Can surrogates really know patients' wishes?

Just more than half of surrogates get it right

Making end-of-life decisions for incapacitated patients most often falls to surrogates chosen by the patients, or to next of kin. But a recent review of the literature indicates that surrogates are only slightly better than physicians at making decisions that the patient would make if he or she were able.

Terri Schiavo's dilemma—being the center of a struggle over end-of-life decisions between her husband, who Florida state courts recognized as her legal next of kin for decision-making purposes, and her parents—has prompted many Americans to sign living wills and talk to their next of kin about what end-of-life care they want or do not want.

But according to David Wendler, PhD, of the National Institutes of Health's Department of Clinical Bioethics, a review of studies that provide data on how accurately surrogates predict patients' treatment preferences indicates that while those discussions are certainly better than doing nothing, surrogates in the studies predicted patients' treatment preferences with 68% accuracy.

Surprisingly, they did even worse after discussing treatment preferences with the patient.

Surrogates get it right sometimes

The analysis, conducted by Wendler and colleagues at the National Institutes of Health and Johns Hopkins University, examined 151 role-playing scenarios that were presented to 2,595 patient/surrogate pairs.1 The "patients" in some cases were healthy subjects asked to assume they were incapacitated and others who were actual patients not at the end of life.

In the studies, the next of kin or surrogate decision makers were asked to decide whether the patients would want certain medical interventions, often ones that physicians deemed necessary to keep the patients alive. The patients were asked, independently, what their preferences would be.

A sample hypothetical scenario was: "You recently suffered a major stroke, leaving you in a coma and unable to breathe without a machine. After a few months, the doctor determines that it is unlikely that you will come out of the coma. If your doctor had asked whether to try to revive you if your heart stopped beating in this situation, what would you have told the doctor to do?" Patients and surrogates were free to choose to accept or reject the proposed treatment.

Leaving the decision up to doctors resulted in a rate of 63% accuracy at predicting what the patient would want. Surrogates or next of kin fared only slightly better at 68%.

Of the 16 studies reviewed, 12 assessed the errors surrogates made when they misjudged patients' preferences. Three studies found that they erred by choosing interventions that the patients said they would not want. One study found that surrogates erred by withholding interventions the patients would want, and eight studies found no consistent trend in surrogates' miscalculations.

The studies showed only one percentage point difference in the accuracy of patient-chosen surrogates (69%) and legally assigned surrogates (68%).

"In general, we thought the way to explain people not being as accurate as you hope they are was that they are unwilling or reluctant [to make the decisions]; but we found you really can't say one way or the other," says Wendler.

Talking about preferences doesn't help

It would be natural to assume that if patients and their surrogates talk about preferences beforehand, the surrogates' decision making would be more accurate.

Natural, maybe — but wrong, according to the two studies that examined whether discussion improved accuracy.

One study found no change in accuracy, while the other found a slight but statistically significant worsening of surrogates' accuracy after they talked with patients about their wishes.

"It was a little surprising that talking didn't change accuracy, and we don't really know why that is," says Wendler. "Our hypothesis is that the kinds of discussions you can have tend to be very general and not sufficiently rich to capture the complexity of real cases."

Wendler says patients should be encouraged to think about what is important to them and what they envision happening in certain circumstances.

"A lot of time, it's not clear," he says. "If you like to talk, read, and walk, then would it be acceptable to you to be on a ventilator? Also, medical probabilities are involved [in real-life situations], and doctors can't tell you what those are for sure."

According to Cynda Rushton, DNSc, RN, FAAN, clinical nurse specialist in ethics at Johns Hopkins Children's Center in Baltimore, a positive benefit to emerge from the Schiavo case is that the number of people who are considering the importance of designating health care agents and preparing living wills has jumped dramatically.

"We are encouraging living wills and for patients to converse with their surrogates early and often," she says.

Rushton says a study conducted at Johns Hopkins, which examined the process of selecting and communicating with surrogates, found that while many had definite ideas they wanted surrogates to adhere closely to, others said they wanted family members to do what they thought was best under the circumstances.

Maryland is in the process of revamping its advance directives, to encourage people to indicate how they want their living wills interpreted — literally or by surrogates applying their best judgment at the time.

Wendler says surrogates were most accurate in making decisions in situations where the patient was in good health before becoming ill and the intervention was relatively sure to return him or her to good health.

"They're good at predicting in those situations, because that's easy," he says. "If you're a healthy 50-year-old and you get pneumonia and you need to be on a respirator but will recover, everyone will say 'yes' to that.

"The ones that give you trouble are the [exercises] where you start radically changing their cognitive state. That's when the preferences aren't as clear and surrogates misestimate it. But also, people have doubts about what their own preferences would be in those circumstances, for example, if they have Alzheimer's.

"For me this suggests that we need more data on what people really care about," says Wendler. "Argument points to using next of kin [for end-of-life decisions], but if you find out that the next of kin is not that accurate, then you have to ask yourself what the impact is for using family members as surrogates, and whether the impact is positive or not. Legislating [next of kin as surrogates] is a reasonable way to go, but you never know. These are hard decisions to make, and it may be a burden."


1.Shalowitz DI, et al. The Accuracy of Surrogate Decision Makers: A Systematic Review. Arch Intern Med. 2006;166:493-497.