Applying analytics to anomalies
We began this stage by developing hypotheses for analyzing the data and applying models to validate them. These included a series of risk-weighted analytics that were used to test the hypotheses.
Then, working from the assembled SQL database and local compliance event spreadsheets, we performed trend and outlier analysis to identify potential irregularities in the event processes, parties, and fields that could require a deeper dive or red flag identification. Examples included:
• Anomalies between the various event populations provided, such as missing events, missing fields, or conflicting key fields
• Anomalies around event costs or counts based on key fields associated with the allegation, such as irregularities in average event costs by type, business unit, business owner, organizer, and participants.
We also performed a trend analysis of anomalous activity by parties cited in the allegations, including HCOs, doctors, other HCPs, and other third-party businesses and individuals.
Finally, to support our findings we conducted a follow-up technical analysis of invoices, payments, contracts, and other documentation.