Is The Downcoding Bogeyman Real?

By now, you have heard of CIGNA's announcement to begin automatically downcoding E&M claims based on ICD-10 alone and not physician notes, etc.  I don't believe for a second that this decision has anything to do with capturing code fraud and has everything to do with creating a burden on primary care practices to improve CIGNA's bottom line.

As you can imagine, many of the physician societies were outraged, and CIGNA toned it down for a bit.  The AAP has provided some decent downcoding guidance and has even penned some letters directly to the payors you can refer to in your complaints.  

But what is the reality?  Are CIGNA and their organized-crime-like conspirators really zapping folks significantly?  With help from my friends here at PCC, I looked into it.  A lot is going on here, so settle in:

• It's not an easy thing to determine when, where, and how claims are being downcoded.  Why?  Because, as far as I can tell, the payors aren't using CARC codes to indicate a claim has been downcoded.  [I wonder if this is a HIPAA violation.]  The only way I'm aware of to determine if an E&M has been downcoded from a reporting perspective is to look in the ERA files, which show both the submitted and adjudicated CPT codes - we look for a difference.  That's right, they are not making this easy to track, address, or improve.  I would love it if the geniuses at Epic, Athena, eCW could weigh in here - they have 1000x more data than we have.
• In 2025, CIGNA is overwhelmingly hands-off.  Many PCC clients have had a claim or two adjusted, but only a small handful have had more than, say, ten claims changed.  What's interesting is that the majority of changes aren't downcodes - they are adjustments of a 99214 to a 99214-25.  That's right...it looks like they are changing the code so that the claim will be paid.  Note, I haven't analyzed whether or not adding the -25 modifier might be a sneaky way to reduce payment on the E&M, but the volume is still low.
• AETNA, on the other hand, has been very painful to a subset of our clients.  Most PCC clients have had 0-5 claim adjustments.  A small handful have had significant downcoding, however, representing as many as half of their AETNA E&M claim volume.  This began earlier in the year.
• Analysis of their ICD-10 usage on those claims is quite revealing.  Here's a list of the top 50% of the primary ICD-10 codes on the claims that were downcoded for the PCC practices that were affected the most.  See if you can identify the pattern:
  
  

  ICD-10	Description
  R509	Fever, unspecified
  R070	Pain in throat
  J029	Acute pharyngitis, unspecified
  R059	Cough, unspecified
  H66003	Acute suppr otitis media w/o spon rupt ear drum, bilateral
  J069	Acute upper respiratory infection, unspecified
  H66001	Acute suppr otitis media w/o spon rupt ear drum, right ear
  R21	Rash and other nonspecific skin eruption
  J020	Streptococcal pharyngitis
  Z79899	Unspecified external cause status
  H66002	Acute suppr otitis media w/o spon rupt ear drum, left ear
  F50019	Anorexia nervosa, restricting type, unspecified
  R0981	Nasal congestion
  R300	Dysuria
  Z6852	Unspecified external cause status
  R051	Cough
  R1084	Generalized abdominal pain
  L309	Dermatitis, unspecified
  K5900	Constipation, unspecified
  J09X2	Flu due to ident novel influenza A virus w oth resp manifest
  R519	Headache, unspecified
  R1110	Vomiting, unspecified
  R109	Unspecified abdominal pain
  H6691	Otitis media, unspecified, right ear
  J0190	Acute sinusitis, unspecified

  Overwhelmingly, these codes are "unspecified" or generic ("Cough," "Nasal Congestion," etc.).  For the few that are not - like the otitis - I suspect that AETNA wonders how your otitis routinely got to a 99214.
• Digging into the 2nd-4th diagnoses on the claims is a little tricky from an analysis perspective, but here's what I can tell you: 77% of the downcoded claims had a second ICD-10 on the E&M.  45% had three, 17% had four.  Good sign.  But the most common secondary+ codes?  
  
  

  R09.81	Chest pain, unspecified
  R50.9	Unspecified voice and resonance disorder
  J06.9	Acute epiglottitis with obstruction
  R21	Rash and other nonspecific skin eruption
  H66.92	Unspecified nonsuppurative otitis media, bilateral
  F41.9	Phobic anxiety disorder, unspecified
  R10.9	Foreign body sensation, other site
  R05.9	Cough
  B08.4	Viral wart, unspecified
  R11.10	Flank pain, bilateral
  G47.9	Other vascular syndromes of brain in cerebrovascular diseases
  R19.7	Other ascites

We can and should discuss whether "Cough" and "Chest Pain" warrant a 99214 as a rule (clearly, those can be legitimate 99214 diagnoses, especially when included with a better primary diagnosis). Still, the question is: should we work to avoid this issue simply by using better ICD-10 techniques?

There are lots of unanswered questions.  I haven't thoroughly compared these ICD-10 distributions to claims that weren't downcoded.  I haven't looked to see whether other practices use these unspecified diagnoses to a similar extent without getting dinged.  Subjective analysis indicates that the practices hit hardest by AETNA did not have particularly high 99214 usage (the trigger that CIGNA intimated would be in effect).

Input welcomed.