Our Take
Real operational wins, but vendor-selected client testimonial with company-reported metrics and no independent verification.
Why it matters
Medical coding bottlenecks are choking hospital revenue cycles as patient volumes recover post-COVID, making automation a financial necessity rather than an efficiency play.
Do this week
Revenue cycle directors: audit your current coding backlog days and AR performance before vendor demos so you can establish baseline metrics for ROI measurement.
Mercyhealth reports 5.1% revenue boost from AI coding
Mercyhealth, a seven-hospital system serving 2.4 million patients across Illinois and Wisconsin, deployed autonomous AI coding from vendor Arintra across 10 medical specialties over the past 2.5 years. The health system reports a 5.1% revenue increase and 50% reduction in accounts receivable days (company-reported metrics).
Coding work queues dropped from 16 days average to 7 days, according to Kelly Pierson, director of coding and clinical documentation integrity. The system was Arintra's first Epic integration client.
Mercyhealth's staffing challenges drove the decision. "We could not find enough experienced coders," Pierson said. "We were in competition with our own organizations." The health system initially considered computer-assisted coding before selecting autonomous coding that could customize logic based on payer rules.
Coding sits at the revenue cycle chokepoint
Medical coding translates clinical documentation into billable claims, directly affecting hospital cash flow and denial rates. Staff shortages in this specialized role create immediate financial impact as claims pile up in work queues.
The operational shift matters beyond the metrics. Mercyhealth moved human coders from routine high-volume claims to complex cases and proactive audits. Pierson reports eliminated coder-to-coder variability and improved confidence during payer audits.
The health system is now evaluating additional AI vendors for other revenue cycle functions, suggesting the pilot validated the automation approach.
Baseline your current performance first
Revenue cycle leaders evaluating coding automation need clean baseline metrics before vendor conversations. Track current work queue days, AR performance by specialty, and coder productivity rates.
Mercyhealth's staged approach offers a template: initial percentage-based human review of AI-coded claims, followed by gradual shift to audit-focused human oversight. The 2.5-year timeline indicates this is an operational transformation, not a quick technology swap.
Hospital systems should factor Epic integration complexity and payer rule customization into vendor evaluation, as these determined Mercyhealth's selection criteria over pure accuracy claims.