Key Takeaways
- CMS’s Medicaid program integrity includes both fee-for-service and managed care integrity efforts, reflecting the measurable scope of Medicaid payment types covered
- HHS-OIG reported 68% of audited Medicaid payment samples included some form of improper payment in the OIG audit summaries where methodologies found errors in claims (improper payments and program integrity findings)
- $29.6 billion in estimated improper payments for Medicaid managed care in FY 2022, representing the modeled improper dollar amount for the managed care portion
- 6.2% was reported as the improper payment rate for Medicaid in 2021 (improper payments as a share of total program payments in the improper-payment estimate)
- In 2023, identity theft complaints filed through IC3 included medical/health related themes; the report provides counts by complaint type and flags medical identity theft as a category with measurable totals
- Medicaid covered over 90 million people in 2022 in the United States (measured enrollment reported by CMS)
- In 2020, the Medicaid program’s payment integrity results included an estimated $4.6 billion attributable to underpayments, measuring another component of improper payment dollars
- In 2021, Medicaid improper payment rates were reported as 6.2%, measuring the estimated proportion of Medicaid payments that were improper
- $24.4 billion in improper payments were estimated for Medicaid overall in 2019 (CMS payment integrity reporting), measuring the total modeled improper dollar value
- 7.3% of Medicaid provider claims reviewed in a GAO case study were found to contain billing errors, measuring the observed error prevalence in a sample-based review tied to improper billing
- A GAO review found that 33% of states did not fully implement provider enrollment integrity steps, measuring gaps in controls that can allow improper or fraudulent billing
- In a GAO case involving questionable billing, 1 in 5 claims (20%) in the reviewed sample had billing errors related to coding or billing compliance issues, measuring a concrete observed error rate in that study context
- In 2021, 65% of surveyed organizations reported that they used provider/network data analytics to detect fraud, measuring adoption of data-driven detection approaches relevant to Medicaid integrity
- In a 2022 ACFE survey, organizations with fraud detection programs were found to lose 50% less than those without such programs, measuring the quantified benefit of detection controls
- In 2022, 44% of healthcare organizations reported that they had implemented risk scoring for claims/provider monitoring, measuring adoption of risk-based controls used to prioritize investigations
Medicaid improper payments and fraud controls remain significant, with billions at risk and many states still missing key integrity steps.
Related reading
01 · Category
Improper Payments4 stats
Improper Payments Interpretation
02 · Category
Improper Payment3 stats
Improper Payment Interpretation
03 · Category
Fraud Typologies3 stats
Fraud Typologies Interpretation
More related reading
04 · Category
Detection & Controls3 stats
Detection & Controls Interpretation
05 · Category
Cost & Recoveries2 stats
Cost & Recoveries Interpretation
06 · Category
Industry Overview5 stats
Industry Overview Interpretation
Medicaid improper payments: rate and dollars over time
Improper-payment measures show both a share-rate of improper payments and large dollar estimates across multiple years.
Cite This Report
This report is designed to be cited. We maintain stable URLs and versioned verification dates. Copy the format appropriate for your publication below.
Emilia Santos. (2026, February 13). Medicaid Fraud Statistics. Gitnux. https://gitnux.org/medicaid-fraud-statistics
Emilia Santos. "Medicaid Fraud Statistics." Gitnux, 13 Feb 2026, https://gitnux.org/medicaid-fraud-statistics.
Emilia Santos. 2026. "Medicaid Fraud Statistics." Gitnux. https://gitnux.org/medicaid-fraud-statistics.
Sources & references
20 datasets cited across this report · attribution is report-level
+9 additional datasets cited (not shown individually)

