Two years of audits across 11 hospitals showed that as adoption of Carta Healthcare's Lighthouse increased, registry‑field accuracy improved
SAN FRANCISCO, Oct. 27, 2026 /PRNewswire/ -- Hospitals using AI to support clinical data abstraction for surgery registries maintained audit accuracy above 99% across every data field, according to two years of results released today by Carta Healthcare. The study followed 11 hospitals in the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) through the 2025 and 2026 audit cycles. As hospitals adopt AI‑assisted abstraction to reduce cost and time, the central question is whether the AI output can be as accurate as a human.The data proves it does.
In 2026, hospitals using Lighthouse, Carta Healthcare's AI‑assisted abstraction platform, achieved an average audit score of 99.36%, which is slightly better than the 99.32% average at hospitals continuing with manual abstraction. Every participating site met or exceeded the NSQIP benchmark of 98%, demonstrating that efficiency gains from AI can be achieved without compromising data quality.
Manual audit review typically covers a sample of cases. With AI supporting review inside the NSQIP workstation, the Carta Healthcare team checked 100% of data fields in every cycle. That coverage showed when the same error appeared across many cases, so the team could build abstractor education around it instead of correcting one chart at a time. Abstractors reviewed and corrected their own flagged cases, and the improvements lasted into the next cycle.
Erin Tobaison, Senior Data Abstraction Manager at Carta Healthcare, presented the findings at the ACS Quality, Safety & Cancer Conference in Orlando, where this year's theme was "QI Powered by AI."
"The technology changed what we could see, and the people changed what we did about it," said Tobaison. "When an abstractor understands why a field was flagged, they fix the pattern and not just the case. The scores held above 99%, and the team got better at the same time."
The proof leaders asked for
Health system leaders have spent the year asking for evidence that AI can work with clinicians in control. In a Carta Healthcare survey of members of The College of Healthcare Information Management Executives (CHIME) released in August, 81% rated clinician involvement in the design and deployment of AI as mostly or vitally important, and respondents named hallucination as the top risk of AI running without oversight. In a national survey of healthcare leaders released in July, 92% said deep healthcare domain expertise is critical in an AI vendor, and 36% said peer case studies with measurable outcomes would help accelerate adoption.
The NSQIP audits are that kind of case study. Under the Hybrid Intelligence model, clinical abstractors reviewed and approved every case while AI served as the force multiplier that made full-field review possible. Average accuracy stayed above 99%.
"Nobody adopts AI to match what they already had," said Brent Dover, CEO of Carta Healthcare. "You adopt it to do more, and you need to know what that costs you in accuracy. At these 11 hospitals it cost nothing. Leaders have told us all year they want clinicians in control and proof that holds up under audit, and this is what that looks like. Clinicians at the helm with AI as the force multiplier kept every site above the benchmark, reviewed every field, and came away with a better way to improve."
The full findings are available here.
About Carta Healthcare:
At Carta Healthcare, we believe high-quality data is the foundation for better healthcare outcomes. Traditional clinical data abstraction is labor intensive, time consuming, and costly. Our Hybrid Intelligence approach keeps clinicians at the helm and uses AI as the force multiplier, pairing credentialed clinical abstractors with advanced technology to deliver accurate, actionable data while reducing costs and saving time. That model has earned the trust of hospitals and health systems nationwide, reflected in 100% customer retention. Carta Healthcare has been recognized with the Merit Award for Best Use of AI in Healthcare, named to the CB Insights Digital Health 50, and honored with the BIG Innovation Award for AI Excellence. Discover more at carta.healthcare
Media contact:
Supreme Communications for Carta Healthcare
[email protected]
SOURCE Carta Healthcare

Share this article