ANN ARBOR, Mich. -- Dying in America is an expensive process, with about one in four Medicare dollars going to care for people in their last year of life. But for African Americans and Hispanics, the cost of dying is far higher than it is for whites.
And despite years of searching for the reason, no one has quite figured out why.
A new study by a University of Michigan Medical School team tried to get to the bottom of this expensive mystery with the most detailed study to date. The team published their findings today in the Journal of the American Geriatrics Society.
Unlike other kinds of health cost disparities, they didn't find that the difference could be explained by differences in patients' income, education, medical conditions or other individual factors such as use of hospice. The differences also persisted after they took into account general health spending patterns in a dying person's area.
The team also took into account specific factors related to the unique circumstances that are present at the end of life - the first time this has been done. These indicators of patients' own preferences still didn't explain the differences in cost.
As a result, the researchers conclude that something bigger, rooted in the healthcare system as a whole, must be going on. After all the other factors were figured in, the cost of that last six months of life was $7,100 more expensive to the Medicare system for blacks, and $6,100 more expensive for Hispanics, compared with whites.
"We need to look harder for the causes of disparities in end-of-life care costs, and look at factors on the provider level and health system level, including family dynamics that may come into play," says Elena Byhoff, M.D., M.Sc., who led the study while she was a Robert Wood Johnson Clinical Scholar at U-M and is now at Tufts Medical Center.
Byhoff and her colleagues, including U-M end-of-life care researchers Kenneth Langa, M.D., Ph.D. and Theodore Iwashyna, M.D., Ph.D., hope that their study will add to the broader conversation about how to encourage good patient-provider-family interactions in patients' final years, including the difficult but important questions about advance planning for end-of-life care. Whites were twice as likely as blacks and Hispanics to have an advance directive document in place in the study.
New Medicare coverage for doctor visits that include such discussions may help, Byhoff says. But providers must also act on patients' wishes as spelled out in advance directives, and families must understand and honor those wishes when the patient can no longer speak for him or herself.
Previous work by Langa and others has shown that end-of-life care costs tend to be lower for patients who have spelled out their wishes to limit treatment ahead of time.
A long-term look
Patient preferences have been seen by many as a key factor in end-of-life cost disparities, but are hard to study. The U-M team found a way to take them into account in their analysis by using data from the Health and Retirement Study, based at the U-M Institute for Social Research.
They gathered data from interviews with more than 7,100 seniors over age 65 with traditional Medicare coverage who took part in the long-term, nationally representative study and died sometime during a 14-year period that ended in 2012.
They also used HRS interviews with the seniors' survivors, which allowed them to find out if the death had been expected, if the patient had discussed their end-of-life treatment preferences with their survivor, and if they had a formal advance directive in place when they died.
Then, the researchers matched each senior's survey data with his or her Medicare data, from doctor visits and hospital stays to prescription drugs and home, hospice and nursing home care. HRS participants consented to this analysis as part of taking part in the study.
The researchers used an End-of-Life Expenditure Index developed by the Dartmouth Institute to correct for regional variations in care costs for people in their last year.
In all, 78 percent of the people studied were non-Hispanic whites, 14.5 percent were non-Hispanic blacks, 4.7 percent were Hispanic and 2.8 percent were members of other racial or ethnic groups.
At first glance, without correcting for any differences between groups, the Medicare costs for black patients in their last six months of life were 35 percent higher than for whites, and costs were 42 percent higher for Hispanics.
Those differences in cost were cut in half after the researchers took into account a wide range of demographic, socioeconomic, geographic and health status differences among the people in each racial and ethnic group. But still, the final months of a black person's life cost the Medicare system 20 percent more than those of a white person, and the difference was 21 percent for Hispanics.
That left 'patient preferences' as the last factor to take into account. The researchers bundled together the presence of an advance directive, discussion of end-of-life treatment preferences, and the fact that a death was expected, to reflect ways in which patients could express their preferences. They did not have access to the advance directives themselves.
But even after factoring these in, the disparities persisted. The Medicare system paid 22 percent more for the care of a dying black senior, and 19 percent more for the care of a dying Hispanic senior, than they did for a white senior who matched them in more than 20 ways.
The persistence of disparities even after taking into account so many factors makes end-of-life care different from other types of care, where factors such as income, education and ZIP code explain much of the difference in outcomes and spending between racial and ethnic groups.
Finding out what factors make the most difference in the last months of life will mean more research. But in the meantime, the researchers hope their findings will help encourage more patients, providers and families of all racial and ethnic backgrounds to start the conversation before it's too late, and uphold the wishes of the dying when they're known.
In addition to Byhoff, Iwashyna and Langa, the study team included John Harris, M.D., M.Sc., a former RWJ Clinical Scholar at U-M. All of the authors are current or former members of the U-M Institute for Healthcare Policy and Innovation.
Reference: Journal of the American Geriatrics Society, DOI: 10.1111/jgs.14263