Variable Prices Per Test for EGFR Testing (Medicare, 2015)

In CY2015, CMS had a flat and fixed fee schedule price for EGFR testing of $329.18 (for AMA CPT code 81235, common variants.)  This gene is commonly tested in lung cancer and some other cancers.

Using the CMS 2015 provider database, a couple labs received up to double the fee schedule price, per EGFR test.   This is one entry point into an unusual policy quirk at Medicare.

• For several years. the MolDX Program has had an online article requesting that labs using Roche Cobas kits for EGFR testing add modifier -22 to claims.
• Then: No public data on actual prices paid were available.
• Now: The various CMS Open Data projects now provide a window into upgraded payments for some labs for CPT code 81235 in 2015 and 2016.
• Extra payments via -22 modifier could help labs beat PAMA fee cuts.
• But: Extra payments (as much as 200%) for some CPT codes on the CLFS, to certain labs, may raise concerns by labs elsewhere that it is difficult to compete on a level playing field (e.g. Mayo in Minnesota; GeneDx in Maryland).

click to enlarge

More details after the break.

Very Brief Blog: CDC Deck on Trends in National BRCA Testing (August 2017)

In August 2017, CDC released an 18 slide deck on US utilization of BRCA testing (through 2014).  Online here.   Around 2007, rates per 100,000 were twice as high in the Northeast as elsewhere although the difference has since narrowed.  The information also appears as a September publication in Genetics in Medicine, here.

US PAMA data published by CMS on September 22 gives a less sophisticated but interesting snapshot of BRCA testing.   Labs reported to CMS they had about 100,000 payments for BRCA full sequencing in 1H2016, suggesting an annualized rate of about 200,00 per year.   The CMS PAMA data misses some labs (including hospital based labs and physician office labs) but might be a fairly good sample of large, independent BRCA-testing labs.   CMS itself pays for circa 20,000 BRCA full sequencing cases per year. 

CMS Publishes Physician Fee Schedule Policies for CY2018 (DPP, Biosimilar, AUC)

On November 2, 2017, CMS published annual rulemaking for PFS policies for CY2017.    The homepage is here.   The final typeset version appeared November 15, 2017, 82 FR 52976, here, 396 pages.

 The CMS homepage with lots of tables and addenda is here.  CMS press release here, fact sheet here.  See also a fairly detailed discussion at Lexology, November 17, here.

Diabetes Prevention Program begins on page 53254 and runs through page 53357, about 100pp or 25% of the annual rule.  This is the first major program that began as an Innovation Center demo and is being expanded by rulemaking into a nationwide benefit.  It's for brick and mortar DPP only.   Questions on future offerings of "virtual DPP" or how MA plans will handle DPP/VDPP are to be clarified in "future guidance." (p. 53235, bottom right.)

Biosimilar guidance has been under discussion again, as CMS proposed to categorize biosimilars in one "bucket" HCPCS code which the biosimilars industry opposed.   CMS will begin categorizing each biosimilar under its own HCPCS code (page 53348ff).

• Under "bucket" coding for biosimilars, CMS would likely have had lower prices, due to drug-on-drug competition to create a profit margin for physicians, by one-upping each other to give the physician more margin under the CMS payment price for the quarter.  
• However, a sufficiently adverse marketplace would mean biosimilars would stand down and not enter the US marketplace at all, a genuine concern.   See RAPS, October 30, here.
• CMS acknowledges that its change to biosimilar-specific coding will "address concerns about a stronger marketplace...encourage innovation to bring more products in the market."
• Note that while CMS sets payment policies only for itself, it also controls all HCPCS codes.  So long as CMS staff had refused to even create unique HCPCS codes, any efforts to get a better market in other payer venues was a non-starter. 
• Two other notes:
• It's unusual that a change this large could be accomplished without a regulatory change.  The requisite regulation, 42 CFR 414.904(j), said that all biosimilars coded together would get an average price.  It did not say whether CMS must, or must not, issue codes that group more than 1 biosimilar together.
• It's unusual in that the proposed policy change was quite short and the final policy change and discussion was very long.  CMS revealed intentions and rationales in the final policy that couldn't be guessed from the July proposal.
• See July proposal, November final, and Regulation in one PDF here.
• For a fascinating article on the early impact of the existing "bucket" pricing rules for biosimilars, see Adam Fein at Drug Channels, 9/2017, here.  See also various links from within that article.  (Like here and here and here.)

Added 1/2018:  Novartis says "US desperately needs biosimilars [at lower prices] to create fiscal space for new drug products," here.

Lab Industry 

CMS discussed comments about the PAMA data collection program for labs at 53581ff.  40 comments were received.  CMS notes drily that "most comments received were out of scope." 

The complex but interesting area of Appropriate Use Criteria requirements, applied to advanced imaging test orders, begins at 53187ff.  This is an invention of PAMA that is still being implemented.  I have always wondered if the same idea could show up some year in context of molecular tests.  Except for PET scans, which are around $1000, most advanced imaging (PET, CT) is $300-600 at CMS prices.   This is easily in the range of many genetic tests and below the range of many proprietary MAAA tests.

Tidbits

For an article on some revisions to Telemedicine services, here.  Also remote monitoring cardiology testing, 53063.   Flow cytometry RVUs, discussed at 53059; pathology consultation 53060; tumor immunohistochemistry, 53061.  Discussion of using PAMA Section 220 authority to collect data on real world pricing, using independent contractors, 52998 (bottom right).

Very Brief Blog: How Will MolDx Handle the Transition to Single Price Tier Two Codes?

UPDATE:  BY DECEMBER 2017, MOLDX HAD TAKEN DOWN 

ITS SPECIAL TIER 2 GENE BY GENE PRICES.   

IT WILL PROBABLY USE 

FLAT PAMA PRICES FOR TIER TWO IN 2018.  

Update here.

When the new genetic test coding system was created in CPT Year 2012, and first used by CMS in CY2013 for gapfilling finalized at the end of 2013,  a new coding concept called Tier 2 codes were introduced.  These are CPT codes for nine "levels" of analysis complexity.   Each CPT code in the series 81400,81401, 81402...81408 has a long list of gene names associated with it.  Work levels progress from one exon, to two exons, to three or four exons, and so on. 

MolDx provides a table of meticulously determined individual prices for each single one of the covered genes under Tier 2.  (Most genes listed in the AMA Tier 2 set are not covered by MolDx).   See the June, 2017 version of this fee schedule here or in the cloud here.  Like snowflakes, almost all are different (there are 7 pairs of matching prices among 49 genes, or 42 prices for 49 genes.)

Through PAMA, CMS has set rates for each Tier 2 code to be effective on January 1, 2018.  The second-to-right column, "Pmt", is the new 2018 PAMA rate.   The "Ave Pmt" column is the observed average payment rate at CMS during 2016.

Here's a brain teaser for MolDx watchers.   What will MolDx do when these fixed fee rates become effective?

1. MolDx trashes its elaborate special Tier 2 fee schedule series and just uses the new Tier 2 codes.   
1. This is the obvious choice; it's what non-MolDx MACs will do.
2. This seems best aligned with the MolDX statement of work, which is only to price codes which "do not have CLFS prices," and the Tier 2 codes will have CLFS prices.  SOW here.
2. MolDx could require the Tier 2 codes to be used with "reduced services" or "enhanced services" modifiers tagged to Z codes and continue to offer special pricing lower or higher than the official Tier 2 code price.   
3. MolDX could ...not recognize the Tier 2 codes and throw all the services onto unlisted codes matched with Z codes.
1. This is the "make crazy" option but they have been very liberal with local usage of Unlisted Code in the past.

I think the answer will be #1, and the snowflake-like pricing of Tier 2 codes will be tossed overboard, but if it is #2 or #3 you heard it here first.

CMS Posts State Level Payment Data for Part B for CY2016: MoPath

CMS annually posts both national payment data by CPT code (here) and by state (here). 

Specific to genomics, I've done an initial analysis of 2016 national payment data here.

It's difficult to analyze the state payment data since there are over 50 files.   However, I've pulled the MoPath code series for Northern California, Southern California, and Utah into one spreadsheet, because these three areas total the lion's share of US MoPath payments.   In the cloud, here.

2016 Data by State: BRCA as a Case Study

       Update:  I've published expanded data and charts in March 2018, as a survey that reviews 2014 2015 2016 together,  here.

While it would take a long time to assess all the MoPath codes by all the states, I've pulled just the BRCA related codes, here, as a case study:

Click to enlarge

Let's analyze as BRCA1/BRCA2 testing, with or without a larger panel of genes.   In Northern California, CMS paid 2,963 claims for 81432 (zero claims for sister code 81433) totaling $2.7M, at an average price per claim circa $1000.    In Southern California, in contrast, almost all claims came in as code 81162 (the consolidated BRCA1-2 sequencing and DupDel code), with payments of about $18M for only twice as many patients.

Turning to Utah, claims came in variably under code 81162 and 81211 at about a 2:1 ratio.   It's not possible to know how many patients are represented by the code pairs 81211+81213;  for this table, I assume there were 3,613 unique patients.  With that assumption, there were about 12,000 patients getting BRCA sequencing in Utah with payments of $31M and the average price just a little higher than in Southern California. 

In round numbers -- with some assumptions -- if all the patients getting BRCA1-2 sequencing (with or without a larger panel)  had been paid at the average Northern California rate, total CMS payments would be around $23M.  If all the patients had been paid at the Southern California or Utah rates, total payments would have been closer to $58M. 

Medicare has published provider specific payment data for up to CY2015 (here).  As of CY2015, there was only one 81211 provider in Utah, although that may not be the case in 2016.  As for SoCal, that's a puzzle.  There were only a few dozen payments for 81211 in SoCal in CY2015, but over 7,000 payments for 81211 in SoCal in CY2016.   At the same time, payments for unlisted code 81479 in SoCal dropped from $62M in CY2015 (of which $31M to Ambry Genetics) to $51M in CY2016.

Can CMS Do Anything?

CMS has long applied a panel pricing rule in lab medicine, so that components of a panel can never get paid more than the total panel.  CMS could apply that to BRCA testing.   If it was a leg fix, it looks like it would score $35M savings per year or (as a first guess) $350M savings over 10 years. 

Bonus: Pie Chart

As best as I can reconstruct, this is a pie chart of the five different ways to code BRCA testing at five different net prices, using CMS 2016 data.  2016 was the first year that comprehensive code 81162 was available, as well as gene panel codes 81432, 81433.   I've shown earlier, in PAMA data for commercial payers in 1H2016, the use of 81162 was around 1%.   The "blue slice," 81162, paid about $2500; the "orange slice," 81213+81211, paid about $2700 but required not employing a published CMS CCI edit against code stacking 81213+81211.   The yellow slice (81432+81433) pays about $1500 but not in California, where a MolDX edit classifes 81433 as an "excluded" code (here) and payment is for 81432 alone at $932 (blue slice).

click to enlarge

Quorum Consulting Acquired by Navigant (NYSE)

San Francisco's leading boutique pharma and healthcare consultancy, Quorum Consulting, will be acquired by Navigant for an undisclosed amount.   Navigant is traded on the New York Stock Exchange with a market cap of about $800M.  2016 revenue was about $1B.  Press release here.

In August 2015, Avalere, another life sciences/pharmacoeconomics consultancy was acquired by Inovalon for $140M.

Truven health data analytics was acquired by IBM in February 2016 for $2.6B.

In November 2017, Optum completed its acquisition of Advisory Board.  The deal was valued at $1.3B. 

In other news, longstanding Silicon Valley pharmacoeconomics consultancy Cedar Associates has apparently taken down its website.    John Hornberger MD is now listed as Senior Fellow at Genomic Health on LinkedIn as of October 2017.

CMS Publishes CY2018 OPPS Rule; Loosens Bundling on Hospital-Outpatient ADLT Tests

HOW TO COMMENT ON 14 DAY RULE ERRORS:  SEE NOVEMBER 21 BLOG HERE.

_________

On November 1, 2018, CMS released the "public inspection" version of the CY2018 Hospital Outpatient rulemaking.   The rule is 282 pages long.  See final publication here, 82 Fed Reg 52356, November 13, 2017.   See discussion at Lexology, here.

For for addenda documentation, here.  CMS press release here, fact sheet here.

 For the complementary rulemaking, the 2018 Physician Fee Schedule, see here.

This blog focuses on changes to the 14 Day Rule in diagnostics.  For a general article on the 2018 rule, see HealthcareDive, here.  Changes to 340B and more.

Note:  An early version of this post said the rule applies only to ADLTs.  Nope.  It applies to Type A, not Type B ADLTs, and it applies to "molecular pathology tests currently excluded from bundling" but not (confusingly) to genetic tests in the GSP series.  

All this is: if I read the rule correctly, or if the CMS rule really says what CMS intended.  There may be errors in the CMS text.  

For example, in the 2018 OPPS excel spreadsheet called "Addendum B," which have payment instructions for each code, all regular lab tests are "N" - not separately paid - while genetic/genomic tests are generally A, including "GSPs," but in the MAAA series, tests are erratically code as status N, E1, or A.  Net-net, I think CMS misspoke when it writes the rule doesn't apply to GSPs.  I would think it will have to go through PLA codes and at least allow payable status to genetic ones, when it comes to its senses.

CMS did not discuss exclusion of PLA or GSP tests in July proposed rulemaking!

For Genomeweb's effort to cover the rulemaking, here (subscription).

The rule loosens Medicare's bundling rules, put in place in 2007, on hospital outpatient specimens.  The discussion is lengthy and runs from page 707 to 737. The logic and rules are a little complicated but here goes:

• Since 2007, hospital outpatient lab tests can only be billed by the hospital, even if they are sent to an outside reference laboratory.   This rule is in place holds for 14 days after the date of tissue or blood procurement in a hospital outpatient center.  In essence, any test performed is "a service of the hospital" and must be billed by the hospital.   
• For example, a Genomic Health Oncotype DX test could (I'm simplifying a bit) be billed only be the hospital if performed within 14 days of an outpatient lumpectomy, but if performed +14 days on the paraffin block, it could be billed by Genomic Health.
• Since about 2013, CMS has bundled all clinical laboratory tests to the hospital outpatient event (e.g. lumpectomy, office visit, etc) - except human genetic/genomic tests.   These latter were still billable by the hospital on a line item basis, or, if later than 14 days, by the performing reference lab.
• THE NEW RULE allows Advanced Diagnostic Laboratory Tests, or ADLTs, to be billed Medicare directly by the laboratory, when run on hospital outpatient specimens in the 0-14 day range, as well as "molecular pathology tests currently excluded from bundling."
• Confusingly, the rule appears to exclude Genomic Sequencing Procedures, which are certainly human molecular pathology tests."  
• The rule also appears to exclude PLA tests which in some cases will be genetic tests that are closely equivalent to genetic tests in the genetic code series.

ADLTs are defined under PAMA law, and later CMS policymaking, as sole source tests that are either (A) MAAA type tests or (B) FDA approved or cleared.  CMS states, confusingly, that the new rule applies only to "TYPE A" ADLT tests.

CMS Still Hasn't Certified Its First ADLT Test

CMS loaded some extra policy rules such as the test was independently developed by the laboratory and not licensed.   (I believe these latter rules will prove to be tortuous to apply or may be applied inconsistently.)  And that the test provides "unique information."   (Will CMS view several prostate MAAA tests as providing a unique class of information?   Will something cease to be an ADLT if someone produces a close copy?  Etc.)    CMS notes that tests released from the 14 day rule because they are ADLTs must be certified by CMS as ADLTs, a process I believe is not in place yet.

Is This an Extra Incentive to Fill Out ADLT Paperwork on Your ADLT?

There has been some question whether well-accepted popular MAAA tests that would qualify as ADLTs need to be certified as ADLTs, because it would convert them from 3-year to 1-year pricing updates, which might be unhelpful to the lab.   This new rule provides one incentive to go through the extra work to be certified as an ADLT, because it gives access to the 14 day rule exemption. 

The policymaking applies to hospital outpatient specimens, but not to hospital inpatient specimens.  

   

OVERVIEW OF CMS MOLECULAR PATHOLOGY UTILIZATION IN CY2016

On October 19, 2017, CMS released its CY2016 Part B national utilization files - here.

I've looked at the Molecular Pathology codes, from 81162 to 81599.   (For my CY2015 report, see here.  For my CY2014 report, see here.)

Total Utilization of MoPath Codes

Culling the codes down to the range 81162-81599, I tally $479,894,633 in "payment allowed", which I use throughout this report.  (CMS "payment issued" is slightly less due to deductibles, etc.)   This code series has about 168 codes in the CMS CLFS CPT. 

In the Genomic Sequencing Procedure Codes, there is a "Valley of Death" with no payments on any of the codes 81410-81425, which are 8 codes including some neonatal codes, exome, and genome.

High Payment Codes

The top ten codes paid $387,452,107.   This is 79% of all payments ($387M/$490M).  This goes beyond the 80/20 rule; this is the 80/5 rule (5% of all codes get 80% of all payments).

The highest payment line was Unlisted Code 81479 at $108M.   No wonder CMS has asked its MACs to start reporting on what companies are using unlisted codes for what tests - I predict CMS will issue G codes or require PLA codes for larger volume unlisted codes.   Based on state level data for CY2016, Assurex was probably paid for about 13,000 81479 services at about $2175 each (here).

However, some of the unlisted code volume is driven by special MOLDX rules - for example, we noted that Ambry Diagnostics had about 80% use of unlisted codes in CY2015.  Ambry primarily provides known gene sequencing, rather than proprietary tests like MAAAs, letting us infer that MolDX rules in California probably drove Ambry's high 81479 number.  Over 99% of 2015 81479 payments were in MolDX jurisdictions.

After the unlisted codes at code #1, Cologuard was runner up at #2, with $63M in payments, and Oncotype DX with $61M at #3.  The full table is here:

click to enlarge

Note that only 1 CYP gene makes the chart (and barely sequeaks in at $12M), due to stringent LCDs on CYP testing Medicare rolled out over the past 3 years.  Two CYP codes, 81225, 81226, garnered a colossal $270M payments back in CY2014.

Top 10 Codes...Other 150 Codes = Long Thin Tail

This pie chart shows ALL Mopath spending $470M), with the first ten slices individually and the last 20% as the last large slice.  Note that in this pie chart, and have added the two BRCA coding methods together into one pink slice (81162 & 81211).  After the initial light blue slice, Unlisted Code 81479 (about 99% found in MolDX states), the "fall colors" are all MAAA tests.  (Much of the light blue slice is also MAAA tests under unlisted code, e.g. Genesight etc.  A best-guess would be that the circle from about 1:30 to about 8:00 is all MAAA tests).

click to enlarge

High Service Volume Codes

 The top 10 codes by unit volume had 511K services, or 61% of the total 838K services.  These top ten most heavily utilized tests had $256M in payments, or 52% of payments.

The most frequent test was Cologuard, at 123,768 services.  (339 Medicare patients per day sent their little package to Wisconsin.)    The second most frequent was Unlisted Code 81479, with 73,573 services, followed by Tier 2 code 81401 at 58,478 services. 

The full table is here:

click to enlarge

Tier 2 Codes

CPT's quirky Tier 2 codes tallied $19,936,026 in payments, or just 4% of CMS payments.  Tier 1 81400 had 31K uses while Tier 2 81401 had 58K uses.   In total there were 125,000 uses of Tier 2 codes, but only a couple hundred for the two highest levels.    "Ave Pmt" is observed average 2016 payment; while "Pmt" is the predicted 2018 PAMA payment.  The full Tier 2 table is here:

click to enlarge

The Tier 2 codes were contractor priced until now, but will have PAMA prices in CY2018.   The PAMA prices generally don't diverge too much from the average MAC prices under the local pricing system.   MOLDX has elaborate prices for different tests within each Tier 2 level.  These may be replaced (sunsetted) by one-size-fits-all pricing when the Tier 2 CLFS prices become law on 1/1/2018.

Tumor Panel Codes Perk Up A Little

Tumor panel codes had virtually no usage in 2015 in CMS data, but pick up a bit in 2016.  There were 3,585 uses of 81445, 3,366 uses of 81450, and 1,062 uses of 81455.   These are respectively codes for 5-50 somatic tumor genes, 5-50 hematopoietic tumor genes, and >51 tumor genes of any type.

MAAA Tests

The MAAA code series 81490-81595 tallied $215M, or about about 45% of all mopath spending.

Medicare BRCA Data Diverges from Private Payer (PAMA) BRCA Data

A few weeks ago, CMS released its 5M line database of private payer lab test price data for 1Y2016, collected as part of the new PAMA price setting process.   For BRCA testing, which has a very complex coding system, at least in 1H2016, private payer use of CPT Code 81162, a comprehensive BRCA testing code, was almost nil (about 1000 cases out of 200,000 cases).   

However, CMS data is more diverse.  There were 16,665 cases of 81162, for $41M in payments.  It's used for more than half of CMS CY2016 payments, while it was closer to 1% of private payer CY2016 payments.  This could be because private payers had based contracting running across 2015 and 2016 on existing codes (e.g. 81211) whereas CMS has no long term contracts, and correct coding rolls over and kicks in anew with each January 1st. 

But quite a few other labs and MACs were still very actively using an unbundling or code stacking approach under Medicare, billing about 9,000-10,000 cases of 81211 (BRCA sequencing) and also 81213 (BRCA uncommon dup del).   In CY2016, the difference between coding 81162 (about $2500) and 81211+81213 (about $2700) was small.  However, the target prices for these codes widen under PAMA (to $1600 vs $2900 (!!)), making the choice of coding in the future more impactful.

There is also a hereditary breast cancer panel codeset:  81432, $932, for 14 or more BRCA related genes, and 81433, $600, for the Dup Del analysis.   CMS paid for 4,336 cases of 81432, but only 900 cases of 81433.   This is likely because of a MOLDX edit that classifies 81433 as an "excluded" code.*   This issue becomes more important as PAMA takes hold, because the price of 81432 will drop at 10% a year in the next three years.   Of course, it also makes no logical sense that BRCA1+BRCA2 analysis pay $2900 while that full service fits inside of a panel that pays in total $900 or less.  Two eggs for $20, a dozen eggs for $2.  You can't explain that to your sixth grader.   

___

MolDX document M0047 V15.  MOLDX classifies 81433 (breast cancer panel dup del) as an "excluded gene" but does NOT classify the parallel code 81213 (breast cancer gene dup del) as an excluded gene.  Data shows that this drives payment to California labs doing BRCA testing to about $900 whereas in some other states Medicare pays about $2500-2700 for BRCA testing.  CMS has created a webpage for CY2016 State-Level payment data, but the files are not active yet. CMS will eventually post provider-level data by CPT code for CY2016, but not until about June 2018.

In short, we're really seeing four ways to price BRCA testing at CMS:  81213+81211 (most costly); 81262, which CMS seems to use more than privates; 81432+81433; and 81432 alone (where 81433 is banned-by-MolDX).   This results in BRCA pricing, in real 2016 data, from around $900 to $2700 or 3X.

Very Brief Blog: CMS Releases National 2016 Part B Data

CMS releases three kinds of national Part B data:  First, national utilization for the prior year, typically in November.  Second, utilization by "carrier" (basically, by state) for the prior year, also typically in November.  Third, after an 18 month delay, CMS releases data by CPT code by individual provider by name.  So 2016 data appears in about June 2018.

CMS has released CY2016 national data, here.  I've also put in the cloud, here.  This is a zip file folder with many small files with sequential series of CPT codes and their national utilization and total national spend.   A window into some CY2016 lab data appeared a few weeks ago in an OIG report.  Now you can see all the CY2016 national lab data.   

CMS has set up the webpage for state specific CY2016, but the links don't work yet.  Here.   

For completeness, eventually CY2016 provider specific data will appear here, but not til about June 2018.  I've also summarized the data sources in one blog, here.

____

The molecular codes from 81162 to 81599 inclusively tally about $479M.   For a breakdown, see my blog entry here. 

Very Brief Blog: Digital Therapeutics Coalition Formed

According to an article in MobiHealthNews, a new coalition of startups in the "digital therapeutics" space has formed.   This is another example of how digital is transforming healthcare delivery in ways that go far beyond EHRs or telemedicine, which are represented by older trade groups.

The article is online here.   Sean Duffy, CEO of Omada, is prominently quoted.

Recently I noted that the CHRONIC care act has bumped around the Hill for several years, and includes a section improving telemedicine benefits and how they are paid/accounted for in Medicare Advantage plans.   That section of the bill will save Medicare $80M a year (!!!), as scored by the conservative CBO, so there's absolutely no reason it can't pass.  Still, it doesn't get done.   Hopefully, the Digital Therapeutics Alliance (DTA) will help the policy system get unstuck.

   

Very Brief Blog: Talk on Diabetes Insulin Pumps & Medicare Policy (Intl Diab Tech Mtg, Bethesda, Nov. 2, 2017)

I'm giving a talk on Thursday, November 2, 2017, in Bethesda at the Diabetes Technology Society international annual meeting.   The conference website is here. 

My presentation is in a series on reimbursement, from continuous glucose meters, to diabetes telemedicine and prevention, to artificial pancreas.   I'm posting the deck I will present, on insulin pump policy: here.

Some key points:

• Medicare is a defined benefit program, and insulin pumps and their insulin are under the benefit for DME (including wheelchairs, etc). 
• There's one legacy pump code, E0784, priced at about $5400 regardless of brand or pump capabilities.
• You have to be doing demonstrably badly enough on insulin before Medicare covers a pump.  
• Due to 2003 rules, insulin pricing for pump-based insulin had gotten badly out of whack with the current market, only about 30% of market price.  However, it was reset to market price on 1/1/2017.
• The GAO did a report strongly supportive innovation for new technology in the DME benefit.  CMS's reaction was lukewarm or less.   Here.
• CMS rejected a special new code for artificial pancreas (e.g. very smart insulin pumps) in 2014, but is reconsidering now, with a decision expected in November 2017.

NEHI Holds Webinar on KYMRIAH Pricing; Releases White Paper on Value Based Oncology Drug Pricing

On October 24, 2017, NEHI released a 33 page white paper on new approaches to value-based oncology drug pricing.  And on October 26, NEHI held a webinar on the topic, featuring speakers from Novartis and Harvard Pilgrim and focusing on outcomes-based billing and pricing for KYMRIAH (list price $475,000). 

See also: value based white paper from Duke with Health Affairs essay; links at bottom.

CMS Still In Background
The presentation doesn't feature any speakers from CMS/CMMI, who told NEHI they weren't ready to publicly discuss the in-progress indication-based and outcome-based pricing model.   I had earlier noted that while CMS and Novartis released "joint press releases" in early September about KYMRIAH, many journalists missed the fact that the CMS statement said next to nothing except for some vague, general positioning about hoping to support biopharma innovation (here, here).  It was noted there is nothing yet about the project on the CMMI website.

Indication Based J Codes?
According to Novartis in the new webinar, CMS will create different J codes for different indication and seems to have agreed with Novartis about only billing for successful patients.   As I pointed out in September, details have to be hammered out since hospitals, not biopharmas, generally bill CMS for drugs, and CMS most commonly sets Medicare policy centrally while leaving many aspects of Medicaid policy to the states.  The question of hospital markups (e.g. billing $950,000 for $475,000 of Kymriah) similarly arose a couple times in the webinar.

Links

The webpage for the white paper and presentation is here.  You'll find one link for the PDF and one link that takes you to the video on Vimeo.  For more details on the sixty minute presentation, here.

In addition to this October 24 white paper on oncology drugs, NEHI has a home page for all its white papers on price innovation initiatives, here.

See also: Duke & Health Affairs, October 30, 2017.

For an essay from the Margolis Center at Duke, by Daniel, Leschly, Marrazzo, and McClellan titled "Advancing Gene Therapies and Curative Health through Value Based Payment Reform."  Here.  Includes links to the Duke-based "Value Based Payment Consortium," here.  

Like NEHI, this center also has an October 2017 white paper on the path to Value Based Payment for medical products (23pp), here.

The National Pharmaceutical Council's 2016 white paper on this topic (15pp) is here.

___
While the original approval of Kymriah was very an extremely rare pediatric condition, Novartis has filed for FDA approval for diffuse large B cell lymphoma in adults, a much more common condtion (here).

A Visual View of The Weird Intersection of BRCA Coding and PAMA Pricing

In an earlier blog we presented lab-specific CMS payment data for BRCA testing (here) and how to explain CMS's topsy-turvy BRCA pricing to a sixth-grader (here).

PAMA and the various combinations of 8 BRCA-related codes makes it even weirder than it used to be. 

I've assembled the BRCA services and coding options in the picture below, along with PAMA data for utilization and newly released PAMA median prices.  (Note, I've doubled the PAMA utilization to extrapolate from the sixth-month survey to a 12-month year.)  Click to enlarge, review, then continue.

click to enlarge

As observers have noted, under PAMA, the price of the most common BRCA billing code, 81211, actually rises somewhat (up by $200 to $2395) while the price of BRCA uncommon Dup Del analysis stays about the same, code 81213.   The consolidated or comprehensive BRCA billing code 81162 falls from $2503 to $1615, so it may be a less popular option than it was before. 

But wait, there's a whole 'nother level of weirdness, as the pricing for single-gene BRCA1 sequencing plummets to $75 and single-gene BRCA2 sequencing plummets to $185.   (Note that these codes got only a couple dozen uses in the PAMA year, whereas BRCA1-2 sequencing was about 200,000 uses.) 

This means there are several ways to build up coding for exactly the same major BRCA services: "Full sequencing of BRCA1 and BRCA2 and analysis of both common and uncommon deletions."

The most code-efficient way (which seems like correct coding to my eye) is 81162, paying $1615.   However, look harder and you'll find the highest sum comes from billing as BRCA 1-2 sequencing (81211, $2395) and uncommon deletion analysis (81213, +$553), totaling an impressive $2948, almost double the value we get for using the parent code 81162. 

But look closely and there's a third option, to order BRCA1 sequencing alone, with its common dup-del analysis, 81214, a frugal $75, and then order BRCA2 sequencing, 81216, for a still budget-friendly $185.  You're almost done: just order 81213 to wrap up those uncommon DupDels ($553).   Your shopping cart gives you all those BRCA services, soup to nuts, and costs you only $813, half the price of 81162, and only about a quarter of the most costly coding option for the same sequencing reports.  It's like those sales that get you one-way to Europe for $250.

To my eye, the simple coding is 81162.  I've said that.   But if you're going to break up 81162 into 81211+81213 (if you're gonna say that is kosher coding), why not do it again, and break up the 81211 into its component codes of 81214+81216?

More About BRCA Coding: Quest, AMA CPT, NCCI, and 3 Year Spending Options

Quest indicates that BRCA coding is ambiguous, allowing either 81211+81213 unbundled stack coding or use of the comprehensive code 81162.  See the Quest website here (sampled 10/30/2017; screen shot below).

There's no question that some of BRCA coding proposal was "tabled" by the powers that be when it reached the September 2017 AMA CPT meeting:

There's also no question that there is a public, national CMS NCCI Edit against unbundling and code stacking of 81211+81213, although it's not applied in some MACs (see here):

The next table shows CMS spending on BRCA testing in 2018, 2019, 2020 under three different coding approaches.   I've assumed 20,000 cases per year, about the 2016 rate.  Click to enlarge.

click to enlarge

CMS mostly pays for BRCA testing the most costly of all possible ways, by code stacking 81211+81213, despite CCI instructions not to pay this way.   This means the upcoming three year spend will be $176,930,400.   However, if CMS only enforced its own CCI edit, already on the books, and required use of comprehensive code 81162, spending would fall to $122,109,000, or a savings of $54,821,400. 

This is a big deal.  There's never supposed to be multiple ways to code the same service with AMA CPT codes.   (And I haven't even raised the issue of submitting gene panel AMA CPT codes where they are a better fit than code stacking individual genes.)

It's easy to imagine that CMS could simply throw up its hands at the brouhaha over AMA instructions and ambiguous use of 81211 and 81162, and many multiple coding methods.  CMS  could deal with the AMA problem and remove these two codes easily from the CMS fee schedule, because they are redundant.  Then CMS providers would not have comprehensive coding options, but, providers could account for full BRCA sequencing and full DUP DEL analysis by coding 81214+81216+81213, for total CMS payments of only $114,824,200, or CMS savings of $62,106,200 in three years.

HHS Posts 17 Internally Generated Innovation Projects

In the last few weeks, the head of CMS has written a WSJ op ed calling for a reboot of the CMS Innovation Center (here).  And the new head of AHRQ has promised a vision for a new "AHRQ 2.0" - here.

Here's another approach to innovation at HHS.   See a set of 14 innovation programs approved by HHS for the coming year.   This is an initiative of the HHS IDEA Lab (here).   The IDEA Lab fielded 79 proposes from across HHS - CDC, NIH, FDA, CMS, etc.   The proposals were rigorously reviewed and scored, and 14 were chosen.

See the list of proposals here.   One directly involves CMS, and happens to be in the laboratory field, headed by Katherine Szarama PhD of the Coverage Group, Eunice Lee of FDA, Brandi Kattman, NIH.   The project promises to reduce wait time between NIH, FDA, and CMS decision points (see full abstract at link). 

Another CMS project looks at the issue of hospital readmissions in sickle cell disease.

Very Brief Blog: New Leader Promises New Direction at AHRQ

On October 17, 2017, the new leader of the Agency for Healthcare Research and Quality (AHRQ) published a detailed blog on his plans for innovating and rebooting the agency.

The blog is here.   Open access coverage at Fierce Healthcare is here.  The new director, Gopal Khanna MBA,  headed a project called "FRAMEWORK" at Illinois HHS, part of an innovation incubator. 

Khanna has a lengthy Wikipedia page (here; cloud here; LinkedIn here).   He took the leadership role at AHRQ in May, 2017.

Khanna's four talking points are:

1. More development and integration of data platforms - Digital Health meets AHRQ.
2. Catalyze "Learning Healthcare Systems."
3. Engage operational leaders; convert innovation into practice & operations.
4. Patient-centered care.  Integrate federal, state, community, and other levels.

At CMS, CMMI is also seeking a reboot, a sort of vision for CMMI 2.0 (here).

AHRQ and USPSTF

AHRQ also supports the US Preventive Services Task Force (USPSTF), whose recommendations affect required health insurance preventive services (here).   

USPSTF released updated cervical cancer screening guidelines in draft form recently, and is expected to release new guidelines for BRCA screening in the coming weeks.  To date, USPSTF has not dealt (a) with BRCA screening in women not at elevated risk or (b) breast cancer related gene panel screening, which has become common.