Medicare Billing Integrity Detection System

Composite Risk Score Framework for Hospital Billing Pattern Analysis

$31.70B
Medicare FFS Improper Payments
FY 2024 (7.66% rate)
27.5M
Medicare Discharges Analyzed
2019 to 2023
101
High Risk Hospitals
3.5% of 2,911 examined
94.3%
Model Accuracy
Detection performance
42,708
Excess MCC Cases
Annually identified
$213.5M
Study Estimated Overpayments
101 hospitals (95% CI: $198.3M to $228.7M)
Data Completeness
99.3%
Cross Validation
91.3%
Precision
89.7%
Recall
91.2%
Risk Stratification Distribution (2,911 Hospitals)
Composite Risk Score Components
Geographic Distribution of High Risk Facilities
Model Performance Metrics
Risk Stratification Results
Risk Category Score Range Hospitals Confirmation Rate
Very High Risk 90 to 100 1 100%
High Risk 75 to 89 100 88%
Medium Risk 50 to 74 1,403 12%
Low Risk 0 to 49 1,407 2%
Key Billing Pattern Indicators
Metric Value Status
MCC Rate Excess 20.1 percentage points above peers (64.1% vs 44.0%) Significant
Temporal Persistence 0.84 coefficient (99 of 101 facilities across 3+ years) Persistent
Bootstrap Confidence Interval $198.3M to $228.7M (95% CI, 1,000 iterations) Validated
Statistical Significance p < 0.001 Strong
PEPPER Concordance 88% agreement with CMS benchmarks Validated
Model Validation Metrics
Metric Result Interpretation
Accuracy 94.3% 2,743 of 2,911 hospitals correctly classified
Precision 89.7% 131 of 146 flagged facilities confirmed anomalous
Recall 91.2% 131 of 144 actual anomalies detected
F1 Score 90.4% Balanced precision recall performance
Cross Validation 91.3% 5 fold stratified sampling average accuracy
Bootstrap CI $198.3M to $228.7M 1,000 iterations with stable estimation
Key Analysis Statistics
Data Analysis
27.5 million Medicare discharges analyzed (2019 to 2023)
Facilities Examined
2,911 acute care hospitals comprehensively evaluated
Validation Method
Bootstrap Resampling (n=1,000) with 5 fold Cross Validation
Federal Context
324 defendants in National Health Care Fraud Takedown ($14.6B schemes)
GAO Impact
$675B financial benefits from addressing high risk areas over 17 years
Data Sources
MEDPAR (47 variables), Hospital Compare (23 metrics), AHRF, PEPPER Reports, OIG Work Plan
Important Note: This system identifies statistical anomalies in billing patterns. Identified patterns indicate facilities warranting further investigation; actual improper payment determinations require clinical review and appropriate enforcement procedures by authorized agencies.