RecovryAI Completes Patient Phase of Pivotal Trial for AI-Powered Postoperative Virtual Care Assistant
Key Insights
RecovryAI (search) reached last patient out in its 15-month pivotal trial involving 525 hip and knee replacement patients across three U.S. orthopedic centers, generating over 43,200 AI-patient clinical conversations.
The study's primary endpoint measures concordance between the Virtual Care Assistant's clinical classifications and independent assessments by experienced orthopedic surgeons, with more than 129,000 individual physician assessments generated.
The company holds FDA Breakthrough Device Designation and plans to submit a De Novo classification request later this year, seeking authorization as a Class II medical device.
RecovryAI (search), a clinical AI company developing patient-facing Virtual Care Assistants (VCAs) for postoperative recovery, announced today that its prospective pivotal study has reached last patient out, marking the completion of patient participation and clinical data collection. The milestone brings the company one step closer to its planned FDA De Novo submission for what could become one of the first authorized patient-facing clinical AI devices capable of performing postoperative assessment and triage without requiring a licensed provider in every interaction.
The 15-month study was conducted under institutional review board oversight at three U.S. orthopedic centers: OrthoArizona in Scottsdale, Kansas Joint & Spine in Wichita, and Mercy Medical Center in Baltimore. A total of 525 patients completed the prospective study, generating more than 43,200 AI-patient clinical conversations for independent physician adjudication.
How the Virtual Care Assistant Functions
During each interaction, the VCA evaluates whether a patient's reported recovery is consistent with the expected postoperative course and provides appropriate guidance. When the system identifies a potential deviation from the expected trajectory, it alerts the monitoring care team with relevant clinical context and synthesis. This safety-first division of labor is intended to expand clinical capacity and allow providers to focus their time and expertise on the patients who need them most.
"Last patient out is not the finish line, but it is a significant milestone in a rigorous, multi-site pivotal trial," said Scott Walchek, co-founder and CEO of RecovryAI (search). "It reflects an extraordinary effort by our clinical partners, patients and team, and gives us the evidence base to begin the final adjudication and statistical work supporting our planned FDA De Novo submission."
Study Design and Endpoints
The study's primary endpoint measures concordance between the VCA's clinical classifications and recommendations and the independent assessments of experienced orthopedic surgeons. Each eligible conversation is presented in a blinded format and reviewed independently by three physicians, generating more than 129,000 individual physician assessments. The resulting dataset will provide a direct measure of concordance and form the statistical basis for the company's planned De Novo submission to the U.S. Food and Drug Administration later this year.
With last patient out achieved, RecovryAI (search) will proceed with final physician adjudication, database lock, and the study's prespecified statistical analysis.
"We participated in the pivotal trial because postoperative demand is outpacing available clinical capacity," said Dr. Ken Schmidt, an orthopedic surgeon at OrthoArizona and principal investigator in the RecovryAI (search) pivotal trial. "We're confident that AI can safely carry a meaningful share of that workload, but without rigorous evaluation and FDA authorization, it cannot earn our trust as part of patient care."
Retrospective and Supplemental Analyses
The prospective study is paired with a retrospective analysis involving 579 additional patients, bringing the total patient population evaluated under the protocol to 1,104. The retrospective component supports a secondary endpoint examining the VCA's potential effect on clinical workflow and downstream healthcare utilization.
The protocol also includes a supplemental stress-testing dataset of approximately 1,500 conversations designed to evaluate the VCA against uncommon and clinically complex patient presentations that are difficult to capture at meaningful scale in a prospective study. These conversations were created through a proprietary controlled-generation framework developed by RecovryAI (search) to simulate combinations of clinical conditions, comorbidities, communication styles, and levels of health literacy. Each generated conversation is anchored to a real, physician-adjudicated clinical interaction and independently validated for clinical plausibility before being used to evaluate the VCA.
Regulatory Pathway and Clinical Rationale
RecovryAI (search) received FDA Breakthrough Device Designation for the VCA earlier this year and is pursuing De Novo authorization as a Class II medical device. The study's primary endpoint, sample-size methodology, and physician-adjudication framework were informed by an earlier pilot study and refined through numerous pre-submission interactions with the FDA beginning in 2024. The pivotal study was conducted as a Non-Significant Risk study, with the investigational device operating under physician oversight throughout.
"There is a meaningful difference between AI that supports a clinician and AI that engages a patient with clinical guidance. The latter must be evaluated as a medical device, with evidence sufficient to establish that it performs safely and reliably," said Dr. Benjamin Ma, Chair of the Department of Orthopaedic Surgery at the University of California, San Francisco. "RecovryAI (search)'s prospective study, independent physician adjudication and FDA pathway reflect the diligence required for patient-facing clinical AI to earn a place in clinical practice."
RecovryAI (search)'s first VCAs are being developed for patients recovering from total hip and total knee arthroplasty, with plans to expand across major procedural domains subject to regulatory authorization. The VCA has not been authorized by the FDA and is not currently commercially available for the proposed indication.