ChatGPT can draft seemingly empathic answers to patient questions on medical portals. This may help clinicians respond more efficiently to their patients’ concerns.
Online messaging provides a new avenue for patients to reach doctors and for clinicians to give ongoing support to patients. At the same time, doctors are overwhelmed by messages and struggle to meet the demand. It’s hard to write accurate and empathic messages when there are so many messages that require a response.
What if you could answer a patient message with the support of an AI chatbot? Chatbots can read the patient’s message and draft a response. This may be especially helpful in filling in time-consuming details. For example, you may recommend the patient take a medication – chatbots are able to write details about how to take it and what side effects to watch out for.
A recent study evaluated the effectiveness of ChatGPT, an AI chatbot, in providing responses to questions asked by individuals to AskDocs, a public Reddit forum where doctors respond to user medical questions. The study compared answers written by doctors to automatically generated answers from ChatGPT.
Medical professionals preferred ChatGPT responses to physician responses nearly 80% of the time. Good or very good quality responses were 3.6 times higher for ChatGPT than physicians.
Surprisingly, Chatbot responses were also rated as significantly higher in terms of perceived empathy; empathetic or very empathetic responses were 9.8 times higher for ChatGPT than for physicians. While doctors are often rushed, ChatGPT has no time constraints and can generate text associated with empathy.
Healthcare professionals have the responsibility to ensure that patients receive exceptional care. While AI chatbots should not and cannot replace human doctors, they can support doctors to create high-quality and empathic communications with patients. Chatbots have now reached the point in terms of sophistication where we may be able to integrate them into patient care.
This piece expresses the views solely of the author. It does not necessarily represent the views of any organization, including Johns Hopkins Medicine.