High utilization and AI-driven coding accuracy weigh on Highmark Health's earnings
Highmark Health's earnings took a hit in the first nine months of the year. Two forces led the charge: higher-than-expected medical utilization and tighter, AI-powered coding accuracy that reduced revenue capture.
If you work in health insurance or provider finance, this combo is familiar. More people are using care, and the codes attached to that care are being scrutinized with far more precision.
What's behind the utilization surge
Deferred care is still working its way through the system, with more members showing up sicker and needing higher-intensity services. Specialty drugs, behavioral health, and outpatient procedures are adding steady pressure to medical loss ratios.
Even small shifts in site-of-care or referral patterns can compound costs across a large membership base. That shows up quickly in quarterly results.
How AI-powered coding accuracy changes the revenue picture
AI tools are improving coding accuracy by catching unsupported diagnoses and trimming ambiguous or low-evidence conditions. That lowers risk scores and reduces risk-adjusted revenue, especially where documentation doesn't fully support chronic conditions.
For plans active in risk-adjusted programs-Medicare Advantage and ACA marketplaces-cleaner coding can mean immediate revenue headwinds if documentation and clinical validation lag. For context on risk adjustment mechanics, see CMS guidance on risk adjustment programs here.
Implications for payers and provider partners
Expect tighter margins, more scrutiny on documentation quality, and increased tension between revenue integrity and compliance. Plans will push for stronger provider education, better evidence capture at the point of care, and more consistent problem-list hygiene.
Providers will need to align clinical notes, problem lists, and encounter documentation with medical necessity-without slowing throughput or adding friction to workflows.
What to do next
- Recalibrate pricing, reserves, and benefit designs with current utilization trends-not last year's assumptions.
- Strengthen utilization management with member experience in mind: site-of-care optimization, specialty drug strategies, and high-cost case management.
- Tighten documentation quality: focus on specificity, chronic condition persistence, and clinical validation linked to medical necessity.
- Audit your AI coding stack: require transparent rules, physician-in-the-loop review, version control, and clear audit trails.
- Expand value-based contracts that share risk and reward documentation quality, care coordination, and outcomes.
- Align provider incentives and education with the documentation you expect to see-short modules, real examples, fast feedback.
Practical guardrails for AI-assisted coding
Set policies that separate suggestion from decision: AI proposes, clinicians and coders confirm. Track model drift, run dual coding on samples, and compare outcomes against human-only baselines.
Most importantly, connect coding suggestions to clear evidence in the chart. If a diagnosis can't be defended in an audit, it shouldn't drive revenue.
Bottom line
Higher utilization is lifting costs while better coding accuracy is flattening revenue. The near-term fix isn't a single tool-it's tighter operations, stronger documentation, and smarter contracts.
If your teams need targeted AI upskilling for documentation, coding oversight, or automation fundamentals, explore focused programs at Complete AI Training and the latest AI courses here.
Your membership also unlocks:
