Friday, May 25, 2018

CMS Posts Final Expanded Code List for June 25 Crosswalk/Gapfill Meeting

In April, CMS announced June 25, 2018, as the date for the summer public comment crosswalk-gapfill meeting for 2019 CPT lab codes.  May 24 should be the final update of the code list to be debated, as the list should be finalized 30 days ahead of the June 25 meeting.

At this webpage, scroll down for Code Lists and choose the one Updated May 24, 2018.   (The direct link straight to the zip file should be here.)

There are an eye popping 100 agenda items now, including a set of PLA codes released by AMA on June 1, through 0061U.   The update adds about 18 PLA codes (agenda 83-100).

  • Registration is required by June 11 via a registration site.   
  • If submitting a presentation, it has the same due date, but you submit to an email, not the registration site.   
  • If you have someone in your party who is a FOREIGN NATIONAL, contact CMS staff far in advance as there are special procedures to follow.   

CMS will hold a two day meeting of its lab test advisory panel July 16-17 by national webinar.

For more details see my earlier post.

Wednesday, May 23, 2018

CMS Releases Provider & Lab Specific Utilization by CPT Code for 2016

Beginning in 2014, CMS has annually released comprehensive data of Medicare payments to each physician and lab in the U.S. based on CPT codes. 

On May 23, 2018, CMS released this data for CY2016 for Part B data.  Data for each year from CY2012 to CY2016 are available.

  • See the CMS database webpage here; click on CY 2016.

Or on this webpage, click on "Interactive Dataset."  Finally, get to a webpage where you click "View Data" box to see the full dataset in a web view.  You can filter and view data on line, or filter to a subset of data and then export (download) in Excel.
  • As a sample view into the database, I filtered for labs using any of these codes (81162, 81211, 81213).  81162 is the comprehensive code for BRCA 1&2&DupDel analysis.  81211 is BRCA 1&2 sequencing and 81213 is BRCA 1&2 DupDel analysis only.  (Note: These codes are being markedly revamped for the CY2019 codebooks).  
    • Total Part B BRCA spend was $67M.
  • Alternately, you could filter for the whole code range of Mopath CPT codes, which gives a 1500 line Excel spreadsheet (180 kb).  
    • Total Part B Mopath spend was $463M.
In the cloud:
  • See an Excel spreadsheet of this filtered BRCA data in the cloud, here.
  • See an Excel spreadsheet of all MoPath data in the cloud, here.
  • See a 100MB Excel spreadsheet of all Lab Code data in the cloud, here.
  •    For the latter, Google allows download but probably not view.


Total payments for these BRCA codes was $67M.   Total payments to Myriad Genetics were $32M, or 47% of the Medicare total BRCA payments.  The data is granular: for example, there were exactly 4,218 payments to Myriad for code 81213.

click to enlarge

CMS did an elaborate NCD for NGS testing in cancer last winter, finalizing it in March.  For CY2016, there were only 6,889 services for gene panel testing under codes 81445, '50, '55.  Payments totalled $4.0M, or less than 1% of CMS MoPath payments.   Of that, almost 75% went to only two payment lines.   These were Caris 81445 ($792,000) and Genoptix 81450 ($2.1M). 

In addition, Foundation Medicine was paid under different NPIs for its Cambridge MA and Morrisville NC labs.  While it billed a few small codes (81275, Kras, $78,000) from Cambridge MA,  its payments were largest for 81479, $3146, for 622 units of service, totalling $2,124,752, billed in North Carolina.   Added with the 81445-55 data, this brings 2016 Medicare payments for gene panel testing in tumors to only 7500 uses among circa 30M patients in Medicare Part B.


All Mopath codes tallied $463M. 

Exact Sciences Cologuard and Genomic Health Oncotype DX Breast were nearly tied for the top position, at $62M and $60M respectively.  Exact was paid for 123,729 patients.  These two tests at these two companies garnered about 25% of all Medicare mopath payments. 

Next, Crescendo Vectra test had 49,702 payments for $29M.  Assurex test was paid via the "81479" unlisted code, with 13,062 payments for $2,181, totalling $28M.  

12 codes were paid $10M or more, and these 12 codes garnered 65% of all payments for MoPath.

click to enlarge

Download the 100 Mb spreadsheet if you like.  This is all CPT codes in the 80,000 series X all submitting providers.  The spreadsheet has about 90,000 rows.   I tally payments at $5.8B. 

80053, Comprehensive Metabolic Panel, had the most billings one code at one provider location, at 1,319,197 services provided by Labcorp, with a submitted charge of $45.26 and average payment of $9.10.

I get the highest average "submitted charge" for a lab test as $15,759.52 for code 81445 from Caris.  CMS average payment was $430.55.   Castle Biosciences was paid $1.3M for 330 cases of 84999, submitted charge $8024 and payment $4100.

Very Brief Blog: IJRO Issue Highlights to Growing Intersection of Radiotherapy and Genomics

I recall that back in the 2000's decade there were a few reports of ways that genetic testing might help shape decisions for radiotherapy.  For example, there are some genetic variants that make clinical radiotherapy for more toxic to a few of us than to the rest.

In June 2018, the International Journal of Radiation Oncology, Biology, Physics, aka "Red Journal," has five special articles on the current status of precision medicine in radiotherapy.   While the main articles are subscription only, there is a summary of each of the main articles open access here.

While the HTML won't be perfect, I've clipped full titles and links from the table of contents after the break.

Topics include:

  • An overview by Kirsch (4pp)
  • A second op ed by Marks et al. (3pp)
  • An association-supported review and report on genomically guided radiation therapy by Hall et al. (8pp)
  • A report on big data and radiation therapy by McNutt et al. (7pp)
  • An article on the intersection of advanced imaging and radiation therapy by Jaffray et al. (7pp)

Monday, May 21, 2018

Very Brief Blog: Girish Putcha Deck on Clinical Utility & Payers 2018

Girish Putcha MD is both the Chief Medical Officer of Freenome and the Director of Laboratory Science at MolDx.  He presented an interesting 42-page deck at the Pathology Executive War College in New Orleans on May 2.   The deck is "branded" and logo'd as Palmetto MolDX" while also stating that "all opinions expressed are my own."  See the open access deck online here.

He emphasizes the importance of careful definition of an intended use and intended use population.

He highlights three special issues arising in the past year as (1) tissue agnostic indication for Keytruda; (2) 510(k) clearance for MSK IMPACT test; and (3) the CMS NCD for NGS in advanced cancer.

Regarding tissue agnostic indications and Keytruda, he highlights the very limited amount of data available on the Keytruda labeling at the time of tissue agnostic accelerated (provisional) approval.

Humorously, he introduces the CMS with a slide showing the famous "Serenity Prayer" (Grant me the patient to accept what I cannot change...)   He notes that tests like MSK IMPACT had guaranteed coverage in the NCD draft version (albeit with burdensome CED) and have only potential LCD coverage in the final NCD.   He notes that MSI and TMB are "inside" the FMI F1 CDx test, but "not" listed as CDx per se.   

He makes a number of points on regulation of lab tests:

  • He asks rhetorically, if CMS can require FDA approval for tests, can MACs do so (if they want to).
  • Regarding "regulation," he notes that Exact Sciences with an FDA approved test has a market cap of $6B, much higher than companies with LDT tests (Genomic Health has a market cap of $1B).  
  • He notes that physicians said CLIA '88 would be very bad with patients and trash access to lab tests, but it didn't.  
He closes by asking for robustly researched, regulated, and evaluated tests with appropriate reimbursement to support the same. 


For reinforcing recommendation for more highly validated and also higher reimbursed diagnostics, see Hayes 2014 here.


Section 216 of the 2014 PAMA law set in motion a repricing of the Medicare lab fee schedule to market average prices based on trienniel surveys.  However, it also created a new test category called Advanced Diagnostic Laboratory Tests (ADLTs), which must be sole-source tests and are priced and repriced annually.   CMS released instructions for the ADLT process in March 2018 (here).

On May 21, 2018, Foundation Medicine issued a press release that the FoundationOne CDx test had been awarded ADLT status from CMS (here).   Under AMA Proprietary Lab Analysis (PLA) code 0037U, the test will be priced at $3500 for three quarters beginning July 1, 2018.  Thereafter, it will be priced annually at its median private payer price.

The stock was up 7%, reflecting a market cap bump of about $200M.   A bit surprising, since the ADLT status was a shoo-in as far as I could tell.  Much of that was between 10 am and 1 pm, so it took the market a while to absorb the news.  (No need for nanosecond stock trades.  You could have slept in, bought at 10, and sold at 1.)

CMS updated its ADLT homepage (here) with a new PDF document, "List of Approved ADLTs," which current has one test listed, 0037U (CMS PDF here).

Small Puzzles Remain

FDA Approval is the Simpler of Two ADLT Routes to ADLT Status
Statute and CMS regulations create two types of ADLTs, the first being a sole source algorithmic test (e.g. MAAA), and the second being a sole source FDA approved or cleared test.   FoundationOne CDx is in the second group, where CMS approval as an ADLT should be pro forma.   For MAAA-type ADLTs, CMS created rather complex rules about the history of the test's development, licensing, and "uniqueness" that haven't been tested yet.*   CMS has stated, however, that ADLT applications will be reviewed by its standing laboratory advisory committee.

Initial Payment Clearly at "List Price" Per CMS
Multiple sources in CMS guidance state that the initial 3 quarters of an ADLT test payment will be at "list price."  For example:

  • Rulemaking at 81 FR 41100 (June 23, 2016) created 42 CFR 414.522, stating that "The payment rate for a new ADLT during the new ADLT initial period is equal to its list charge." (Here, PDF page 65).   
  • CMS also notes that SSA 1834(A)(d)(1)(B) "defines the term 'actual list charge' to mean the publicaly available rate on the first day at which the test is available for purchase by a private payor."  (Here, PDF page 5, 81FR41040).   
  • In its March 2018 guide to laboratories making ADLT applications, CMS stated similarly "Once a new ADLT initial period begins, payment for the new ADLT is made based on its actual list charge amount for the entire duration of the new ADLT initial period [3 quarters]."  See here, PDF page 22.   

I put this under the "small puzzle" column because in its press release, FMI states CMS will provide reimbursement at $3500 during the initial period.  Generally, publicly available sources have quoted the FMI list price as considerably higher, e.g. $5800.   I've also seen FMI's average payer payment stated at circa $2400, and I've noted that you can impute its overall net average payment closer to the range of $1000 per test (e.g. 10,000 clinical tests for $10M per quarter clinical revenue).  For PAMA calculations, CMS does not count "zero" payments. 

One answer to the puzzle - reading the instructions scrupulously - is that FMI could have set the F1 CDX list charge at $3500 literally on day one, and a higher price (say, $5800) on day two forward.  There is a clawback if CMS payments per test during the first three quarters prove to be too much higher than the later-calculated private payer rate during the same period; the tactics and strategies for this get complex (here).

Nerds see here for eye-crossing license-based exclusions from ADLT status; these apply to MAAA-type but not FDA-type ADLTs.


Sunday, May 20, 2018

Very Brief Blog: New Report on Rate of Closings and Mergers in US Oncology Practices

As reported by ASCO POST on April 23, 2018, the Community Oncology Alliance released its 2018 practice report - showing the continuation of a dramatic shift of oncology care towards hospitals and hospital-owned practices.

See ASCO POST here.   See the 2018 report (PDF) here.   The topic is inevitably tied to 340B drug purchase programs; see NEJM 2018 here and Congress Energy & Commerce Committee 2018 here.

In highlights,
  • 423 clinics closed
  • 658 acquired by hospitals
  • 168 merged or acquired other than hospitals
  • 359 more clinics were "struggling" and
  • 45 clinics stated they were referring Medicare patients elsewhere.
In the figure below "blue" dots are practices acquired by hospials and "purple" dots were merged or acquired.  

Consolidation of oncology providers can also affect practices for detailing and providing oncology drugs and diagnostics.


Saturday, May 19, 2018

Very Brief Blog: FDA Retracts Opportunity Offered Only to Mark McClellan's Institute

On May 17, 2018, John Carroll's blog Endpoints News collated and discussed a recent story that emerged from Politico and WaPo and other sources.

Mark McClellan, former head of both the FDA and CMS, and Scott Gottlieb, current head of FDA, are two of the stars of the Washington health policy world.

A few weeks ago, FDA published a request for proposals with the government prerogative of making it a "sole source" offer - to McClellan's Margolis policy center, located at Duke with a DC branch.

FDA reversed course "after it became aware" that potentially other centers could apply for the funding, which is to help FDA convene public stakeholders and experts on matters of interest, such as improved structured risk-benefit assessment.    As WaPo noted, the award budget will be $4.2M.
  • See Endpoints News, here, May 17.
  • See Politico, here, May 16.
  • See WaPo, here, May 16.
    • WaPo notes that FDA made a single source grant to Brookings Institute Engelberg Center in 2013 (here).[*]
In WaPo, Jerry Avorn, a Harvard health policy expert and frequent contributor to NEJM and JAMA, stated his surprise that FDA would view the topic as one leading naturally to a sole-source RFP.

The Actual RFP

The original RFP (RFA FD 18 013) was withdrawn from its online status but is available on Google Cache (as of 5/20/2018) here and a cloud copy of the original here.  Note a simple remark on page 8 that entitie(s) eligible to apply are: "Duke University - Margolis Center."

It was reissued as award FD 18 025, online here.  The new application cycle is May 17-July 18.  See "Part 2, Section I," for research objectives, specific research interests, and the anticipated approach. (I've also posted a cloud copy here.)   The award budget is stated as about $850K in each of 5 years (FY2018 - which will nearly be over by the award date), 19, 20, 21, 22).

Topics include - well, almost everything the FDA does, in a lengthy bullet point list, from "enhancing regulatory science and expediting drug development" to "enhancing the use of real world evidence" and "the patient's voice in drug development and decision-making."

The two areas receiving the most highlight, in paragraph form, are enhancing PFDD - patient focused drug development - and structured risk-benefit assessment.

Structured Risk Benefit

While "regulatory science" is a very broad domain, the core principle is to make risk/benefit decisions, such that benefit is greater than risk.  This is tortuous for several reasons.  For one thing, risks and benefits are incommensurate - Viagra has one set of benefits, but a completely different sets of risk (e.g. stroke).[**]  Scientifically, when is one great than the other?   For a second thing, trial design and statistics are completely different.  Benefits are usually very tightly defined endpoints with hard statistics (survival increases 8 weeks ± 1 week.)  Risks are random, unpredictable, and have little statistical weight because they are ad hoc and observational - two patients had a heart attack, one had new migraines, and one had a stroke.  

Traditionally, FDA advisory boards see a 100 slide presentation of benefits, get a coffee break, see a 100 slide presentation of risks, then chat for a while and vote.   That's hardly "structured."  FDA has posted some "qualitative structured" templates, but the ones I've seen are hardly more than putting benefits in a list on the left and risks in a list on the right - presumably what people were doing in their heads anyway.   (The technique was published by Benjamin Franklin in 1791).  

For some current FDA thinking on structured risk benefit, see here here here and here.  For an 18 page 2017 Duke/Margolis white paper on the topic, here.

I wrote about the topic in this blog in 2014 (here).  I revisited in 2015 regarding flibanserin, not to promote flibanserin but to note how openly disparate the "risks" and "benefits" were.


[*] The 2013 award was to study a range of bullet-point initiatives (similar to the current award) with special focus on patient reported outcomes, on structured risk/benefit, and on standardizing and evaluating REMS.  That award amount was $700K for one year, similar to a one-year budget in the new grant.
[**] In case it's not obvious, the risk benefit "equation" would be highly subjective, such as whether 12 sexual encounters per year are worth (say) 3% risk of stroke.  (Discuss and explain your answer.)

Friday, May 18, 2018

Very Brief Blog: CMS's Web-Friendly, Searchable Hospital DRG and Payment Data

CMS has number of web-searchable, big data provider databases.  I've written about the one for physicians and labs.  There's also one for ordering physicians for DME and one for hospitals.

Home page for inpatient hospital data is here.  (For other options, see the column at left on this CMS page).  As of May 2018, the most current data year is 2015.

I briefly explain how to get into the data and download it in a short footnote.[*]

If I've got this right, for Heart Transplant, DRG 001, there were 1,633 cases among 74 institutions.  7 hospitals had 40 or more Medicare heart transplants, the top four being Advocate in Illinois, St Lukes/Baylor in Texas, Barnes in Missouri, and Cedars in Los Angeles.  $360M was paid, for an average of $220,711 per case (standard deviation about $60,000).  

The famous charge:payment ratio ranged from 2:1 for Tufts to 12:1 for Temple.  All sorts of tidbits - e.g. Mother Mayo in Minnesota had 26 cases, but Mayo Arizona had 28.

Recently there have been new calls for more transparent Medicare data, such as hospital charges and drug prices, and admittedly, this is stuff for experts or at least people facile with Excel.  But still, it's publicly available today on open websites and can be accessed in a few seconds as I've shown above.


For a May 16 investigative article in the Houston Chronicle on the #2 volume center, St. Luke's/Baylor, see here.

I believe the physician data is all CPT codes and the DRG data is all DRG codes.  The Outpatient data for some reason lists only 28 of the many APCs, though.

[*]  Home page for hospital data is here.   Click on "Interactive Dataset...FY2015."  At the next page, click on the box "View Data."  You've now got a table-format web view of what's a huge dataset.   Let's filter it.  Click on the "Filter" box.  Click, "Add a new filter condition."  Change DRG Definition to "...Starts With" and type 001.  It now displays all heart transplant DRG utilization by hospital.  Click on "Export" and export as Excel.  Finally, I immediately save the CSV file as XLS.

Very Brief Blog: Genomic Health Announces Date of Service Investigation

In its May 9, 2018, 10-Q filing with SEC, Genomic Health notes it is cooperating with an investigation into Medicare "date of service rule" compliance.
  • 10-Q here.
    • Search for: date of service; several hits in the 64 page PDF document.
    • GHDX is up 6% in the week since mentioning the investigation.
    • The May 9 document was subsequent to its May 2 investor call.
  • Brief Genomeweb report here.  I believe it's open access.
In recent news, from March, there was also a compliance investigation of Myriad Genetics billing (Bloomberg here; also here).   As often occurs, there is a parallel shareholder lawsuit (here).  Myriad share price is up 20% since mentioning the investigation.

Also in March, Natera Inc. settled a Department of Justice investigation into compliance with a $11M payment.  Press release here. Natera share price is up 26% since the press release.

Wednesday, May 16, 2018

Very Brief Blog: Access the Medicare Drug Spending Dashboard (Part B, Part D, Medicaid)

In the past week, much news about how the Administration may reduce drug costs, including several speeches by HHS Secretary Alex Azar.

Mentioned is a newly updated "Drug Spending Dashboard" at the CMS website.

  • Access it here.  
    • Note different pages available for Part B, Part D, Medicaid.
  • Brief article at here.

Brief Blog; Medicare Opens NCD Process for CAR-T Therapies

On May 16, 2018, CMS opened a National Coverage Determination process for CAR-T therapies, based on a request from United Healthcare dated February 22, 2018.
  • The NCD tracking sheet is here.
  • The UHC letter is here.
    • The letter specifically refers to creating a "level playing field" for Medicare Advantage plans in regard to these costly drugs.
    • The letter proposes on-label coverage with a flexibility for important valid uses that may not be on the FDA label.   
  • Coverage at BioPharmaDive, here.
The timetable includes an initial open comment period (on the topic; there is no draft NCD yet) from May 16-June 15, 2018.   CMS will hold a "MedCAC" public meeting on August 22, 2018.

The draft NCD is expected February 26, 2019, and the final NCD (after a comment and review period) is expected May 17, 2019.

Of note, CMS had a relatively extended and skeptical range of criticism in April 2018 inpatient rulemaking, in response to a request for new technology add-on payments (NTAPs) for Kymriah and Yescarta.  See the relevant pages in the cloud, here.  (Somewhat scarily, on page 20189, CMS referred to the diseases under treatment as "hematopoietic carcinomas."  Ouch.  Carcinomas are normally other classes of cancers, excluding leukemia and lymphoma.)

The last NCD and MedCAC on a cancer drug covered Provenge for on-label uses in 2011, here, here.

I've clipped the CMS announcement of the NCD process below.  The lead analyst is Katherine Szarama, PhD; the lead medical officer is Lori Paserchia MD.

Brief Blog: National Academies Releases a Report on Paying for Costly Illness; Chance to Revisit Drug Pricing Report 2017

On May 16, 2018, the National Academy of Sciences released a 71-page ebook on "Financing and Payment Strategies to Support High Quality Care for People with Serious Illness."   Find the ebook here.  The underlying workshop was held in Washington at NAS last November, here.

May 2018: Financing and Payment Strategies EBook

Returning to the opening topic of the new May 2018 ebook on health financing.  The "Financing and Payment Strategies" ebook, planned and executed more cohesively under one administration, is much shorter and has chapters like "challenges and lessons of fee-for-service" and "challenges and lessons from global budgeting arrangements."  The ebook does not have panel consensus recommendations but includes several pages of individual recommendations as remarked on in the course of the meeting, by individual participants.  

December 2017:  More Ruckus Over NAS EBook on Drug Pricing Options

With all the visibility to last week's speech by President Trump on U.S. drug pricing, we should recall that National Academies also released a 235-page ebook on drug pricing last November 30, 2017.   That report was based on a December 13, 2016 conference (e.g., planned under Obama administration), here, here

The November 2017 release included a one-hour conference (video here).  Find a unique cloud transcript of the videoconference, here.   On December 13, 2017, there was an Energy & Commerce hearing on drug pricing; website here, Politico here, WaPo here.

Find a summary of the Making Medicines Affordable November 2017 release, citing six trade journal articles, condensed at Kaiser Health News, here

Drug pricing recommendations included:
  • Accelerate market entry, including generics and biosimilars
  • Consolidate government purchasing power and apply it
  • Improve drug valuation methods
  • Greater transparency of cash flows in drug supply chain
  • Discourage DTC advertising
  • Modify & mitigate cost burdens for patients
  • "Eliminate misapplication of funds in federal discount programs" e.g. 340B
  • Target incentives for rare diseases only to rare diseases
  • Align physician prescribing with value
The drug pricing book took one year to appear, longer than the usual NAS timetable of four to six months.  The ebook contains a 20-page dissenting view by Michael Rosenblatt and the late biopharma entrepreneur Henri Termeer.  The dissent chapter appeared, in part. in NEJM here.  Rosenblatt is a physician executive at Flagship Pioneering.   The ebook also contains a shorter, 5-page "minority view" with remarks such as, "Patients are left at the mercy of coverage and pricing decisions that are completely unknown to them."  

Jump from December 2017 to May 2018.    
In Mid-May 2018, NPR highlighted 3 facets of the new Trump proposals as (1) more transparency for supply chain channels and PBM discounts, (2) shift some costly drugs from Part B (average sales price) to Part D (negotiated Part D plan price), and (3) make it much easier to find Medicare and Medicaid drug prices.  (These prices are often available today on CMS websites and fee schedules most easily found by experts.)  
A transcript of a speech by HHS Secretary Alex Azar on the drug topic May is here; a second speech on May 16 is here.  Trade article here.  Both speeches & trade article in the cloud here (7000 words.)  Redline comparing the two speeches is here.


NAS notes that both books are conference reports of diverse panels and do not represent positions of NAS itself.

Since both ebooks are about health pricing, it might be of interest to compare with the May 16, 2018, New York Times article on the US healthcare price explosion after 1980 by economist Austin Frakt - here.

The same week as the NAS report on healthcare costs and strategies, the Urban Institute issued a plan for subsidized insurance that would reduce the uninsured by 16 million, here.

Tuesday, May 15, 2018

Very Brief Blog: Illumina Acquires Edico for $100M

Last year, I had the chance to chair two conference sessions and write an article on the topic of digital genomics, companies dedicated to the digital layer of services rather than running a wetlab at all.

Edico Genome was one of the companies discussed.  This week, it's been acquired by Illumina for $100M.  Over the past several years, Edico had raised $32M in venture capital. 

Edico enables both research and clinical technologies; for example, it's been used by Rady Children's Hospital in landmark applications of rapid genomic sequencing in very ill babies.  (See Farnaes et al., 2018, here.)

  • San Diego Union Tribune, here.
  • Xconomy, here.
  • Genomeweb, here.

Illumina has a market cap of $38B and cash and equivalents of $2B.  

Very Brief Blog: OIG Says CMS Telehealth Payments Often Incorrect

There have been many voices calling for better Medicare coverage of telehealth, including some small legislative fixes in the recent spring Fiscal Year Budget Bills. 

However, others have been concerned that telehealth and digital delivery modalities could be subject to incorrect billing or worse.   OIG issues a report on this topic in April 2018, finding a 30% error rate.  Telehealth remains, however, a very tiny fraction of Medicare spending, despite having risen from $61,000 in 2001 to $17M in 2015.

  • The OIG report is online here.
  • Trade press here.

With Telehealth spending of $17M in a budget of about $600B, telehealth is about 1/300 of one percent of each Medicare healthcare dollar.

Very Brief Blog: GAO Issues Report on CMMI Activities & Productivity; Other CMMI News

On April 25, the GAO released a report on the progress and results of 37 CMMI models.   GAO's bottom line was that of 37 plot projects, just 6 were shown to have successfully reduced spending (while maintaining or improving quality). Even fewer (2) were expanded so far.  $5.6B has been spent since the creation of CMMI by the Affordable Care Act in 2010.
  • The GAO report is online here (54pp).
  • Trade press at Advisory Board here, Healthcare Informatics here.
  • More trade press at AJMC here and  RevCycleIntelliegence here.
This blog wrote about the GAO report on CMMI Medicaid pilots in February 2018 here.   (GAO found lack of rigor in reporting.)  

Other CMMI Recent News

Reboot & Comment
CMS/CMMI launched a "pause and reboot" project in October 2017; and a few weeks ago in April 2018 released public comments and some ideas for moving forward such as primary care direct contracting (here).

CMMI Request for Information on Direct Contracting Primary Care
Find more info about this here.  Comments accepted til May 25.

$800M Budget Cut
In May 2018, OMB proposes to cut $800M from the CMMI budget. here.  CMMI has had a ten year budget to 2019, with several billion unspent and without a spending plan, so this is more a clawback of stored funds than a cutback in programming.

New Head Adam Boehler
The Obama era head of CMMI stepped down in Summer 2017; finally in April 2018 Adam Boehler was officially named the new CMMI director.  Here.  (Boehler's name had circulated in this context since December.)   Boehler founded Landmark Health, a novel venture funded delivery system for the chronically ill (more care at home) that aimed to support Medicare Advantage and other health plans.

Brain Drain
A March 2018 article in Politico discussed "brain drain" at CMMI, here.

Rigor at CMMI? Or Not?A May 2018 article in NEJM decries the shift from randomized trials to voluntary demos at CMMI, here.   I would add that since CMMI must by statute demonstrate its programs save money and raise quality, if they are not rigorous, CMMI may be shooting itself in the foot at evaluation time.

Slow Start at Diabetes Prevention Program
This is the CMMI funded pilot-to-expansion program.  Stumbling, per an article in Kaiser Health News, April 2018.  Here.  Update in May, here.


More generally on alternate payment models -

For a March 2017 JAMA article on value based purchasing ("Time for a reboot?") by Jha, here.

For a May 2018 editorial on measuring bundled payment programs in Annals of Internal Medicine by Provonost at Johns Hopkins, see here (trade press here.)    For a NEJM article on the poor performance of "performance measures," MacLean et al., here

For a comparable to Landmark Health, maybe Hometeam (here, here).

Thursday, May 10, 2018

Very Brief Blog: Tissue Type Distribution in FMI F1 CDx FDA Data

Based on its PMA approval last November, the Foundation Medicine F1 CDx test has two types of indications.  The first set is indication(s) for several CDx genes, generally obtaining the CDx authorization by comparison to an FDA-approved legacy single gene test in 150-200 blocks.   The second indication is for pan-solid-cancer tumor "profiling" which is "for use by a physician."   This second indication gets FMI pan solid tumor coverage under the recent CMS NCD for NGS tests in oncology.

On pages 34-35 of the FDA P170019B approval, FMI lists its tissue comparability data (document section 10.)   A large scale retrospective analysis was conducted used 80,715 specimens from 43 tissue types, "to demonstrate that genomic profiling can be performed on DNA derived from FFPE specimens of any tissue type."   The document states that "the data set for analysis consisted of routine clinical samples analyzed using the F1 LDT from March 25, 2015 to March 13, 2017."

I typed the FDA data into Excel to sort it.   FMI provides technical success data for each of the 43 tumor types after DNA extraction, library capture (LC) and hybridization capture (HC).   The vast majority of specimens pass all QC steps, but note that they had presumably been screened or submitted according to some instructions about the expected number of slides or blocks or percent tumor.

Assuming the 80,000-odd blocks were routine clinical workflow, that's about 10,000 cases per quarter over 8 quarters, which more or less foots to FMI statements about clinical volume.   The top ten tumor types, reaching 75% of specimens, are shown below.   Note that I've left the data exactly as in the FDA tables except for merging pancreas together with Whipple resection.

FMI F1 CDx PMA 170019B p 35 table 24

Three cancers - lung, liver, and brain - along with "lymph node" samples, are about half of all specimens.   The top ten together are about 75%, so the next 30 are only 25%.   So that's from the perspective of concentration.

You can also look from the perspective of dispersion - thirty-some types of cancer all log in at less than 2% each.  18% of samples had rather vague denominations like "lymph node" or "effusion."

For comparison, the annual cancer deaths (a proxy for advanced cancers) are lung, 150,000; colorectal, 50,000; pancreatic, 45,000; breast, 40,000; liver, 30,000; prostate, 30,000; leukemias, 25,000.   Brain cancer would be farther down, around 15,000.  In the FDA data, lung and liver FMI tests are about the same, but the incident US death ratio is 5:1 favoring lung.  Colon and brain FMI tests are about the same, but the US death ratio favoring colon is over 3:1.

The original FDA document is here, see pages 34-35.  My typed and sorted table is in the cloud here.

The FDA doesn't seem to require 80,000 blocks in 43 tumors to merit the pan-cancer label; the MSKCC IMPACT test also is labeled as pancancer based on about 10,000 blocks in 17 cancers. 


Update: A numeric error was corrected a few minutes after the original post.
ACS Cancer Facts 2018, here. Odds of dying of 20 cancers (e.g. testicular cancer, odds, 1 in 5094, here.

Monday, May 7, 2018

Genomeweb Scoop: Discusses FDA Response to IVD Reform Legislation

In March 2017, the House released a 215-page discussion drat of a bill that would substantially overhaul the FDA's approach to diagnostics, both LDTs and IVDs.    See the bill online here.

On May 4, Turna Ray of Genomeweb published a detailed article described an FDA technical response (comment) on the legislation.   As of a couple weeks ago, the word was that the Hill was still waiting for this FDA document to arrive.  The subscription Genomeweb article is online here.  The journal obtained, analyzed, but didn't post the FDA document.

According to Genomeweb, FDA would encourage higher positioning for its "precertification" program based on lab or manufacturer reporting/recording of analytical and clinical validity.   FDA was concerned about some details of the bill such as the breadth of its grandfathering clause.

A March 2018 Genomeweb article captured Commissioner Gottlieb's remarks at the ACLA annual meeting, where he also noted he was favorable to a long time legislative overhaul which would specialize diagnostic regulations away from the general regulation of medical devices like implants.

FDA is well along in its plans to create an all-new soup to nuts digital health pre-certification program in 2018 (here), and the future diagnostic pre-certification program would be analogous but fitted for diagnostics.   FDA first rolled out elaborate genetic test pre-certification or self-certification in germline test authorizations crafted for 23andMe several years ago.

Friday, May 4, 2018

CMS Publishes Full Code List For New Lab Codes; Big Revisions to BRCA

CMS holds an annual public meeting to allow public comment on its pricing options for new, and revised, AMA CPT laboratory codes.   CMS publishes an initial list (mostly PLA) codes in early April; CMS has now published a full list of 81 codes.   BRCA codes are scheduled for extensive revision and CMS repricing.


  • See the CMS lab meeting webpage here.
  • See the full Excel spreadsheet (as a zip file) titled "as of April 30 2018," here.
    • The full Excel spreadsheet should have 81 agenda items.
    • (Note they've left up, for now, a partial spreadsheet posted a few weeks ago.)
Registration for the public meeting is open until June 11 (click the "register" link).   Note there are also explicit instructions for your one slide Powerpoint presentation format, click on Power Point Presentation Template.  

Both registration and slides are due June 11; note that the meet could fill up prior to that date due to limited audience space (for people) and limited agenda time (for slides).

What's On Deck?

The agenda lists 2 tests for pricing reconsideration, 81334 (RUNX1, in AML) and 81326 (PMP22, in Charcot-Marie-Tooth neuropathy).

Items 3-29 (26 items) are PLA codes, the new quarterly "proprietary lab analysis" category of code.  These are diverse codes; some human genetic, some microbial, some MAAA and some FDA approved tests.

Items 30, 31, 23 are "administrative MAAA" codes (such as 0011M).

Items 33-81 are Category I genomic CPT codes for the most part and mostly single genes (80X01 is a dihydrotestosterone chemistry code).  

Agenda items 45-49 are revised BRCA codes.
  • 81X78 will be BRCA1, BRCA2, full sequencing.
    • 81X79 will be BRCA1, BRCA2, full dup del analysis (e.g. large rearrangements)
  • 81X81 will be BRCA 1, full sequencing and 81X82 will be BRCA 1, full dup del analysis.
  • 81X83 will be BRCA 2, full dup del analysis
The "new codes" agenda doesn't list deleted codes, since a deleted code does not require pricing, so the status of some existing BRCA codes like 81211 isn't explicit from the CMS CPT listing.   In its public report on the February 2018 AMA meeting, the AMA summarized as follows:
"Accepted addition of five new codes to report full sequence analysis, full duplication/deletion analysis [for] BRCA1 and BRCA2; revision of code 81162 to include full duplication deletion analysis; revision of 81216 to include HUGO gene name; deletion of 81211, 812131, 81214."  The codes-affected list was 81162, 81211, 81212, 81213, 81X78, X79, X81, X82, X83.
AMA could, possibly, make alternations between the February meeting public report and the code book going to press in summer.  Basically, from the above information, it looks like they've left in place 81162 (BRCA1, BRCA2, full seq + full dup del) but if a lab is going to report only sequencing or only dup del, it could use 81X81 alone or 81X82 alone instead of 81162.   (I would strongly predict a CMS edit not to use 81X81+81X82 together as stack codes.)   PAMA has impacts in all this, because 81162 drops in price under PAMA while 81211 rose in price under PAMA, at least, until it's deleted.

Very Brief Blog: Who Invested in Theranos?

While Theranos' valuation once exceeded a billion dollars, it also absorbed and spent $600M in cash.  Where did that come from?   Wall Street Journal provides a listing based on court documents.

High profile investors included the Walton family (Walmart; $150M), Rubert Murdoch (Fox; $121M), the DeVos Family (Betsy DeVos, Education Secretary; $100M); the Cox family (media; $100M).   Runners-up include Carlos Slim, the Mexican billionaire, and others including the ex-chairman of Bechtel and the owner of the New England Patriots

My library of a couple hundred media Theranos links, 2014-2016, is here.

Thursday, May 3, 2018

Very Brief Blog: Gilead to Invest $90M in Verily's "Immunoscape" Immuno Cellular Profiling

Update:  On May 14, Google Ventures (GV) and Third Rock fund a Broad Institute spinout, CELSIUS, for $65M to do similar work on sophisticated molecular map analysis of immune system for disease definition & drug development.  John Carroll's Endpoints article here.  

Providing contract research to pharma is a big industry, and includes not only clinical trials but also R&D.   Gilead, planning to move further into drugs for inflammatory diseases, announced it is investing $90M in research at Verily.   The platform will be Verily's "Immunoscape" - a name that was trademarked only last December.  Here's what it does:  "The Immunoscape platform represents the next-generation of tools to understand the immune system, harnessing the latest in genomic and immunologic laboratory technologies and leveraging Verily’s ability to curate and analyze health data at scale."

For comparison, based on their latest investor call, Foundation Medicine's annual pharma revenue is the the neighborhood of $100M.

The fact this big molecular science deal lands at Verily - part of Google Alphabet - rather than a traditional life science lab or in an offshoot from academia is telling.  Possibly the actual data - for example, high sensitivity flow cytometry and RNA expression - is relatively straightforward today and the secret is in the resulting big data or machine learning analysis.   If so, put this one in the category of the impact of digital health modalities on genomics.  (See my December 2017 article).  For another perspective, $90M would swamp the budgets of academic labs looking at immune profiling and expression across several autoimmune diseases.

Read more at Genetic Engineering News and at Forbes.   Head of the lab project is Charlie Kim PhD, who moved to Verily from his position as assistant professor at UCSF (Linked In here).  (To use the perspective theme once again, imagine booking a $90M grant as a beginning assistant professor.)

Currently, the home page for Verily Projects lists four broad investment and research categories - sensors, interventions, health platforms and population health, and precision medicine.   Each category has 3-6 projects, of which Immunoscape is one project out of 6 in the precision medicine category.


Verily and Gilead are located in close proximity to each other on the SF peninsula.

Very Brief Blog: Is "Transformation of Healthcare Systems" Mostly Hype?

While we hear much about alternative payment models and health system reform, is much really happening?

U Penn / Wharton health policy gurus Lawton Burns and Mark Pauly are doubtful, as expressed in a new article in Milbank Quarterly.  It seems to be open access.   Find it here.

The takeaways:

  • Policymakers seek to transform the US health care system along two dimensions simultaneously: alternative payment models and new models of provider organization.
  • This transformation is supposed to transfer risk to providers and make them more accountable for health care costs and quality.
  • The transformation in payment and provider organization is neither happening quickly nor shifting risk to providers. The impact on health care cost and quality is also weak or nonexistent.
  • In the longer run, decision makers should be prepared to accept the limits on transformation and carefully consider whether to advocate solutions not yet supported by evidence.

Very Brief Blog: FMI Definitely Becoming an ADLT; Definitely Has Pan-Cancer NCD Coverage

On May 2, 2018, Foundation Medicine held its Q1-2018 quarterly earnings call, which is available online here.   See also summary at Genomeweb here.  Q1 quarterly revenue reached $53M, up from $26M one year earlier.

Definitely Pan Cancer Coverage at CMS
Based on explicit analyst questions, FMI confirmed that CMS had confirmed that it has pan cancer coverage in solid tumors.   FMI pointed to its label, which has both an indication set for several CDx genes and also a clearly stated FDA indication for pan solid tumor gene profiling.   The analyst emphasized that the NCD interpretation was perplexing to observers.   I called it a "Rubik's cube" problem in a recent blog and white paper -- but my analysis then  seems to concord with FMI statements this week, and by proxy, with CMS.  See that blog here.

Definitely Applying for ADLT Status
FMI also announced it was definitely applying for ADLT status.  It will be paid at local contractor rates since then; at least one announcement stated its local contractor rate with Palmetto was $3415.  (In the earnings call, FMI referred 3 times to Palmetto but never to the MAC in its home region, Massachusetts.) 

ADLT rules allow local MAC payment rates until an ADLT decision is made by CMS (here, likely July 1), then, 9 months of payment at the list rate, and then, payment reset annually to a market rate.  There is a clawback option if the 9 month list rate is too high above the retrospectively confirmed first year market rate.  CMS released ADLT instructions in mid March (blog with CMS links and my discussion, here).   I pointed out that labs have a lot of options for controlling the first-year data that results in the second-year median rate (here).

Definitely Shooting for TMB in an FDA Liquid Biopsy Test 
FMI emphasized that its ACT liquid biopsy test now entering FDA review would include tumor mutational burden (TMB) information.

I'm not a card carrying regulatory expert, but based on recent FDA approvals, getting a pan-cancer label out of the FDA isn't a shoo-in and it's a white space to do it in the liquid biopsy sphere.   As part of their November 2017 FDA reviews with pan-solid-cancer indications, FMI provided data from 80,000 blocks across 43 cancers, and MSK IMPACT provided >10,000 blocks across 17 cancers.  (See my Rubik's cube article cited above).

Revenue Per Test Unit
Biopharma revenue tests were reported as quarterly 7,184 units for $27.8M revenue ($3869 per unit) and one year ago 1,802 units for $9.5M revenue ($5272 per unit.)    Clinical tests were reported as 21,861 units for $18.8M revenue ($860 per unit) versus one year ago 13,900 units for $11.6M revenue ($833 per unit).   Note that revenue recognized per clinical test unit may lag as test volume grows ahead of collection cycles.  Still.  FMI noted it has switched from cash- to accrual-based clinical revenue.  If I read correctly, they state that cash-based Q1 clinical revenue would have been a bit higher, $22M, than the accrual-based Q1 clinical revenue figure.

Share Price
FMI's share price is about $70, up from about $50 just before the draft NCD was released in November.  However FMI's share price peaked at $86 in February, before the NCD was finalized.  Market cap is $2.6B. 

Brief Blog: New Cycle of Activity on "Cancer Sequencing Hype and Reality"

Science blogger Derek Lowe has a new article May 2 titled, "Cancer Sequencing Hype and Reality."  The essay is predicated on a recent debate at AACR between David Hyman of MSKCC and Vinay Prasad of OHSU.   It tracks a report on the conference published in Science by journalist Jocelyn Kaiser, here.

As Lowe tells it, in the debate, Prasad focused on the fact that only about 15% of lung cancer patients have an FDA-actionable mutation and only about half of them have a corresponding therapy response, therefore, precision medicine has too much "hype."   I'm not sure; I don't think it's a surprise to anyone in the field - it's common knowledge that a (fairly small) subset of lung cancer patients have EGFR and ALK and then ever rarer mutations.   Maybe it's "hype" that there is "hype"...

See Hyman's presentation in AACR video online, here.  Hyman's team showed that larotrectinib, an investigational drug, was active against TRK fusions in a wide range of tumors (NEJM, 2018, here.)

See an article in Medscape here that reviews Prasad's April 2018  JAMA paper on "percent of US cancer patients that benefit from genome driven oncology."

I haven't updated it but around 2015-2016 I had a blog inventory of fifteen or twenty articles, at that time, skeptical of precision medicine (here).


PubMed currently has 1300 citations for "HYPE" (here) only a few of which are the gene carbamoyl dehydratase (HYPE). 


In Nature Reviews Clinical Oncology, Kimmelman & Tannock (2018) argue that precision medicine subpopulations can be so small that any outcome and comparison studies are necessarily very limited, with large error bars, thus generating paradoxical "imprecision within precision;" here.

Very Brief Blog: NAS Meeting on Genomic Health Disparities, June 27, 2018

The National Academy of Sciences has a standing committee called "Roundtable on Genomics and Precision Health," here

They've announced a new public workshop, to be held in Washington on June 27, 2018, with the title, Understanding Disparities in Access to Genomic Medicine.  See the webpage here.   I've clipped the homepage in this blog below the break.  I was very happy to have the opportunity to serve on the organizing committee for this workshop.

Track them at #GenomicsRT and #GenomicsDisparities

The group's prior workshop was November 1, 2017, on "implementing genomic screening programs in health care systems."   That webpage here; the proceedings of that autumn workshop were released as a detailed eBook on March 16.