Survey: Severe Stroke Prognoses Differ Depending on the Doctor

A physician has to make many quick decisions when confronted with a stroke patient.

Should treatment be intensive or comfort-based, or somewhere in between?

A physician will recommend a more intensive approach for a patient who may be able to rehabilitate, but it's all about comfort if the chances for recovery are low.

The stakes for these decisions are even higher in cases of brain hemorrhage, a severe type of stroke after which about one-third of patients die within a month. A group of neurologists is encouraging colleagues to slow down, though, before heading out to the waiting room to discuss treatment recommendations with family members.

"There's a lot of variability across centers in terms of how these patients are treated," says U-M neurologist Darin Zahuranec, M.D., principal investigator of a new physician survey published in Neurology. "We wanted to see what role, if any, the physicians may play in that variability."

Huge variation in survey results

Researchers surveyed 742 physicians across the U.S. on two sample cases, describing two patients with a brain hemorrhage. The investigators selected characteristics they thought would elicit varying responses from doctors: differing patient age and severity.

The responding physicians had to predict the 30-day mortality rate and recommend treatment intensity for each case.

"The range of predicted mortality was from 0 percent to 100 percent in most of the cases," Zahuranec, a member of U-M's Institute for Healthcare Policy and Innovation, says. "I was surprised to see that level of variability among physicians."

Treatment recommendations also ranged widely among physicians, with some recommending comfort measures only and others suggesting full intensive treatment for the exact same patient. 

Quality metrics at individual centers often require physicians to document a prognostic score in the chart while treating stroke patients. These prognostic scores are meant to provide a standardized assessment of severity. But the effect of showing doctors these scores has not previously been tested.

To see whether having a model would change the physician recommendations, the research team included prognostic scores for some patients but not for others.

"We wanted to see if you approach treatment differently if you're being told the patient's chance of recovery," Zahuranec says.

In the most severe case, when the score suggested 0 percent chance of recovery, physicians were more likely to recommend comfort measures only. In a mild case, physicians who saw a better chance of recovery were more likely to recommend full intensive treatment with the goal of rehabilitation.

The survey also revealed neurosurgeons tended to be more optimistic about mortality predictions than neurologists, and physicians who saw the most brain hemorrhage cases were slightly more pessimistic than those who saw fewer cases.

Researchers tried to control for physician personality, such as religion or optimism, but none of those measures predicted the physicians' treatment recommendations.

Effect on patient families

Brain hemorrhage patients are not usually able to participate in these initial discussions about their case, so physicians often end up sharing their prognoses with family members, who are eager for the doctor's insight before making a care decision.

Knowing physicians approach these decisions differently, the researchers recommend that clinicians keep the variability in mind before sharing their prognoses with stroke patients and families, who are already dealing with a high-stress situation.

"Do we really know what we think we know when making these predictions?" Zahuranec asks. "My hope for physicians is that we really understand the impact of our prognostic statements."

He recommends physicians first discuss with the family what a good recovery would mean, and what would be important for that individual patient, and then tailor the predictions and recommendations to those wishes.

"These situations are always very difficult for the family and for the physicians," Zahuranec says. "One thing that can make things easier is when the family has a clear understanding of what the patient would want."

This research was funded by the National Institute on Aging and grant support to the Michigan Institute for Clinical & Health Research.

Disclosures: Zahuranec receives research support from the National Institutes of Health (KAG038731). Co-author Angela Fagerlin receives research support from the NIH, the Patient-Centered Outcomes Research Institute, the European Union and the VA Healthcare system. Co-author Brisa Sánchez receives research support from the NIH. Co-authors Meghan Roney, Bradford Thompson and Andrea Fuhrel-Forbis report no relevant disclosures. Senior author Lewis Morgenstern receives research support from the NIH.