Detecting Duplicate Claims via Machine Learning

THE BACKGROUND

Duplicate claims accounts for significant percent of all billing errors across medical insurance industry. In addition to direct revenue loss for paying the duplicate claim expenses, there is administrative waste and cost of reprocessing a claim for insurance provider. And if that claim must be reviewed by claim examiner, that cost rises even further. Soothsayer’s client had a rule-based scoring system that was very conservative leading to lot of false positives and consequently lot of unintended efforts and costs in their review. The review process was partially automated using an expensive licensed tool, but still required a lot of manual effort as examiner had to go through tons of documents and procedures to adjudicate. Client had a team of more than 200 employees to just review these suspected duplicate claim pairs.