NEMT trips with no matching medical claim: how Medicaid finds them and the proof that stops a recoupment

Medicaid pays for a ride because it gets a member to covered care. So the fastest way for an auditor to test a transportation claim is to look for the care. If a rider was driven to a clinic on March 3 and no medical claim exists for that rider on March 3, the trip lands on an exception list. Federal and state auditors have run this match on NEMT for at least 25 years, and the trips that end up refunded are the ones where the transportation company cannot show where the rider went or why.

How the match works

The test starts in the claims data. Auditors pull every paid transportation claim for a period and look for a claim for a medical service or product billed for that rider on that date. Anything left over is “unmatched.”

The raw list is always too long, so the next step is cleaning it. In Indiana’s 2001 audit, an initial listing for fiscal year 1999 put more than $1.3 million in transportation claims on the unmatched list. Working with the state, federal auditors stripped out trips where Medicare had paid the visit, trips billed under a parent when the child was the patient, and visits whose medical claim was still pending or had been denied. What remained for one quarter was 4,104 claims worth $157,948.

Then the auditors sample and dig. They ask the transportation company for its trip record, read the destination, and contact the medical provider at that address to ask whether the rider had an appointment. In the 2021 Massachusetts audit, auditors checked federal claims data for each sampled ride and sent a letter and questionnaire to the medical provider listed on the rider’s PT-1, the form that documents the need for the ride. Under North Carolina’s rules as the 2016 audit described them, counties could verify a visit by “calling the provider or obtaining the provider’s signature on a verification form.”

The legal hook is simple. Federal rules define Medicaid transportation as travel the agency finds necessary “to secure medical examinations and treatment” (42 CFR 440.170). New York’s inspector general turns that into audit criterion 5 in its ambulette and taxi protocols, both revised July 22, 2026: the medical service could not be corroborated, and claims for trips that were not to or from the location of a covered Medicaid service will be disallowed.

What the audits found

The share of unmatched trips that turned into refunds varied widely. These are the federal and state audits that tested NEMT claims against a same-day medical service, and what each turned up.

Two patterns run through the reports. First, a flag is not a finding. In the 2001 Indiana sample, 35 of 100 unmatched trips had adequate proof of a real visit. Second, the trips that failed usually failed on paper. Indiana’s 2020 audit lost 17 claims because the four companies involved had closed, moved, left the program, or could not find the records. In the 2021 Massachusetts audit, 62 of 100 sampled lines lacked enough documentation: brokers had no driver fact sheet for 38, and the sheets for 24 more were missing items such as addresses, the date, or the appointment time.

The work is not finished. On May 28, 2026, HHS OIG announced another round of state reviews (series SRS-A-26-028) to test whether the states it picks paid NEMT claims by the rules, including the rule that providers keep records backing each service. It expects to finish in fiscal year 2028.

Real visits that leave no Medicaid claim

A trip can be honest and still have nothing to match. The audit reports and state responses name these cases:

• Medicare or other insurance paid the visit. For full-benefit dual eligible riders, Medicare pays first, and CMS still requires states to provide the ride to a Medicaid-coverable service Medicare pays for. MassHealth told the state auditor in 2025 that a Medicare or commercial claim paid in full may never cross over into its system.
• Bundled and weekly codes. MassHealth pointed to G2067, a methadone treatment code that covers seven days in a row and carries only the first date of service. Six of those seven daily rides can have no same-day claim.
• Prepaid packages and follow-up visits. In Indiana’s 2001 sample, some trips went to mental health visits in a prepaid package billed earlier, and others to follow-up gynecology, prenatal, or dental visits that Medicaid had already paid for.
• Group sessions. The 2022 New York City report noted that the lack of a claim record “does not mean” the rider received no Medicaid service, giving group therapy as an example.
• A billing error at the clinic. MassHealth also noted that a provider’s billing mistake can put the visit on the wrong date.
• The rider arrived but was not seen. Indiana’s auditors accepted trips where the rider went to a valid appointment and left because of long lines or expected delays.
• The clinic canceled on arrival. Illinois lets the transportation provider bill both legs when the doctor or clinic canceled and the rider learned of it only at the destination.
• The visit was a child’s. Indiana removed trips billed under a parent or guardian when the medical service was the child’s.

Each of these cases holds up best when the trip record names a specific facility and someone there can confirm the date.

Where trip proof breaks down

The reports show the same weak spots again and again.

1. Vague destinations. In Indiana’s 2001 sample, 24 trips had destinations that auditors could not tie to any provider, most of them at or near large medical buildings. Auditors searched the claims data for nearby providers and found none billing that day. OIG’s response was direct: a general building name or street address does not document a medical visit.
2. Unrecorded stops. In the 2012 New York audit, one rider had visited two providers that day. OIG confirmed the visit only after the state named the second provider, which the transportation company had never mentioned.
3. Records that name no provider. In the 2022 New York City audit, the transportation manager’s records did not show which provider the rider was taken to, so four payments could not be confirmed either way.
4. Missing records. A closed business, a move, a fire, or a lost box of trip sheets turns sampled claims into overpayments. Indiana keeps providers responsible for those records for seven years after each service.

What to record on every trip

Build the destination proof into the trip record while the ride happens. CMS describes the main job of NEMT records as proving that every claim paid for a ride the member needed to reach covered care, or to come home from it (SMD 23-006). In practice:

1. The facility, not just the street. Record the facility name and full address, including the suite. Where your state and broker allow it, add the provider or department. North Dakota is the exception to watch: its manual says to leave the appointment provider’s name off for privacy and record only the facility name and address.
2. The appointment time. It ties the drop-off to a scheduled visit and is one of the fields Massachusetts auditors looked for on fact sheets.
3. Every stop. A pharmacy stop or a second appointment belongs on the record as its own destination.
4. A signature that fits your state. Minnesota’s trip record rule accepts the rider’s signature or the medical provider’s signature certifying the rider arrived, and requires the driver to note when both were requested and refused.
5. What went wrong, in writing. A clinic cancellation, a rider who left before being seen, or a provider who sent the rider elsewhere should go into the trip notes with the time.
6. The order behind standing trips. Keep the order or authorization for dialysis, therapy, and methadone series with the trips it covers.

Our guide to NEMT documentation requirements lists the other fields a full trip record needs, and a trip audit checklist lets you sample your own trips the way an auditor would.

Answering an unmatched-trip request

When a broker, the state, or a federal auditor sends a list of trips with no matching medical claim, treat it as a records request with a deadline.

1. Sort the list by destination. Repeat destinations such as a dialysis unit or methadone clinic can often be cleared together.
2. Gather both legs of every listed trip, with the signatures and notes that go with them.
3. Name the reason for the gap. Medicare paid, weekly bundle, group session, canceled on arrival: say which case applies.
4. Ask the medical provider to confirm the visit. A letter on letterhead listing the rider and dates is the strongest answer. In the 2012 New York audit, OIG interviewed the practitioners behind six disputed claims and reversed its finding on two.
5. Never fill gaps after the fact. Send what was recorded at the time and explain what is missing. The Medicaid audit guide covers response deadlines and appeals.
6. Return what you cannot support. If a trip truly had no covered service behind it, the 60-day overpayment rule applies, and Medicaid recoupment explains how repayment and offsets work.

Keeping trip proof in one place

An unmatched trip is won or lost on the record you kept that day. In HealthRide, each trip holds the addresses it ran between, when each pickup and drop-off happened, the GPS-recorded miles, and a signature captured on screen. When an auditor sends a list of dates, download a trip log covering them from reports in CSV or PDF.