Use the National Denial File to Assess "Claim by Claim" Coding of PLA Codes

 While some PLA codes have high dollar volumes and specific coverage under LCD or NCD, most of the 600 PLA codes do not have policies - so they are under "claim by claim" review, as CMS says.  Let's cal lit CbC.

I think that CbC review usually means either (A) autopay or (B) autodenial, with relatively few codes getting true manual review of actual medical records.

Let's test it with the Denials Database (here).

I filtered 2024 for codes ending in "U" - PLA codes.

This gave about 3400 lines downloaded into XLS.  (Be sure to convert from csv to xls soon.)

Column i (eye) is submitted claims, column L is denied claims.  I mean a new row "S" which is denied/submitted.

Because CMS shows 0 denied by "*", the percent denied when zero comes ot as "value!" rather than "0".

ZERO DENIALS

About 2700 lines had 0 denied (suggesting autopay).

I deleted claims with 0-10 claims (CMS had already deleted all of those that were not 0,)

This left 541 lines. About 83 lines had >100 claims, up to 629.

 Total services were 31,017 and total payments were $13M, average $419.

10-20% DENIALS

About 132 lines had 10-20% denials.

Total services were 75,660 and total payments were $27M, average $363.

45-55% DENIALS

Only 25 lines had 45-55% denials.

Total services were 2961 and total payments were $452,000, average $153.

80-90% DENIALS

Only 20 lines had 80-90% denials.

Total services for these persistent hopeful labs were 5015 and payments $130,000, at $26.

100% DENIALS

Here, we pop up to about 245 lines with 100% denials.

Total services for these scalawags were 22,730. 

Actually, they're not scalawags, because about 80% of these submitted with the GZ modifier, which expects "denial." In the other group, nearly nobody used GZ modifier.

We have no payments, but charges were $35M, so under normal circumstances, payments might have been in the $15M range.  If so, implied adjucated payments for the 22,000 services might have been $650 each.  

##

100% Denials

Some may have been submissions of known-to-deny claims, such as for appeal rights.

Others likely represent tiny percentage of all claims for large test, falling in a single denial line (condition) where by definition we've sorted for "100% denials."