Multi-Method Detection: Providers Flagged by Multiple Independent Systems

Why Multi-Method Detection Matters

Any single fraud detection method has limitations. Statistical tests can flag legitimate outliers. Benford's Law can misfire on providers with standardized fee schedules. Velocity checks may catch providers who genuinely expanded services. But when two or three independent methods all flag the same provider, something unusual is almost certainly happening.

If one test has a 10% false positive rate and another independent test also has a 10% false positive rate, the chance both would flag the same innocent provider is just 1%. Add a third independent method and the probability of a triple false positive drops to 0.1%. This is the power of multi-method detection.

Detection Methods in This Dataset

Most Common Method Combinations

Key Takeaways

• ▸246 providers were flagged by 2 or more independent detection systems, making them the highest-confidence anomalies in our dataset.
• ▸30 providers were flagged by 3+ methods — the probability of a triple false positive in independent tests is extremely low.
• ▸The most common combination is Statistical + Velocity, found in 96 providers.
• ▸The top 50 multi-method providers account for $23.87B in total Medicaid spending.
• ▸Multi-method flags are the strongest signal available, but they are still screening results — not verdicts. Each case requires individual review.