When CMS introduced POS 10 ("telehealth provided in patient's home") in January 2022, most practices treated it as a relabel — "oh, we used POS 02 before, now we use 10 sometimes." That casual treatment is why we expect a wave of telehealth audits over the next 18 months. The codes mean different things and the reimbursement reflects that.
What each code actually means
POS 02: telehealth provided other than in patient's home. The patient is at a clinic, hospital, or other originating site that is NOT their residence.
POS 10: telehealth provided in patient's home. The patient is physically located at home (apartment, assisted living, hotel) at the time of the encounter.
The most common error
Practices use POS 10 on autopilot for every telehealth visit because the reimbursement is better. The problem: if the patient was at a SNF, an inpatient rehab, or even at work in their employer's office, POS 10 is wrong. Documentation has to show the patient was at home.
Documentation that holds up in audit is one short sentence in the note. Something like: "Patient confirmed telehealth visit conducted from her home in [city, state]." That's it. We've seen audits succeed and fail on whether that single sentence is present.
Modifier 95 vs modifier 93 — and why it matters
Modifier 95 = synchronous audio + video. Modifier 93 = audio-only. They are not interchangeable, and several payers (notably United and parts of Cigna) reduce reimbursement on modifier 93 to 80–85% of modifier 95. Some plans don't reimburse modifier 93 at all for certain CPTs.
If a video call drops to audio-only mid-visit, the modifier reflects the majority of the visit. Document the technical reason for the drop. Auditors look at this when comparing claim modifier to the EHR audit log.
Three checks to run this month
- 1Pull last quarter's telehealth claims by POS code. If 95%+ are POS 10, you have a default-coding problem. Real practice mix is usually 60/40 or 70/30.
- 2Sample 20 telehealth notes. Count how many include a sentence confirming the patient's location. If under 80%, that's your documentation gap.
- 3Compare modifier 95 vs 93 reimbursement per payer. If you're surprised by what you find, you've been losing money on audio-only without knowing it.
Telehealth coding isn't hard. It's specific. The audit risk isn't from intentional fraud — it's from coding by habit instead of by what actually happened in the visit. Take 30 minutes to look at your data. You'll either confirm you're fine or find a problem worth fixing while it's still small.