How it usually reaches a practice
Most healthcare fraud matters begin with data rather than with an accusation. A payer or its contractor compares your claims to those of similar providers, and something stands out: a code used more often than peers use it, visits of a length that repeat, a supply billed alongside a procedure. Others begin with a complaint from a former employee, a patient, or a competitor. What arrives first is a records request, an audit finding with a demand for repayment, or a notice that claims are being reviewed before payment. A pattern that looks unusual in a dataset frequently has an ordinary explanation in the charts, such as an unusual patient mix, a template in the record system, or one staff member's habits.
Where a billing problem becomes something else
The line reviewers look for is not whether a code was wrong but whether the documentation supports it, who directed the billing, and what happened when someone raised a concern. Emails about production targets, instructions to staff, and unresolved compliance reports carry more weight than the claims themselves. One thing deserves a warning of its own: do not correct, complete, or tidy records after a request arrives. Electronic record systems keep an audit trail of every edit, and late changes tend to become the most damaging part of the file, whatever the intention behind them.
What to gather and what runs in parallel
Collect the request or audit letter, the specific claims and charts at issue, your billing and coding policies, contracts with any billing company or coder, compliance committee records, prior audit results, and the payer agreements themselves. Appeal windows in those agreements move quickly and are easy to lose while everyone is gathering documents. A payer audit, a contractor review, a whistleblower suit, and a government inquiry can all touch the same claims on separate schedules. In a first conversation we work out which of those you are actually in, what the charts support, and what should be said in writing first.