Working hackathon MVP · Investigation support
Give investigators a prioritized path through public Medicare data.
Integrity is a hackathon-proven foundation for explainable provider-risk decision support. It produces traceable signals for human investigation and does not make fraud findings.
Maturity and proof: Integrity is a company-led ACT-IAC hackathon winner built on public data. The metrics below document the hackathon build and analytical validation. They do not establish federal production past performance, an agency fraud determination, or realized savings.
The hackathon build packages ingestion, scoring, and an interactive investigation dashboard in one Go binary. Its evaluated corpus combined 22 CMS datasets and 10 years of history, scored 1.47 million providers, and constructed an 82.6-million-edge analytical graph from processed public records. Twenty-seven features across eight categories feed composite, ML-corroboration, and gradient-boosted-tree scoring methods.
Out-of-sample validation matched 37 flagged providers to documented enforcement cases involving over $500M in alleged or resolved fraud. Those providers were never in the training data and were flagged from billing-pattern anomalies. Average detection lead time: 6.9 years before OIG exclusion. Read the full case study for methodology, backtest results, and out-of-sample validation details.
What the Hackathon Build Includes
- Public data only: No PHI, data use agreements, or claims-level access required
- Live weight adjustment: Investigators tune feature weights in real time and watch scores re-rank immediately
- Full transparency: Every score traces to source data, SQL queries, peer groups, and weights so investigators can inspect how it was produced
- Single binary: DuckDB, in-process GraphWizard, and all assets compile into one deployable system
- GraphWizard integration: 6 graph algorithms calculate features associated with shared relationships, prescribing isolation, and exclusion proximity
What Production Still Has to Prove
A production engagement would define the authorized public-data corpus, investigative workflow, validation measures, explanation needs, access controls, accessibility criteria, deployment environment, and operating responsibilities. The hackathon results are a starting point; usefulness and controls must be tested with the people who would operate and oversee the configured capability.