Computer-assisted physician documentation (CAPD) solutions help close gap in outpatient care, giving clinicians a more accurate and detailed picture of every patient’s chronic condition and ensuring health systems receive appropriate reimbursement. Dr. David Jordahl from MyMichigan Health shares his experience implementing an AI-powered tool to help clinicians improve clinical documentation integrity and boost RAF scores.
In the previous articles in this series, we’ve looked at the evolving outpatient reimbursement landscape and the critical importance of suspecting, capturing, and recapturing Hierarchical Condition Categories (HCCs). In this final installment, we’ll explore how one health system uses AI to improve clinical documentation integrity and boost Risk Adjusted Factor (RAF) scores.
MyMichigan Health is a non-profit health system that serves 25 counties across Michigan. It provides a full continuum of care in various settings, including urgent care centers, home health, virtual care, and medical offices covering more than 30 specialties and subspecialties.
Dr. David Jordahl is a Family Physician and Senior Council Chief at MyMichigan. He spoke to us about how computer-assisted physician documentation (CAPD) is helping the organization adapt to the world of risk-adjusted care.
“I’ve been working with the HCC program for the last couple of years, and we’ve recently implemented an AI-powered tool that’s really helped us move the dial on our efforts,” said Dr. Jordahl.
The solution provides in-workflow guidance for clinicians to identify patient risk and intervention opportunities while accurately capturing and recapturing HCCs. “It’s enabled us to improve HCC capture and RAF scores without adding manual work for our coders,” Dr. Jordahl explained. “They can rely on the AI to help clinicians make the right decisions and capture the correct codes.”
Overcoming adoption challenges
Clinicians often have specific ways of working which they’ve grown accustomed to over many years in practice, and it can sometimes be difficult to persuade them to adopt new technology. However, Dr. Jordahl stressed the importance of working through adoption challenges to achieve the desired results. He also recommended starting with smaller groups of clinicians and expanding over time.
“Initially, we focused on primary care,” said Dr. Jordahl. “We have 100 clinicians on the solution now, and our strategy is to expand by targeting specialties that are most likely to have an impact on RAF scores. We’re looking at areas like cardiology, rheumatology, and endocrinology, where there are opportunities to capture additional and more specific HCCs.”
“Before deploying the solution, we ran some sessions with clinicians, educating them on HCC coding in general and on how to use the tool,” he continued. “While we’ve had some struggles engaging clinicians in using the tool, we’ve seen incremental improvements in HCC capture, as evidenced by our RAF score.”
Improving RAF scores with AI
The AI-powered tool gives MyMichigan’s clinicians and CDI team a broader picture of each patient’s chronic conditions, and the health system’s RAF score for claimed HCCs saw an 11% increase in 2022 compared to 2021. “Our RAF score is improving despite downward pressure due to variances in population, as well as normalization and coding intensity factors,” said Dr. Jordahl.
Importantly, the tool provides AI-driven advice for suspecting conditions. “The increased RAF score demonstrates not only the success of recapturing known HCCs, but also capturing additional HCCs and improving diagnosis specificity,” Dr. Jordahl explained. “To suspect conditions manually would require a lot of heavy lifting by a lot of people. It’s great to have the AI to tell us the things we don’t know, or that might have been missed.”
While My Michigan’s journey to clinical documentation excellence is ongoing, the health system is making excellent progress, giving Dr. Jordahl cause to be optimistic about the future. “Admittedly, we’ve had some struggles along the way,” he said. “However, I’m excited to see what continued developments with this tool can offer us for improving our HCC coding.”