Summary
KFF’s annual analysis of claims denial and appeals data reported by HealthCare.gov insurers for plan year 2024. HealthCare.gov insurers denied approximately 19% of in-network claims and 37% of out-of-network claims in 2024 — a combined average of 20%, or roughly 8.8 million denials across reporting insurers. Fewer than 1% of denied in-network claims were appealed, and of those, insurers upheld 66% of original denials — meaning approximately 34% of appeals resulted in the original denial being overturned. This data is the clearest quantitative documentation in the wiki of institutional exhaustion-as-strategy operating at scale on the American public: the denial machine runs on non-participation, not on being right.
Key Points
- Denial rates: 19% in-network, 37% out-of-network, 20% combined average for 2024. Rates stable vs. 2023.
- Variation by insurer: 3%–36% denial rates across reporting insurers. Among parent companies with 5M+ claims: 8% (Elevance Health) to 25% (Oscar Health).
- Geographic variation: Hawaii 27% (highest state); South Dakota 7% (lowest).
- Denial reasons (in-network, 2024): “Other” (unspecified) 36%; administrative 25%; excluded services 13%; lack of prior authorization/referral 9%; medical necessity 5%. The plurality — 36% — has no stated reason.
- Appeals data (the critical finding): Fewer than 1% of denied in-network claims were appealed. Of appealed claims, insurers upheld 66% of original denials — meaning ~34% of appeals were overturned. Only 5,881 external (independent-review) appeals were filed in 2024. Only 4% of upheld internal appeals proceeded to external review.
- 2023 comparison: Same <1% appeal rate; 2023 overturn rate was 44% (dropped to 34% in 2024).
Newsletter Angles
- The mathematical proof of institutional gaslighting: ~44M claims filed → ~8.8M denied → ~88,000 appealed (<1%) → ~30,000 overturned. ~8.77M denials are never appealed despite a ~34% success rate when people do appeal. The system works because the overwhelming majority of affected people do not participate. This is the exhaustion engine at consumer scale.
- Cigna PxDx context (from related reporting): Cigna’s algorithmic claim-denial system denied 300,000 claims in two months at an average of 1.2 seconds per denial. Human review is physically impossible at that pace. The “denial” is automated; the “review” is a checkbox.
- 36% of denials have no specified reason — the plurality category is literally “Other.” This is opacity as strategy: consumers cannot contest a denial whose reason has not been stated.
- The structural analog to federal accountability: Same architecture as Minneapolis ICE evidence seizure, Epstein file slow release, and Catholic Church statute-of-limitations exploitation. Different institutions, different stakes, identical mechanism — make participation expensive enough that the overwhelming majority of affected people do not show up.
Entities Mentioned
- Kaiser Family Foundation — the research institution behind the analysis
- Cigna — insurer using PxDx AI system (context from related reporting)
- Elevance Health — lowest denial rate among major insurers (8%)
- Oscar Health — highest denial rate among major insurers (25%)
Concepts Mentioned
- Institutional Gaslighting — the architecture of exhaustion-as-strategy at consumer scale
- Mechanical Turk Pattern — the hidden labor of algorithmic denial review (indirectly)
Quotes
Not quote-driven; the data is the evidence. KFF’s analysis framing:
“Fewer than 1% of denied in-network claims were appealed in 2024.”
Notes
Kaiser Family Foundation, January 30, 2025. Annual report on HealthCare.gov marketplace plan data; CMS provides the underlying data under ACA disclosure requirements. The ACA reporting requirement covers only HealthCare.gov plans — employer-sponsored, Medicaid, Medicare Advantage, and off-exchange individual plans are not included. The true national denial figure is almost certainly larger. The <1% appeal rate is the single most powerful statistical data point in the wiki for arguing that institutional architecture depends on exhaustion-driven non-participation rather than substantive merit.