🤖

AI-driven Insurance Denials Overview

Nov 9, 2025

Overview

Article examines rising health insurance denials driven by automation/AI, emerging AI appeal tools, and calls for broader systemic reform.

Rising denials and automation

  • Denials have increased, partly due to AI-driven algorithms making rapid determinations.
  • Class-action suits target UnitedHealth, Humana, and Cigna for algorithm-aided denials.
  • Cigna allegedly denied 300,000 claims in two months, averaging 1.2 seconds per claim.
  • UnitedHealth’s nH Predict alleged to have 90% error rate; very few patients appeal.
  • Doctors’ offices now staff entire teams for prior authorization processing and appeals.

Appeals and patient impact

  • Less than 0.2% of ACA marketplace enrollees appeal in-network denials (KFF).
  • Under Medicare Advantage, <10% of prior auth denials were appealed in 2022 (KFF).
  • Nearly half of US adults faced unexpected bills or copays; many unaware of appeal rights.
  • Delays caused anxiety for 4 in 5; nearly half reported worsened conditions due to delays.
  • Case: ICU physician Deirdre O’Reilly faced $5,000 ER denial; four appeal attempts failed.

Insurer and expert perspectives

  • UnitedHealth spokesperson denies algorithmic use to make coverage decisions; Humana no comment.
  • BlueCross BlueShield of Vermont denies algorithm use; says clinicians apply national guidelines.
  • Michelle Mello: Better algorithms could constructively streamline coding, formatting, and approvals.
  • Andrew Witty (UnitedHealth): Most denials stem from filing errors; 85% avoidable via tech and standardization.
  • Mika Hamer: Human oversight needed; AI limits persist amid broader cost pressures.

Regulatory and policy actions

  • Senate report: UnitedHealthcare, CVS, Humana reject prior auth at high rates using automation.
  • CMS issued new rules regulating prior authorization for Medicare Advantage plans.
  • California law bans AI-only coverage decisions; requires physician oversight.
  • Several states, including Vermont, passed measures to ease prior authorization burdens.
  • Federal and state scrutiny of automated denials is intensifying.

AI tools fighting denials

  • Generative AI tools now draft appeal letters for hospitals and patients.
  • Open-source LLM projects and services help patients request records and contest denials.
  • “Battle of the bots” as providers deploy AI against insurer automation.

Costs and system-level issues

  • Provider administrative costs to appeal denials exceed $7.2bn annually.
  • One in five US GDP dollars is spent on healthcare; overhaul seen as necessary.
  • Standardized industry processes emphasized to reduce errors and speed decisions.

Key figures and entities

Entity/TopicClaim/DetailSource/Context
Cigna300,000 denials in 2 months; ~1.2 seconds per claimLawsuit allegation, ProPublica reference
nH Predict (UnitedHealth)Alleged 90% error rate; few appeals (~0.2%)Lawsuit allegation
Appeal rate (ACA)<0.2% of denied in-network claims appealedKFF survey
Appeal rate (Medicare Advantage prior auth)<10% appealed in 2022KFF survey
Administrative cost of appeals>$7.2bn annuallyCMS data analysis
Medicare Advantage market shareTop three provide ~60% of coverageUS Senate report
UnitedHealth revenue~$300bn in 2024; projected $340bn in 2025Earnings call
Avoidable denials85% avoidable via tech/standardizationAndrew Witty estimate
Healthcare spend1 in 5 US GDP dollarsMika Hamer

Decisions

  • California enacted physician oversight; prohibits AI-only coverage decisions.
  • CMS implemented new Medicare Advantage prior authorization rules.

Action Items

  • Patients can use tools to file records requests and generate appeal letters.
  • Providers and payers to pursue standardized forms/processes to reduce denials.
  • Policymakers to continue oversight on automation and prior authorization practices.