A paper in the recent edition of PLoS Medicine proposes a radical idea: a computer might be better at making end-of-life decisions than a patient’s family. A trio of researchers led by NIH’s David Wendler created a “population-based treatment indicator” from already-published surveys of patient preferences in end-of-life cases. Wendler and company then compared the predictive accuracy of the indicator against that of surrogate decision makers in hypothetical situations. The result: the decisions made by both surrogates and the algorithm matched that of patients about 78 percent of the time. The authors hypothesize that as more specific data on patient preferences are added to the indicator formula, it could become even more accurate at predicting a patient’s wishes than his or her family.
But if such a system were to show up in clinical settings — and that, of course, is an enormous if — would it really be used to shed light on patient preferences? Or would its real use be as a way of making family members feel better about their decisions? The authors suggest such a possible use: “This approach might help to relieve some of the burdens associated with making decisions for incapacitated patients, while allowing family and loved ones to retain final decision-making authority.” The converse could also be true. What happens if a family’s decision conflicts with the computer? Should the clinician tell them? Has the family made the “wrong” choice?
Scientific American has more on the “population-based treatment indicator.”
-Greg Dahlmann
Labels: end of life, health services research