Data-Driven Visual Consent
We developed a method for consenting patients before surgery using personalized risk visualizations and accounting for patients’ preferences. The risk calculator takes the type of surgical procedure and personal information for the patient from the EHR system (or manually entered if not available) as an input. At the output, a list of the probabilities for possible complications is provided. Before they can see any of the probabilities for the possible complications, the patient is prompted to select three of those complications that concern them the most. Those will be added to the three complications with highest risk for the patient, and together represented in the final risk visualization. One of the three possible visualizations that is most appealing to the patient gets plugged in the final risk visualization, which also includes the likelihoods for discharge destinations.
In an evaluation study we found that there is not a single preferred risk visualization, but the visualization approach was perceived as desirable and especially promising for raising risk awareness and stimulating discussion with the surgeon.