Piedmont Health Leverages AI to Improve Efficiency, Quality in the Front End

Allyson Keller, Vice President of the Patient Connection Center at Piedmont Health and a participant in the HealthLeaders Revenue Cycle Technology Mastermind program, shares her thoughts on the potential for AI to improve.

Allyson Keller, Vice President of Patient Connection Center at Piedmont Health, operates at the increasingly important intersection of patient experience and financial access. Her front-end strategy for the revenue cycle is firmly rooted in automation, driven by the dual objective of enhancing efficiency and improving the patient-provider experience.

Keller was a participant in this year’s HealthLeaders Revenue Cycle Technology Mastermind program, which brought together RCM executives from roughly a dozen health systems for a series of virtual and in-person round-tables to discuss key benefits and challenges.

Automation will "help us be more efficient, but also be able to really drive and improve patient experience and revenue,” Keller says.

Some current projects with significant implications for the front-end at Piedmont Health include:

• Optimizing Orders with Interoperability: Patient access and informatics teams at Piedmont are currently collaborating on data transmission for orders to reduce fax volume, which creates a more seamless experience for both patients and ordering physicians .
• Scaling Automated Assistants: The system continues to enhance and scale its automated voice assistant to manage call deflection and containment.
• Improving speech recognition: Piedmont is investigating ways to gain deeper insights into why patients either schedule or cancel appointments, with the goal of improving the conversion rate for outpatient imaging orders to completed appointments.

A Wishlist for the Front End

While actively integrating current technologies, Keller points to two areas where AI could significantly improve front-end operations, particularly in reducing upstream denials.

One is the automated extraction of clinical data. An effective solution that can automatically pull the necessary clinical context from patient charts to complete authorization submissions would enhance accuracy and reduce manual effort.

“Otherwise, we’re relying on nonclinical people to do that,” Keller says.

The second area involves medical necessity and payer policy adherence. Payer policies change with such frequency that manual review is challenging and resource-intensive. An AI tool that verifies services meet up-to-date payer policies could help to prevent costly medical necessity denials down the line, according to Keller.

Efficiency Plus Quality Equals Success

Piedmont broadly considers impacts to efficiency and quality when it considers technology investments.

Success is not just about cost reduction, according to Keller. It is about enhancing patient  access and experience. Use cases must be quality-driven, meaning the technology must perform as well as or better than a human. Key quality metrics include decreasing denials and improving the conversion rate of orders to completed appointments, which ensures patients receive care without delay.

Despite the immense potential for AI, its implementation currently requires significant human effort, especially on the front end.

“It’s going to take a lot of human interaction,” Keller says. “To work with these tools, to get them to do what we need them to… Right now, we’re still in a place that it takes a lot of effort to implement these tools.”