The change that breaks a clean pair without telling you
NCCI v33.0 reports land this week. One category of change produces no warning at all.
NCCI v33.0 reports land this week. One category of change produces no warning at all.
What Changed
NCCI PTP Version 33.0 change reports land this week. Effective January 1, 2027, implementation January 4, under CR 14611 / Transmittal 13946.
What it does to your claims: six working weeks from today to your first real volume under the new edits.
Coding Desk
Three kinds of NCCI change, and only one is dangerous
Every quarterly NCCI update contains three categories. Your team treats them as one list. They behave completely differently.
New edit pairs. A pair that was billable together is not any more. These announce themselves. You submit, you get rejected, you learn. Annoying, visible, self-correcting within a week.
Deleted edit pairs. A pair that was bundled is now separately billable. Nothing breaks. You under-bill until somebody notices, which is a revenue question rather than a compliance one.
Modifier-indicator changes. This is the dangerous one.
An indicator of 1 means a modifier can override the edit when clinical circumstances justify it. An indicator of 0 means no modifier will ever override it. When a pair moves from 1 to 0, your coders keep appending the modifier they have appended for two years, the claim now fails, and nothing in your workflow told them the rule changed.
Worse, a pair moving from 0 to 1 means you have been writing off a bundled service that is now billable with the right modifier. No rejection ever tells you that.
The six-week plan
- Nov 17–21 · Do this: Pull the change report. Filter to modifier-indicator changes only. Cross-reference against your top 20 code pairs by volume.
- Nov 24–Dec 5 · Do this: Review each affected pair with a certified coder. Decide what changes in practice, not just in the table.
- Dec 8–19 · Do this: Load v33.0 into the scrubber. Brief coders on modifier-indicator changes only, since new edits surface in normal work.
- Dec 22–31 · Do this: Holiday coverage. Do not configure anything in this window.
- Jan 4 · Do this: First real volume. Watch week-one rejections closely, because whatever you missed appears here.
Brief your coders on the indicator changes alone. A list of 400 new edits produces no retention. A list of six pairs they touch every day produces behavior change.
Denial Watch
The MIPS electronic prior authorization measure starts with the CY 2027 performance period
If you report MIPS, a new Promoting Interoperability measure arrives January 1: a yes/no attestation on your adoption of the payer Prior Authorization API, or a claimed exclusion.
This applies to MIPS clinicians, hospitals and critical access hospitals. The API obligation sits with payers, and the attestation sits with you.
Two things to do before year-end. Ask your major payers which API they will offer and when, since compliance dates differ by payer type. Then confirm with your EHR vendor that they support the electronic prior authorization standards, because attesting yes without the capability is not an option you want to discover in Q4 2027.
Also Worth Watching
Five telehealth codes in the CY 2027 proposals
The CY 2027 proposed fee schedule would add five telehealth codes covering clinical-staff advance care planning, shared medical appointments, pediatric speech therapy, and vaccine adverse-effect management.
These are proposals. Configure nothing. The reason to read them now is that each one describes a service your clinicians may already deliver and bill some other way, or not bill at all. Shared medical appointments in particular have lived in a coding grey zone for years.
If any of those four describe real volume in your organization, the final rule is worth reading the week it posts rather than the month after.
AI in Practice
The modifier-indicator problem is a configuration problem wearing a training problem's clothes.
Your coders know the modifier rules. What they cannot know is that an indicator changed in a quarterly file they never see. Running retraining at that gap burns coder hours and fixes nothing, which is the most common wasted response to a quarterly update.
CODIO AI resolves NCCI edits, MUE limits and LCD/NCD coverage while the claim is being built, and quarterly updates apply automatically with nothing for your team to retrain on. A pair whose indicator flipped on January 1 resolves correctly on January 4 without anybody reading a change report.
One Thing to Try
Pull your top 20 code pairs by volume this week. Not your top 20 codes. Pairs. Most teams have never produced that list, and it is the only input the next six weeks actually require.