Where Medicare and Medicaid edits disagree

Four kinds of divergence

• Pairs only one programme has. A combination that bundles under Medicaid and not under Medicare, or the reverse. Checking against the wrong file returns nothing, and nothing looks exactly like a clean claim.
• The same pair, a different override. The more dangerous case. A pair can allow no override at all under one programme and permit a distinct-service modifier under the other. The claim is identical; the correct action is opposite.
• Different unit limits. Maximums differ on thousands of codes, in both directions.
• A DME pairing file with no Medicare equivalent. Medicaid publishes procedure-to-procedure edits for durable medical equipment; Medicare does not. There is no Medicare answer to compare against.

The enforcement indicator Medicaid does not publish

Medicare’s unit limits carry a value saying how the limit is enforced — per line, or per date of service, and whether the excess can be appealed. The Medicaid unit files do not carry it. Where a Medicaid limit differs from Medicare’s, whether the excess is recoverable is genuinely unpublished, and inheriting Medicare’s answer would be asserting something nobody has stated. What the enforcement types mean.

State programmes go further

The Medicaid files are a federal floor. Individual state programmes and their managed-care plans add rules on top, and a state edit will not appear in either federal file. A check against the published sets answers what the federal rules say — it does not answer what a particular state plan has layered above them, and no published dataset can.

Practical consequence

“We checked it and it was clean” is only meaningful with the programme attached. A practice billing both needs the answer twice, or needs to know which one it asked. That is why the programme is a selection on the claim rather than a setting somewhere in a profile — it changes the answer, not the label on it.