Physician evaluation after medical errors: does having a computer decision aid help or hurt in hindsight?

Abstract

Objective: The authors examined whether physicians' use of computerized decision aids affects patient satisfaction and/or blame for medical outcomes.

Method: Experiment 1: Fifty-nine undergraduates read about a doctor who made either a correct or incorrect diagnosis and either used a decision aid or did not. All rated the quality of the doctor's decision and the likelihood of recommending the doctor. Those receiving a negative outcome also rated negligence and likelihood of suing. Experiment 2: One hundred sixty-six medical students and 154 undergraduates read negative-outcome scenarios in which a doctor either agreed with the aid, heeded the aid against his own opinion, defied the aid in favor of his own opinion, or did not use a decision aid. Subjects rated doctor fault and competence and the appropriateness of using decision aids in medicine. Medical students made judgments for themselves and for a layperson.

Results: Experiment 1: Using a decision aid caused a positive outcome to be rated less positively and a negative outcome to be rated less negatively. Experiment 2: Agreeing with or heeding the aid was associated with reduced fault, whereas defying the aid was associated with roughly the same fault as not using one at all. Medical students were less harsh than undergraduates but accurately predicted undergraduate's responses.

Conclusion: Agreeing with or heeding a decision aid, but not defying it, may reduce liability after an error. However, using an aid may reduce favorability after a positive outcome.