Tue, 13 Mar 2007 07:51 CDT
A simple formula can predict how people would want to be treated in dire medical situations as accurately as their loved ones can, say researchers.
The finding suggests that computers may one day help doctors and those acting as surrogate decision-makers to better estimate the wishes of people in a coma.
By signing what is known as an "advance directive", people can specify what types of medical care they would want if they lost the ability to make decisions. Many people, however, do not complete such a directive in advance of these critical situations and their relatives or others must then decide on their behalf.
But how well can surrogates accurately predict the wishes of patients? Researchers have previously addressed this question by asking people how they would want to be treated in various hypothetical medical scenarios and, in a separate room, asking surrogates to guess what those responses had been. A review of 16 studies found that surrogates got it right only 68% of the time.
Reason and remember
Bioethicist David Wendler of the National Institutes of Health in Bethesda, Maryland, US and colleagues wondered whether a formula could be used to better predict a patient's wishes. They examined information collected by pollsters and scientists about the attitudes towards medical care held by the general US population.
The data suggested that most people want life-saving treatment if there is at least a 1% chance that following the intervention they would have the ability to reason, remember and communicate. If there is less than a 1% chance, people generally say they would choose not to have the treatment.
"The difference between zero and 1% is all the difference in the world for someone," says Wendler.
His team then looked at subset of the 16 studies in which the medical scenarios were judged to be easier for a member of the public to understand. In these casea, they found that surrogates predicted the patient's wishes more accurately, 78% of the time. But surprisingly, using the formula that people only want interventions if there is a 1% chance of a good outcome had the same accuracy.
Wendler says he was surprised at the formula's accuracy. "I think it's fascinating. At first when you hear it you think 'That just can't be right,'" he says.
He imagines a situation in which a surrogate is told there is only a 5% chance that an incapacitated loved one will survive a life-saving surgery following an auto accident. He says that the relative might predict that the patient would not want the intervention while the formula would predict that they did.
Wendler now wants to collect medical care preferences from people of various ethnic, religious and gender groups, which will help his team refine the formula. He believes that a computer program might one day predict patient's wishes to an accuracy of 90%.
And the tool could take some of the pressure off of relatives who sometimes have to decide whether or not to switch off a patient's life support machine.
Question of ethics
However, critics caution that computer algorithms should never supplant human surrogates. "I believe it would be extremely irresponsible to allow machines to make decisions involving life and death," says Bobby Schindler, brother of Terri Schiavo. Schiavo was in a persistent vegetative state for 15 years until she died in 2005 after doctors removed her feeding tube. Her case sparked huge debate in the US.
"If a person becomes incapacitated, is not dying, and can assimilate food and water via a feeding tube, then I believe that we are morally obligated to care for the person and provide them this basic care - regardless of a computer attempting to 'predict' what that person's wishes might be," Schindler adds.
"Essentially, you would be allowing a machine to determine what is ethical, what is right and wrong, which no machine is able to do."