Changelog

The autocoder applies its radiology claim rules again. A configuration defect switched off every radiology rule, so none of the radiology alerts could appear on a claim. The rules now load and run.

Add-on codes are checked on every radiology claim. When a claim carries an add-on code and none of the primary codes CMS accepts for that add-on, the claim carries the alert "IR Add-On Code Without CMS-Acceptable Primary". The alert names the acceptable primary codes. The check reads the CMS add-on code edit data. It stays silent when CMS defines no primary list for the add-on. This check runs on diagnostic radiology claims as well as interventional radiology claims.

The radiology bundling and documentation alerts can appear on any radiology claim whose codes match the rule, not only on interventional work. The July interventional radiology update lists those alerts by name.

What to look out for: expect these alerts on ordinary diagnostic radiology claims that never raised them before. The checks were not running at all, so a claim that passed clean last month can raise an alert this month. Your organization can ask your Hank representative to retune or switch off an individual radiology rule.

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.

Modifier G8 is now limited to the six codes CMS names for it. Those codes are 00100, 00160, 00300, 00400, 00532 and 00920. On any other code the autocoder removes G8 and applies QS in its place.

An anesthesia line now carries exactly one of QS, G8 or G9. These modifiers can no longer stack on the same line. QS is the default for monitored anesthesia care. G8 wins when the code qualifies for it, then G9, then QS.

Three new review flags. A line that keeps G8 or G9 carries an informational note to verify the documentation. A line with G9 is flagged when no severe heart or lung diagnosis is coded on the claim. A physical status of P3, P4 or P5 is flagged when no supporting condition is coded on the claim.

What to look out for: expect G8 to disappear from cases outside the six approved codes, and expect QS in its place. Expect new review flags on G9 lines and on claims with a high physical status. When the record supports the modifier or the physical status, confirm that the supporting diagnosis is coded on the claim.

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.

A claim that finishes with no primary anesthesia line is held. The hold fires when an anesthesia claim has no anesthesia line that carries a code. It catches a line that was produced and then removed later, and it catches a case where no code was ever found. A labor epidural (01967) counts as a valid anesthesia line and does not cause the hold.

A claim whose first ranked diagnosis cannot be traced is held. The autocoder must trace the first ranked diagnosis to a stated reason for the procedure. It looks at the anesthesia record after the case, then the record during the case, then the pre-anesthesia evaluation, then the surgeon operative report. The alert names the document that it used.

Modifiers are placed in the order that you configure. Your organization sets the order for surgical modifiers and for anesthesia modifiers. A modifier that is not in your list goes after the listed ones, in alphabetical order.

Two more holds started to fire on 2026-07-01. Those two rules never fired before. One holds a claim with no primary anesthesia code. The other holds a claim with a date of service more than one year old.

What to look out for: organizations that bill procedures only (ICU lines, emergency department blocks, blood patches, canceled cases) will see more held claims. So will organizations that take the first ranked diagnosis from a problem list. Every hold above can be turned off or softened for your organization.