Friday, June 16, 2017

Brief Blog: CMS Molecular Payment Data CY2015

As noted in earlier blogs, CMS posts Part B data by provider name and CPT code for 2012, 2013, 2014, 2015.  The data can be displayed in a web interface that is very easy to sort and then download filtered data in Excel.

Although I'd usually post the Excel file in the cloud for you, it feels awkward to do so, because even though the data is publicly available, it contains hundreds of physician names, addresses, and NPI numbers.  However, you can easily find the data yourself or download it yourself [*].

  • You get a file of several hundred KB and about 2,000 lines by filtering the total database for codes beginning in 812, 813, 814, 815.  (If you're really brave, filter for 8nnnn - all lab and pathology codes - and you'll get a 100 mb dataset of 850,000 lines which does still open in Excel for me on a laptop, albeit slowly.)
  Quick View of 2015 MoPath Data
Biggest CPT Code by Provider
The highest claim line was $57M to Genomic Health for Oncotype DX.

Lines Required to Reach 50% of Payments
The second highest was $32M to Ambry Genetics for "unlisted code."  Just 12 claim lines out of 2000 total claim lines reached $240M or 50% of all payments for the year.

$1M or Higher
62 claim lines were at least $1M or higher (78% of all payments).  An earlier blog noted that 1/3 of all payments went through the unlisted code 81479, and 99% of that fell in MolDX states.

Click to enlarge:

More after the break.

California Versus New York
No contest.  California had 57% of all CMS mopath payments, whereas New York had only 0.3%.   New York State has a unique governmental review apparatus, almost like a mini FDA, for genetic test approval before public offering to people in NY State (here).

No Use of Tumor Panel Codes 
Almost none of the $480M went through the somatic tumor codes 81445, 81450, 81455 (less than $150,000).

Biggest Billers for Tier 2 Codes
Once you have the Excel, another simple trick is to sort by (a) code and (b) annual payment.  The biggest provider of 81400, Tier 2 Level 1, was Alpha Genomix in Georgia (16483 services, $2.0M). They were also the biggest provider of 81401, Tier 2 Level 2, (16277 services, $2.3M).  

The top Tier 2 codes, 81407 and 81408 were paid out only to Tennessee pathology groups, with about $400,000 in payments for about 300 services in these Level 8 and Level 9 Tier 2 codes.  The Cahaba MAC paid a Dr. H. for exactly 140 services of 81407 and for exactly 140 services of 81408.

There were about 231 total claim lines for Tier 2 codes, with $22M of payments.  Only 15%, or $3.2M, of Tier 2 payments went through California.

By Code Sequence Set
The gene sequencing codes 81200-81355 totaled $31,709,096.  The next set, HLA codes, totaled $3,497,177.  At 81400-408, the Tier 2 codes totaled $22,079,550.  81479 is next at $166,616,219.  The 81500 MAAA series was $57,499,957.  The GSP series 81410-71 was just being introduced, and was $510,507.

You Can Get All USA Laboratory & Pathology Code Data
I also tried pulling down all 8xxxx series codes.  This initially gave a 100MB file, which opened in Excel.  I deleted superfluous columns like first name and street address.  It then saved as a workable Excel spreadsheet of 70 MB and 880,000 lines, which I was able to open and run (albeit slowly) on a Windows 10 laptop in Excel.

[*]  How to filter for genetic test codes.

At the newly released CMS website for 2015 data, here, click on "ADD A NEW FILTER CONDITION" and then SEE "FILTER" dialog box, and then filter forthe field-name HCPCS CODE (drop down menu) with the method "STARTS WITH" (drop down menu) and starting with with any of these: 812, 813, 814, or 815. See picture.

Having done that and hit return, click the pale blue button EXPORT and download as an Excel CSV file.  (I recommended immediately opening and resaving as an Excel Workbook).  If you've done what I did, it's a file of about 2000 lines.

Finally, I recommended creating one column that CMS doesn't give you, which is services allowed * dollars allowed per service.  This gives you total payment for that provider by CPT code for 2015.  Data can now easily be sorted by dollar volume, provider, state, CPT code, etc.

I also recommended highlighting and deleting excess rows you don't need, for example, provider first name, address, etc.