Fraud Analytical Skills
What this domain tests
Fraud Analytical Skills covers the reasoning habits that turn raw claims data into defensible indicators: benchmark selection, stratification, triangulation, base rates, and professional skepticism. The exam rewards analyses that separate an anomaly from a fraud indicator.
Core rules
- A red flag is a prompt for inquiry, never proof. Indicators must be tested against legitimate explanations.
- Comparison is only as good as the comparator. Match on line of business, specialty, geography, policy size, and time period before calling a value high.
- Use rates, not counts: claims per 1,000 policy-years, paid per exposure, services per 1,000 members.
- Triangulate at least two independent signals (billing pattern plus ownership link plus timing) before escalation.
- The fraud triangle - pressure, opportunity, rationalization - organizes hypotheses; it does not prove anything.
- Document why an indicator was rejected as well as why it was escalated.
Worked example
A chiropractor shows 40 visits per patient per year against a statewide average of 12. Reasoning shown: the statewide figure mixes specialties, so it is the wrong comparator. Rebuild the benchmark using chiropractors in the same region and similar panel size; the peer median is 22. Then examine service mix: 90% of visits bill the same three CPT codes, 31% of visits fall on Saturdays when peer clinics are closed, and 14 patients share one address. The volume alone would be weak; volume plus code uniformity plus scheduling pattern plus address clustering makes a case worth referring.
Traps that cost marks
- Benchmarking against a broader population than the subject's true peer group.
- Treating high utilization as fraud without checking severity mix.
- Cherry-picking the two months that look worst.
- Failing to record the alternative explanations that were tested and rejected.