Artificial Intelligence Chatbots in Surgical Care: A Systematic Review of Clinical Applications

In the context of surgical care where accurate and timely information is essential, artificial intelligence (AI)-driven chatbots offer innovative opportunities for improving patient education and perioperative outcomes.

A systematic review per PRISMA guidelines was conducted to evaluate the application of chatbots within the surgical pathway and assess outcomes relating to patient experience, cost, safety, and clinical recovery. Studies were retrieved from MEDLINE, EMBASE, CENTRAL, and Google Scholar databases (November 2024) and were included if they deployed chatbots in the perioperative timeframe for adult surgical patients.

The review encompasses twelve studies totaling 6,619 patients, featuring rule-based, rule-and-frame-based, hybrid, and generative AI chatbots. Chatbots were used for delivering automated information (66%), answering patient queries (66%), symptom monitoring (16%), facilitating clinic communication (16%), and soliciting patient feedback (8%). Chatbots achieved 60-82% satisfaction on Likert scales, engagement rates of 35-83%, and accuracy rates from 79-99% depending on function. Additionally, they reduced healthcare personnel workload, saving 9-39 hours per 100 patients postoperatively. While no shared clinical outcomes were assessed across multiple studies, individual studies found chatbot use was associated with reductions in opioid use, pain, preoperative anxiety and readmissions. Risk of bias was assessed via the ROBINS-I tool with most studies classified as low risk (67%) and 25% as serious risk due to confounding variables, missing data, and measurement biases.

Chatbots are emerging as a useful tool in surgical care, improving patient outcomes, satisfaction, and engagement. Future research can further explore chatbot models, delivery methods, and surgery-specific applications on clinical outcomes.