondents would prefer mechanical ventilation if it would extend their life by 1 week, and 22% would prefer MV if it would extend life by 1 month. A large proportion of respondents thought the likelihood of returning to normal activity after being on a ventilator was high, according to Dr. Barnato. “There seems to be a correlation between wanting MV and believing that it is effective in returning you to normal activity. I think it's hard for patients to conceptualize what these things mean unless they have a family member who has been in the ICU,” she said.
The spending calculations were done with standardized prices—the End-of-Life Expenditure Index—in each U.S. hospital referral region, and the analyses were adjusted for age, gender, and race.
The original hypothesis, that individual preferences drive the regional variations in end-of-life spending, was not borne out by the data.
In general, participants' concerns about and preferences for end-of-life treatment were not significantly different among quintiles of the End-of-Life Expenditure Index, although in higher-intensity regions respondents were less likely to want palliative drugs that might shorten life. This exception no longer held, however, once multivariable analyses were adjusted for sociodemographic variables and health status.
A major limitation of this study was the 35% nonresponse rate. It is possible that the original hypothesis was not borne out because patterns of nonresponse differed by region.
The survey showed that almost all of Medicare enrollees worry that the amount of treatment they want in their last year of life will not align with what they will actually receive. When asked to choose, most of the participants said they would prefer treatment that would ease pain rather than extend life.