Two things about 92507 are both true and a little maddening. It's the code speech-language pathologists bill more than any other, and it's the one that gets kicked back more than almost any other. The denials, though, usually have nothing to do with the therapy. You can run a solid session, write a tidy note, and still lose the money because a modifier was missing or two services landed on the same date when they shouldn't have.
So most of the gap between doing good work and actually getting paid is paperwork. Which makes it worth getting specific about a few things: what the code covers, what Medicare pays for it in 2026, the modifiers that decide whether a claim lives or dies, and the boring mistakes that quietly drain money out of speech therapy practices.
What CPT 92507 Actually Describes
Here's the official line from the AMA: "Treatment of speech, language, voice, communication, and/or auditory processing disorder; individual." Picture one patient and one SLP working a communication goal, and you've basically got it.
The reason this code turns up everywhere is that it covers a lot: articulation drills, fluency work with a kid who stutters, voice rehab after a vocal-fold injury, expressive and receptive language, cognitive-communication therapy after a stroke. If the session is one-on-one, it bills as 92507. One code, most of the caseload.
What it doesn't cover trips people up just as often. Evaluations go to 92521 through 92524. Group therapy is 92508. AAC device training has its own code, 92609. Timed cognitive work runs through 97129 and 97130. The moment a service stops being individual, skilled, and focused on communication, 92507 no longer fits, and jamming it in anyway is a quick route to a denial.
The Untimed Trap: The Costliest Misunderstanding
If one thing costs practices money over and over, it's this, and it isn't complicated. 92507 is untimed. One patient, one day, one unit. The clock doesn't matter. A twenty-minute session and a sixty-five-minute session bill exactly the same.
People coming out of the fifteen-minute world of PT codes want to stack units on a long session. Can't be done. Medicare's Medically Unlikely Edit caps 92507 at one unit a day, and the commercial payers mostly copy that. Say a school SLP pulls the same student twice in one afternoon for two short sessions. That's one unit, not two. Bill two and it auto-denies, and now you've got an audit flag you never asked for.
There's a narrow exception, and it needs proof. If you actually provided a separate, separately documented service the same day, modifier 59 can split them out. But only if the notes show two genuinely different goals. Reviewers know what one session written up twice looks like.
2026 Reimbursement: The Numbers as They Stand
Now the money. Medicare's national non-facility rate for 92507 in 2026 sits around $85.53, up roughly two percent from last year. That's an average, not a promise. Your actual payment moves with geography, so a clinic in rural Idaho won't see what a clinic in downtown Boston sees. If you want your real figure, your Medicare Administrative Contractor has it.
This year there's an odd wrinkle underneath the rate. For the first time, the fee schedule runs two conversion factors: about $33.40 for the non-APM clinicians, which is most SLPs, and $33.57 for the ones inside qualifying advanced alternative payment models. Both already include the one-time 2.5 percent raise Congress tucked into the budget bill.
And here's some actual good news. CMS rolled out a 2.5 percent "efficiency adjustment" for 2026 that shaved reimbursement off thousands of non-timed codes. 92507 wasn't on the list. Codes sitting right beside it took the cut; this one didn't. If your revenue rides on this single line, and for plenty of practices it does, that exemption is a real cushion against a year that's otherwise pushing rates down.
Commercial payers do their own thing. The 2026 ballpark puts Blue Cross Blue Shield near $96 a session and UnitedHealthcare somewhere between $85 and $90, with Medicaid all over the map depending on the state. Treat every one of those numbers with suspicion until you've confirmed your own contract. They're fine for budgeting and useless for forecasting.
The Modifiers That Make or Break the Claim
More 92507 claims die on the modifier line than anywhere else. Get them right and the claim sails through. Get one wrong and you're writing an appeal.
GN is non-negotiable on Medicare Part B. It tells Medicare the service belongs to a speech-language pathology plan of care, and they want it on every SLP claim. Forget it and the claim won't even process. It's the most common denial in the whole category, and the most avoidable.
KX is about thresholds. For 2026, the combined PT-and-SLP threshold is $2,480 per patient for the year, $70 more than 2025. It's not a hard cap. Medicare keeps paying past it as long as the care is still necessary. The catch is that every 92507 claim after that point needs KX attached, basically you signing off that treatment is still warranted and the chart proves it. And the total sneaks up fast. A patient seen four times a week hits $2,480 in about seven weeks, and clinics without a tracking system usually find out the hard way, through the first denial. There's a separate medical-review threshold sitting at $3,000 further along.
For telehealth, modifier 95 flags a live audio-video session, and the video part isn't optional. Audio-only doesn't fly under 92507 for Medicare. Some commercial payers want GT instead of 95, and the place-of-service code shifts depending on where the patient is. One thing to watch: Medicare's telehealth authority for SLP services has only made it through 2026 on a string of short-term extensions from Congress, so the expiration date keeps moving. Check it with each payer before the session instead of assuming it's permanent.
Two more worth knowing. Modifier 59 unbundles services that Medicare's NCCI edits would otherwise pair up, but it only holds if the documentation keeps the goals separate. Modifier 22 flags a session that ran unusually long or complex; a lot of payers will bump reimbursement 25 to 50 percent for it, though the claim then goes to manual review and wants extra paperwork to justify it. Lean on 22 too often and payers stop buying that your long sessions are unusual.
Documentation: The Note Is the Claim's Defense
A code is only as good as the note holding it up. When somebody audits a 92507 claim, they're really asking whether the diagnosis matches the service, whether the work needed a licensed clinician, and whether anyone bothered to establish the therapy was necessary in the first place. Same questions, more or less, no matter the payer.
The diagnosis piece is the easy one. You need an ICD-10 code for a real, documented communication disorder sitting in the chart, aphasia, dysarthria, stuttering under F80.81, a voice problem, and it has to actually be written down, not just living in your head as a clinical impression. A treatment code with no matching diagnosis reads as a mismatch and gets denied right away.
The skilled-service piece is where people stumble. The reviewer wants to see that this session genuinely called for a speech-language pathologist, not something an aide or a parent with a home program could've run. "Performed speech exercises" tells them nothing. They want the reasoning, the adjustments you made partway through, the calls only an SLP would make.
Necessity pulls it together, ideally with a physician-approved plan of care. Payers keep drifting toward outcomes, which means your reimbursement holds up better when the record shows measurable progress toward goals you actually named. Vague, copy-pasted notes pull audits. Specific, outcome-anchored ones get through them.
Who Is Actually Allowed to Bill It
Staffing catches new practices out. Medicare won't separately reimburse work done by SLP assistants, no matter how closely you supervise them. Students in clinical training count as unlicensed for billing too, so the supervising SLP is the rendering provider and it's their NPI on the claim.
Clinical Fellows are their own situation. They bill under a supervising SLP's NPI, with supervision rules that change state to state and payer to payer. If you supervise CFs, sort out your state's rules and each payer's policy before the first claim goes out, not after one bounces back.
Same-Day Collisions and the NCCI Edits
A lot of avoidable denials come down to billing the wrong codes together on one date. Don't put 92507 on the same day as an evaluation code, 92521 through 92524. Treatment plus evaluation on a single date is a classic audit trigger, and the few exceptions that exist mean writing out the rationale and clearing it with the payer first.
Swallowing therapy, 92526, can share a date with 92507, but only if the notes show two complete, separate services with different diagnoses and goals. CMS is also clear that SLPs shouldn't be billing PT codes like 97110, 97112, 97150, or 97530 as add-ons to 92507. Those belong to PT and OT, and they don't ride along with speech treatment.
The Mistakes Worth Memorizing
Pull back and the costly errors land in a pretty short pile. Billing 92507 like it's timed and stacking units. Dropping GN on a Medicare claim. Forgetting KX once a patient crosses $2,480. Putting 92507 next to an eval code on the same day with nothing to justify it. Notes that don't prove skilled, necessary care. A diagnosis that doesn't match the service. The striking part about that list is how little of it has to do with clinical skill. It's process, top to bottom, which is exactly why you can fix it.
What's Coming in 2027
Keep half an eye on 2027. Medicare's use of 92507 more than doubled between 2017 and 2022, and a jump that size gets the AMA and CMS looking. A Code Change Application covering SLP codes went in front of the CPT Editorial Panel at the late-April 2026 meeting, aimed at the 2027 set, so how individual speech therapy gets structured and billed could look different next year.
None of that changes anything you're doing right now. 92507 is still valid, still billable, still the backbone of SLP revenue. Keep your documentation audit-ready and actually read ASHA's reimbursement updates as they land, so 2027 doesn't catch you flat.
Conclusion
Getting 92507 right is unglamorous work. Modifiers, thresholds, diagnosis matching notes that survive a second look, the back-office stuff that keeps the lights on. The practices that take billing seriously instead of treating it as an afterthought get paid faster, clear audits without the headache, and leave themselves room to grow. Bill it carefully, document it tight, and this one little code keeps doing what it's there to do: paying for the care the patient came in for.
Make An Appintment With Us