New data found that African-Americans hospitalized after stroke had lower mortality rates than did whites, but the study authors also found that the African-American patients received life-sustaining interventions more often and had longer lengths of stay, higher hospital spending, and a lower hospice admission rate.
DR. ROBERT G. HOLLOWAY: “Most patients die after a large stroke and a deliberate decision to withhold or withdraw life-sustaining therapies — most often mechanical ventilation or artificial nutrition, or both. This, in fact, may be excellent quality of care if the choice is truly informed.”
A two-year observational study on hospital mortality associated with stroke raises difficult questions about the quality and intensity of end-of-life care. In a Feb. 1 paper in the Annals of Internal Medicine, investigators reported that inpatient mortality after acute ischemic stroke was lower in African-Americans than whites. The mortality difference was observed at discharge and was persistent through one year of follow-up.
The study included 5,319 African-American and 18,340 white patients hospitalized with acute ischemic stroke between January 2005 and December 2006. Among findings, overall in-hospital mortality was 5 percent for blacks compared with 7.4 percent for whites. Unadjusted mortality at 30 days from all causes was 6.1 percent for blacks and 11.4 percent for whites, and at one year, 16.5 percent for blacks and 24.4 percent for whites. After adjusting for patient and hospital characteristics, the difference in mortality was significant at 30 days (OR 0.69, 95% CI 0.57-0.84) and attenuated but remained significant at one year (OR 0.86, 95% CI 0.77-0.96).
Why would there be a disparity in outcomes? Study co-author Robert G. Holloway, MD, professor of neurology and community and preventive medicine at the University of Rochester School of Medicine and Dentistry, said some analysts might attribute the lower mortality of African-American stroke patients to their tendency to have more small vessel disease, which is less likely to produce the catastrophic brain damage that often results from blockage of a larger artery to the brain.
“That could contribute to some differences in mortality,” said Dr. Holloway, who is also an associate editor of Neurology Today. “We can't exclude that interpretation, but our study suggests that the differential use of life-sustaining interventions may also be at play.”
The investigators found that African-American patients with stroke received life-sustaining interventions more often and had longer lengths of stay, higher hospital spending, and a lower hospice admission rate.
“Black patients as a group tend to be more aggressive in their use of life-sustaining interventions, and less likely to enroll in hospice,” Dr. Holloway explained. “That has been shown in many studies, and I think that may be an explanation for their lower mortality.”
At a time when merely encouraging doctors to discuss end-of-life directives with patients can conjure the specter of “death panels,” the authors proceed with caution when discussing the implications of their findings, Dr. Holloway said.
“The morally challenging and ethically complex discussions we have with patients and their families about their end-of-life preferences likely have a much greater impact on short-term mortality than we think — possibly greater than what we traditionally view as having an impact, like dysphagia screening, deep vein thrombosis prophylaxis, or antiplatelet use,” he continued. “Most patients die after a large stroke and a deliberate decision to withhold or withdraw life-sustaining therapies — most often mechanical ventilation or artificial nutrition, or both.
“This, in fact, may be excellent quality of care if the choice is truly informed,” Dr. Holloway continued. “We have to confront the possibility that well-informed patients who decide to forego life-sustaining interventions may have higher short-term mortality, but an excellent quality of care — possibly better than uninformed patients who decide to have life-sustaining interventions and lower short-term mortality,” he said. “Many patients don't want to be kept alive on machines or by artificial means if [they are] severely disabled and dependent on others. They want to die a comfortable death.”
Indeed, palliative care may offer a bridge between aggressive life-sustaining treatment for the mortally ill and hospice, which focuses primarily on comfort, said Timothy E. Quill, MD, professor of medicine, psychiatry and medical humanities at the Center for Ethics, Humanities and Palliative Care at the University of Rochester School of Medicine, who was not involved with the study.
“Certain treatments prolong life or prolong the dying process, depending on how you view the status of being extremely disabled and dependent on medical interventions for much — if not all — of your continued existence,” he said. “That is the dilemma. Palliative care will support people who want aggressive treatment, but at the same time make sure they get good pain and symptom control. It makes sure there's a rich discussion about the possibilities. I think that's an attempt to get at this conundrum.”
Independent experts in stroke and health care outcomes offered varied interpretations and analysis of the study data. Lee H. Schwamm, MD, who led a 2010 study in Circulation on race/ethnicity and health care outcomes in ischemic stroke, said the study may have been limited by an inability to measure the likelihood of dying based on the severity of the stroke.
“That's an important limitation,” said Dr. Schwamm, vice chairman and professor of neurology at Harvard Medical School and director of TeleStroke and Acute Stroke Services at Massachusetts General Hospital. “In our work, that was the single most important predictor of mortality — the initial stroke severity. Certain stroke types have a much lower mortality. We discovered that blacks and Hispanics were having strokes at a younger age, presumably from poorly controlled risk factors. They have more minor strokes, so they survive hospitalization more than their white counterparts. If we followed them over time, however, we'd probably see that black patients are having strokes more frequently and at younger ages, so their quality of care is actually lower, but paradoxically, because you're looking at only one hospitalization, they appear to be doing better.”
Wuwei (Wayne) Feng, MD, lead author of a 2009 Stroke paper that found greater health disparities among younger stroke patients in the South Carolina “stroke belt,” believes the fact that African-Americans tend to have strokes at a younger age means they will be less likely to have a living will or other end-of-life directives.
“Based on my experience, the older you get, the more likely you are to have a living will, or a do-not-resuscitate directive that says I don't want life-prolonging treatment if I cannot live a good-quality life after a stroke,” said Dr. Feng, a vascular neurology fellow at Beth Israel Deaconess Medical Center in Boston. “Younger patients are less likely to have this, so they are more likely to receive aggressive treatment.”
Age may also have influenced the higher spending on the African-American stroke patients, said Philip B. Gorelick, MD, John S. Garvin Professor and head of the Department of Neurology and Rehabilitation, and director of Stroke Research at the University of Illinois College of Medicine at Chicago.
“African-Americans have strokes at a younger age than the white population,” he said. “The ratio may be as high as 2:1 or more. However, in later years the ratio flips and white populations catch up and even exceed African-Americans. In this study, the age discrepancy is very striking with whites being approximately nine years older than blacks. Since the white patients in the study were older, they may have been more likely to forego aggressive management such as intubation and respirator management, and to have a ‘do-not-resuscitate’ order. That certainly could reduce the costs associated with stroke care.”
But Dr. Holloway noted, in response, that the investigators adjusted for age and stratified by different age groups. “The mortality difference was seen for all ages,” he said.
While interpretations of the lower mortality and higher spending on African-Americans may vary, another group of investigators reported similar findings at the February International Stroke Conference in Los Angeles.
Erin M. Grise, MD, assistant professor of emergency medicine and neurosurgery, and an emergency room physician and neurointensivist at the University of Cincinnati Neuroscience Institute, described how she and her colleagues used the Medicare Provider Analysis and Review database to find racial differences among stroke patients who received a tracheostomy and percutaneous endoscopic gastrostomy (PEG), which they used as a marker for aggressive care.
“What we found was that blacks were more likely to proceed with trach and PEG placement, and to have lower in-hospital mortality, both of which are probably markers for a racial disparity between those who proceed with aggressive care versus those who don't,” she said. “I would say that both my study and the study in Annals indicate that racial differences in the approach to end-of-life care probably have a significant effect on overall mortality. Our studies are limited by the fact that we weren't able to take stroke severity into account, but since evidence states that blacks have less severe strokes, not accounting for stroke severity likely underestimates the true racial disparity in the decision to proceed with aggressive care after stroke.”
These data also suggest short-term mortality might not be a valid measure of quality, according to Dr. Holloway. “What I fear are the possible unintended consequences of using a 30-day mortality measure for stroke,” he said. “Will I feel subtle incentives or pressures to direct patients away from their own preferences so my hospital, my stroke program, or my own mortality statistics will look better? If we really want to challenge ourselves we could develop measures of a good-quality death, which would involve informed patient choice and excellent symptom control. Then death would not necessarily be a marker of failure. Patients and families seem to know this already.” •