Are Medicare audits random or predictable?

Theoretical and applied analyses suggest that audits are most effective when they are random and unpredictable. Researchers have not examined the audit strategies followed by auditors in the Medicare program. The present analysis discusses the purpose of Medicare hospital audits and the role of financial intermediaries as auditors of Medicare Cost Reports. It employs a logit model to test the hypothesis that selection for an audit is predictable. The probability of an audit is then predicted based on a set of hospital characteristics. The results indicate that over the period 1985-1987, the model accurately predicts audit status 61% of the time. The model also predicts the future likelihood of an audit, producing an accuracy rate from 55% to 76% in the individual years studied. Given these findings, it would be advisable for Medicare financial intermediaries to periodically review and alter their auditing strategies.