OIG flagged all 399 sampled optometry E/M lines. The SMRC is now reviewing 99309 and 99310.

If your practice sends anyone into a nursing facility, the level billed on those visits is now a review target. Noridian Healthcare Solutions, the CMS Supplemental Medical Review Contractor, lists project 01-167, a nationwide post-payment review of Part B claims billed with CPT 99309 and 99310. On July 16, 2026, CMS ran a compliance item in MLN Connects titled "Optometry Services at Nursing Facilities: Bill Correctly." Both trace back to an OIG audit issued in December 2025.

What the OIG audit found

OIG's data analysis started from a simple pattern: Part B payments to optometrists for high-level E/M services optometrists don't usually bill. For 2021 through 2023, Medicare paid $4.7 million to 200 optometrists for moderate-to-high-complexity subsequent nursing facility care. The top 15 accounted for 72 percent of that.

OIG then audited those 15. It pulled a stratified random sample of 15 enrollees from each, 225 in total, and reported that every one of them had claim lines that didn't meet Medicare documentation or coding requirements. Those 225 enrollees carried 414 lines of service, and all 399 of the E/M lines billed as 99309 or 99310 failed.

That isn't a projected failure rate. Every E/M line OIG pulled came back unsupported.

Read the dollar figure correctly

OIG's headline recommendation cites $3,059,204 in estimated overpayments. That number is the lower limit of a two-sided 90-percent confidence interval. Appendix D puts the point estimate at $3,272,823 and the upper limit at $3,486,443.

The distinction matters if anyone quotes this in a meeting. It's a projection from the 225-enrollee sample onto the claims of the 15 audited optometrists, and OIG's recommendation limits recovery to the portion inside the 4-year reopening period. It is not a bill, and it says nothing about the other 185 optometrists in the original data pull.

Why nothing stops this at the front end

Here's the part worth reading twice. OIG made three recommendations. CMS concurred with two and didn't fully concur with the third, which asked it to increase claim reviews and develop system edits to prevent incorrect billing of these codes.

In OIG's recommendation tracker, the training-and-guidance recommendation closed as implemented on March 30, 2026. The two that would change how these claims get screened are both still open and unimplemented, with updates expected October 15, 2026 for increased claim reviews and October 14, 2026 for system edits.

No system edit means no front-end rejection. A 99310 the record doesn't support will pay, and the correction shows up later as an Additional Documentation Request. That's the opposite of an NCCI edit that rejects the pair up front, and it decides where you have to catch the problem: in your own pre-bill review, because nothing downstream will.

From audit report to nationwide review

About seven months separate the OIG report from the date now showing on the SMRC project page. Noridian publishes a Last Updated stamp rather than a project open date, so treat that gap as approximate. The observation is ours, not CMS's.

What a self-audit on these codes has to check

Start from every 99309 and 99310 with a CY2023 date of service. That's the population the SMRC named. From there, here's where we'd look. None of it is new. It's the same level-of-service reading any E/M audit applies, pointed at one code pair and one year.

Pick the basis before you defend the level. A note built on time and a note built on MDM get defended with different content, and a record gesturing at both usually establishes neither. If the evidence in the note doesn't reach the level billed, that's a refund conversation, not an appeal.

The pattern travels past optometry

The billing specialty here is narrow. The sequence isn't.

A published OIG audit named a specialty and a code family in December. Roughly seven months later a nationwide contractor review of that exact code family appears, covering dates of service still inside the reopening period. If a specialty you code has shown up in an OIG report, treat that report as advance notice of a medical review rather than a finished story, and pull the records while correcting them is still your own call.

Two limits worth keeping straight. The SMRC publishes no error rate for project 01-167, and an audit of 15 providers selected as outliers tells you nothing about how the other 185 billed. Confirm any descriptor against the AMA CPT Professional Edition the SMRC cites, or the official code reference, rather than an internal cheat sheet that may predate the 2023 change. Insufficient documentation is already a leading reason these levels get taken back, and the SMRC isn't going to read a chart more generously than a MAC does. Whether the note carries the evidence is still the whole question.