Automated AI Voice Calls for Postoperative Follow-up After Arthroscopic Shoulder Instability Surgery

Automated AI Voice Calls for Postoperative Orthopedic Follow-up: A Randomized Trial of SMS Pre-notification and Blinded Validation of PROM Capture.

This prospective, single-center randomized clinical trial evaluates the technical efficacy, accuracy, and operational performance of an automated conversational artificial intelligence (AI) voice agent for postoperative follow-up after arthroscopic shoulder instability surgery. The study examines whether SMS pre-notification improves survey completion during the first automated outbound call and assesses the accuracy of AI-captured patient-reported outcomes compared with blinded human review.

Participants who underwent arthroscopic shoulder instability surgery at least one year prior were contacted using an automated AI voice agent to collect standardized postoperative outcomes. Patients were randomized 1:1 to receive an SMS pre-notification prior to the first automated call or no pre-notification. The AI agent conducted structured interviews and classified call outcomes using a prespecified taxonomy. A hybrid protocol combining automated outbound calls, inbound callbacks, and targeted human follow-up was implemented to maximize data capture while minimizing resource use. All completed calls were independently audited by a blinded human reviewer to assess agreement with AI-extracted outcomes.

Inclusion:

• Age ≥18 years
• History of arthroscopic shoulder instability surgery
• Minimum postoperative follow-up of 12 months
• Valid telephone contact information

Exclusion:

• Previous ipsilateral surgery
• Non-consent