Case Study
Froedtert & MCW Clears a Registry Backlog and Returns Senior Staff to Quality Work with Layer Health
Froedtert & MCW set out to answer a question every quality leader is asking: can AI handle the accuracy bar for clinical registry submission while actually reducing the burden on an already stretched abstraction team? The answer was yes – across every measure they tracked.
57%+
Reduction in heart failure abstraction time
98%+
AI accuracy across the heart failure registry
3x
ROI delivered
About Froedtert & MCW
Froedtert & the Medical College of Wisconsin is a leading academic health system with an active registry program spanning cardiovascular, surgical and neurological specialties. The American Heart Association recognizes Froedtert Hospital and Froedtert Menomonee Falls Hospital with the Get With the Guidelines® Heart Failure Gold Plus achievement award, and Froedtert West Bend Hospital with the Silver Plus award – recognition that depends on the accuracy and completeness of registry data. Froedtert's heart failure registry team faced a backlog that put submission deadlines at risk. To address it, leadership pulled in a senior data manager – a highly skilled, highly compensated member of the quality team whose role is to build analyses, inform care pathways, and support clinical leadership – to work through the cases. Froedtert partnered with Layer Health to evaluate whether AI could meaningfully reduce abstraction time, clear the backlog, and return senior staff to the quality improvement work they were hired to do.
The challenge
Froedtert's registry team faced compounding pressure: growing case volumes, a growing backlog, and a staffing model that couldn't absorb both. Like many health systems, they had no good way to scale.
Froedtert conducted a structured evaluation of traditional manual abstraction versus AI-assisted abstraction, measuring time and accuracy at both the aggregate and individual abstractor level.
Four questions shaped Froedtert’s evaluation:
Can AI meaningfully reduce abstraction time for heart failure registries?
Can AI return high-cost staff to appropriate quality improvement work?
Does AI-supported abstraction maintain clinical-grade accuracy?
Will the abstraction team actually embrace AI-supported workflows?
The solution
Froedtert & MCW partnered with Layer Health to evaluate whether AI could meaningfully reduce abstraction time, clear the backlog, and return senior staff to the quality improvement work they were hired to do.
Layer Health provides pre-filled registry responses supported by evidence from the patient chart, enabling abstraction teams to review and validate recommendations rather than manually searching for every data element.
Froedtert & MCW’s path to transforming registry operations with AI
"The Layer Health team took time to understand our workflows to ensure their solution was impactful. I couldn't imagine going back to our old way of doing things."
Nurse Abstractor, Froedtert & MCW
The results
Time Savings
57%+ Reduction in Heart Failure Abstraction Time: Layer Health provided an evidence-supported first pass – without compromising accuracy.
Instead of searching the chart for every data element, the team reviewed pre-filled recommendations with supporting documentation.
The impact was most visible for Froedtert's senior data manager, whose abstraction time dropped from 68 to 28 minutes per case – a 59% reduction for a staff member whose expertise commands significantly higher compensation than a standard abstractor.
Accuracy
Met or Exceeded Human Accuracy Across the Heart Failure Registry: Layer Health met or exceeded human accuracy at both the field level and in aggregate.
Layer Health met or exceeded human performance across fields requiring both structured data (labs, ADT, demographics) and unstructured data (admit and discharge medications, imaging, cardiac and diagnostic tests).
Froedtert's abstractors could see exactly why Layer Health made each recommendation in the supporting documentation tied to every pre-filled answer. That visibility earned their trust.
And, for a system recognized by the AHA for heart failure registry quality, maintaining that standard while reducing abstraction burden was the baseline requirement. Layer Health met it across the vast majority of measured fields.
Operational Excellence
Backlog Cleared and Senior Staff Returned to their Real Work: A high-cost resource went back to the work she was hired to do.
Froedtert's senior data manager was pulled from quality improvement work to help clear a mounting registry backlog. Layer Health accelerated her throughput enough to clear it – and return her to building the dashboards and pathway analyses that registry data is meant to enable.
The result was both faster abstraction and workforce flexibility: the ability to surge capacity during a crunch and pull it back once the backlog was gone.
Layer Health Surfaced a Workflow Inefficiency: Routine validation revealed a hidden workflow gap.
During implementation, Layer Health's validation process revealed that one abstractor was completing registry fields no longer required for her submission type – a holdover from an outdated specification, adding 10-20 minutes to every case.
Layer Health gave the team the visibility to correct it, compounding time savings across every future case. The same rigor that validates AI accuracy also reveals how human workflows are actually running.
Productivity
Senior Staff Reached Productive Speed in Weeks: Fast enough to clear the backlog before returning to her primary role.
Layer Health shortened the ramp-up curve for Froedtert's redeployed senior data manager. By surfacing relevant evidence and applying registry logic consistently, the platform enabled her to reach productive abstraction speed within weeks of deployment – fast enough to clear the backlog before returning to her primary role.
Looking ahead
The registry results, and the rigor behind them, gave Froedtert confidence to expand the relationship with Layer Health.
When the health system opened a new ambulatory surgical center and needed to screen patients against 74 custom eligibility criteria, they turned to Layer Health.
The same platform that proved itself in heart failure abstraction was adapted for site of care optimization — a real-time dashboard that flags patients who are not appropriate for the ASC, surfacing risk factors typically scattered across the EHR.
The results:
100% of red-flag patients correctly identified
99.6% accuracy across red and yellow flags
As one orthopedic surgeon put it: "This tool is already prompting us to make real clinical decisions."
Froedtert & MCW continue to expand Layer Health across additional cardiovascular and surgical registries, using the heart failure foundation as a platform for broader quality program improvement.
