Changelog

Some EMRs print the CPT code that was selected when the case was booked. The booked code does not always match the procedure that was performed. A common example is a colonoscopy booked under a polypectomy code when no polyp was removed.

The autocoder now treats a printed code number as scheduling data, not as proof. The words in the record still count, including words in the booking or scheduling area. But the autocoder does not expand a bare code number into its full code description to add details the record does not state.

What you will see differently:

• Surgical codes follow the documented procedure. If the booking shows a polypectomy code but the report documents a diagnostic exam with no polyp removal, the claim carries the diagnostic code.
• Anesthesia codes for GI cases no longer inherit polyp removal or biopsy from the booked code. A screening colonoscopy with no intervention codes as a screening exam, with the correct payer handling for the PT modifier.
• Diagnosis codes tied to an assumed intervention (for example, a colon polyp diagnosis) appear only when the record documents the finding.

What to look out for: claims where the booked code and the documentation disagree now follow the documentation. If a procedure was performed but not documented, the claim will not include it. Ask the provider to complete the record, then resubmit the case.

Reference material now matches the date of service. When the autocoder looks up code reference material while it codes a case, it reads only the reference set for the claim's date of service. CPT references follow the calendar year. Diagnosis references follow the October to September fiscal year. Before this change, a case coded for an earlier date of service could see current year material.

This matters most on back dated work: a claim for a service in a prior year is now coded against the code set that was in force on that date, not against the current one.

What to look out for: back dated cases should select codes that were valid on the date of service. When a case carries no date of service, the autocoder still reads every year, as it did before. If a claim needs the date of service to be right for this reason, confirm the date before you reprocess the case.

The autocoder can now build a companion claim for the operating surgeon alongside the anesthesia claim, from the same case documents. The feature is OFF by default and is enabled per organization on request.

What it does when enabled:

• Surgeon-authored documentation (operative notes, procedure reports) is coded into a separate surgical claim with the surgeon as the billing provider.
• A deterministic Medicare fee schedule engine applies the payment rules: multiple-procedure ranking, bilateral handling, assistant and co-surgeon checks, and modifier placement.
• Screening colonoscopy conversions follow the payer rules (PT or 33), and the anesthesia claim and the companion claim are cross-checked so the two never disagree on the screening story. A disagreement is held, not billed.
• Every companion line carries its evidence quote from the surgeon's own document. A line whose evidence cannot be verified in the source document is excluded and flagged.

What to look out for: nothing changes unless your organization asks for the feature. If you want surgeon-side claims from the documentation you already send, contact your Hank representative.

Code 29826 can no longer stand alone as the surgical code on a claim. 29826 is arthroscopic subacromial decompression, an add-on code. An add-on code reports work that is done with a base procedure, so it cannot be the only surgical code on a claim. The autocoder now holds a claim whose only surgical code is 29826. Codes 99100 and 99140 were already held this way, and 29826 joins them.

What to look out for: a shoulder arthroscopy case documented only as the decompression will hold until a coder supplies the base procedure. The base procedure is usually stated in the operative report. Find it, add it, and reprocess the case. If the record documents no base procedure, ask the surgeon for the missing detail before you bill the case.

Cystoscopy coding is corrected. Before this change, the autocoder sent laser stone cases to the tumor codes.

A laser used on a bladder tumor codes to the tumor codes. Transurethral resection of a bladder tumor with a laser codes to 52234, 52235, or 52240. The code follows the size of the tumor.

A laser used on a stone codes to the lithotripsy codes. Cystoscopy with laser lithotripsy codes to 52353. When a stent is placed at the same session, the case codes to 52356 instead. Placing a stent by itself codes to 52332.

What to look out for: the anesthesia code can change with the surgical code, because the anesthesia code follows the surgical procedure. Review your cystoscopy cases that document a laser after this update, and confirm that the record states what the laser treated. A record that says only "laser used" does not separate tumor work from stone work.

A code billed as the anesthesia code must belong to the anesthesia code family. The autocoder now checks the anesthesia code against that family. A code outside the family raises a review alert on the claim. A code that is not a real CPT code at all holds the claim.

A code valid only in 2025 could reach a 2026 claim. The check that validates surgical coding was broken, so a retired code could pass onto a claim with a 2026 date of service. That check works again. The autocoder is also told the target code year while it selects codes.

What to look out for: expect alerts that name an anesthesia code which does not belong to the anesthesia code family. Read the record, then select the correct anesthesia code. Review any recent claim that carries a code retired at the end of 2025.

Codes from a handwritten record now carry a real confidence value. Before this change, every code from a handwritten note was stamped with a confidence of 33. The autocoder now scores the handwriting on how legible it is. Clear handwriting scores high. Genuinely illegible handwriting scores low, and the reason appears in the rationale.

Paper charts produce fewer held claims. The low confidence holds on anesthesia codes, surgical codes and diagnosis codes no longer fire when the autocoder reads the handwriting confidently. An illegible record still scores low and still holds.

Code selection changed on records that mix typed and handwritten notes. A legible handwritten code now competes on equal footing when the primary code is selected. A code that scores below 30 drops out of that selection.

What to look out for: expect fewer review queue items on paper charts. If your team used the confidence value of 33 as a marker for "this is a paper chart", that marker no longer exists. Use the handwritten alerts on the claim instead. They state that a handwritten note is the source.