Lecture Notes: German Health Minister Sets Goal for Fast Digital Advances in German Healthcare

Some readers notice that I am a lifelong German-language hobbyist, and I know more the health system in Germany than any other European country.  Sometimes this pays off, like this week, when I attended a fascinating lecture Germany's Health Minister on dramatic new goals for a digital push forward in the German healthcare system.   The event was a joint Q&A with Dr Karl Lauterbach and Dr Micky Tripathi, the National Coordinator for Health Information Technology at the US HHS.  It was held at the Kennedy School at Harvard on September 22.

Dr. Lauterbach, who holds a PhD in health policy from Harvard, became well-known in Germany during the COVID crisis but did not ascend to the position of Health Minister until the change of  government in late 2021.   As prep, I ran across a 30 minute interview that Lauterbach gave a few days ago, with the German newspaper DIE ZEIT, on AI in healthcare (German; here).   He recapitulated many of the same points in Boston this week.

Lauterbach focused on 4 or more pieces of legislation just now moving through the writing, Cabinet-approval, and parliamentary process in Berlin.   The results would connect doctors' offices, hospitals, and pharmacies throughout German with a unified electronic health record.   

Moreover, there would be an "outer ring" and an "inner ring," per Lauterbach.  The outer ring would be encrypted but patient-specific, and patients (who could opt out entirely) could easily allow transfer of clinic visits and history to new providers.   The inner ring would be anonymized and constitute a regulated secure "Federal Research Space" as a learning health system and for research.  AI could help create retrospective narrative structures from original free text.  

Other features mentioned by Lauterbach included a reform of the research approval process so that high standards of patient safety could be maintained while shrinking the burocractic approval process to 30 days.   Both at DIE ZEIT and at Harvard, he referred several times to 2025 as the goal for putting major pieces in place (this seemed remarkable to me).  

In the DIE ZEIT interview, countries like Israel and Sweden were used as models and goals for electronic health records the great-leap-forward in Germany.   In addition, online, the interviewer noted that Germany was not only working with or learning from nations, like Israel and Sweden, but Microsoft and Google were at the forefront of digital health and AI.   Echoing this exactly, at Harvard, Dr. Tripathi closed by noting in the last few days he was keynoting at meetings where he was preceded or followed by Microsoft or Google.   

##

Context.  It's difficult for me to entirely place this "new plan" in context, as a quick Google search the last five years shows a plethora of articles about prior legislation and prior goals and deadlinees to help Germany "leap forward" from its backward position in digital health records and interoperability ("Digitalisierung.")  E..g. this page is dated 2020.  Compare a 2023 essay here.  Other articles talk about the shaky status of German hospital economics, and Covington just published an article updating us on pending major hospital DRG reforms in Germany, here.

AI Corner

I left the Lauterbach-Tripathi session with a page of fragmented phrases, keywords, and roughly typed notes.  I dumped them into Chat GPT and asked it to edit them into an elegant journalist-style essay.   Here.

Bonus Book

Health care change, let alone health care reform, has always been messy and chaotic and confusing.  See a 2021 book by a Yale professor, Peter Swenson, covering 150 years of zig-zagging US health policy - Disorder.

Journal Club: New JAMA Essay on AI in Healthcare (Gottlieb & Silvis)

A new, open-access JAMA Health Policy essay, "How to Safely Integrate Large Language Models Into Health Care," two authors with both industry and FDA experience discuss pathways for integration of Large Language Models (LLMs) in healthcare. Dr. Scott Gottlieb  served as FDA commissioner and Lauren Silvis is a former senior FDA regulator now at TEMPUS.

The essay highlights the potential of LLM's to enhance patient interactions and healthcare delivery.  Three takeaways:

LLM Potential: The essay underscores LLMs' transformative potential in healthcare, enabling human-like text generation, aiding in diagnosis, and providing ongoing patient support.

Staged Integration: The authors advocate a cautious and phased approach, starting with well-understood conditions, to help ensure safe LLM integration and boost confidence.

Data and Bias: Effective LLM use in healthcare demands improved data sharing, bias detection, and inter-institutional collaboration for accurate and unbiased results.

Very Brief Note: 153 Comments Posted on "TCET" Proposal

 CMS has posted (after a few weeks' delay) the 153 comments on its TCET, emerging technology proposal:

https://www.regulations.gov/docket/CMS-2023-0107

TCET was discussed multiple times at a September 19, 2023, House Energy & Commerce hearing on Medicare and innovation.

https://www.discoveriesinhealthpolicy.com/2023/09/september-19-2023-congress-holds.html

September 19, 2023: Congress Holds Hearing on CMS and Innovation

On September 19, 2023, the House Energy and Congress committee held a hearing on, "Examining policies to improve seniors’ access to innovative drugs, medical devices, and technology."

Find the home page here.  The agenda includes Dr Dora Hughes, chief medical officer for CMS. The E&C website links to the YouTube archive video stream.  Her ten page testimony here.  (Most of it, was a fairly dry review of LCD and NCD and CED processes).

Chair's remarks here.  

The most interesting document is the 6 page hearing memo here.  Remember, for Dora Hughes, this is essentially a hostile audience - the Republican House vs the Democratic HHS administration.  The hearing memo reviews 10 or 12 "legislative fixes" that have been proposed to improve the NCD and other CMS processes.

AI CORNER

I've posted an unofficial auto transcript beginning with an unofficial detailed summary, in the cloud here.

MolDxology: DEX Registry Allows Categorical Browsing

I'm not sure when this started, but I'm pretty sure it didn't used to be here.   The MolDx Z code online registry, DEX (which shows all about a registered test, EXCEPT its actual secret Z code) has always allowed searching by test name and lab name.   

(Test name can be a real wild west, and lab name can even be complicated in case of acquisitions and subsidiaries.)

Now Palmetto DEX allows searching by five different drop-down categories.  At the left-hand side, these are diseases, FDA status, medical specialty (e.g. cardiology), method, and test type.  (Test type is "diagnostic, confirmatory, predictive," etc).  Find DEX here:  https://app.dexzcodes.com/

There appears to be a built-in "and" logic, for example, here I've searched for the Test Type category "DIAGNOSTIC" while I happened to have let the test name "RaDaR" in the name search field.  This excluded quite a few variants of the RaDaR test, but showing one that is classed as "diagnostic."   

The DEX registry is available with free email registration.

##

Bonus

The New York State "Wadsworth" system allows searching for lab tests endorsed by New York State, here.  Article in 360dx.

##

AI Corner:

This blog in French.

ACLA, Others, Comment on Verbose and Confusing Novitas LCD for Oncology

 Last year, Novitas issued a draft LCD for oncology that primarily deferred coverage judgements to compendia like NCCN.  They finalized a far longer, more confusing version of the LCD this past June, withdrew it in July, and reissued it as "draft" in July.  It's accompanied by an extremely long and confusing billing article.   See my July blog about the "re-issue" here.

ACLA, in a press release, flags some of the public comments, including one by a multi-stakeholder group and one by ACLA itself.   Stakeholders quarrel with the deferral of coverage decisions (unless made in a one-off way for a few tests), to the NCCN and other guidelines.

See the ACLA press release here.  It has a number of links in it. These include the multi stakholder letter, the ACLA letter, and its supplement.

The comment letters are worth reading but frankly, I'm surprised they all just take the very long, verbose, repetitive, and confusing policy and article "on its face" and try to respond rationally to it.

For the AMP response, not yet posted, check here.

##

Open Meeting Transcripts Available.

See the LCD open meeting page here.  See the transcript (and audio) listing here. 30,000 words.

##

NCCN Compendia for Drugs - As a Model

If you wanted to use compendia as a guide to coverage (more a floor than a ceiling, though), there's an excellent precedent.   SSA 1861(t)(2) refers to cancer compendia, and CMS rulemaking provides process guidance.  See the CMS webpage here.  The regulatory discussion, in 2009, was just a few pages long here, The process in regulation at 414.930, is here, just a few paragraphs long.  They're written in a few paragraphs of normal, clear professional English.    (For NCCN's request letter to be endorsed, 10pp, here. For CMS's review of it, here.)

AI Corner #1.

Here is a summary by ChatGPT of the multi-stakeholder letter.

The letter, dated September 8, 2023, is addressed to Dr. Patrick Mann and Dr. Alicia Campbell regarding concerns about the draft local coverage determinations (LCDs) for "Genetic Testing for Oncology" by Novitas and First Coast Service Options. The letter highlights fears that the LCDs, if finalized as drafted, would limit Medicare beneficiaries' access to genetic testing for cancer, potentially violating Social Security Act and Medicare Program Integrity Manual requirements. Concerns include default non-coverage for tests not in knowledgebases, ICD-10 code issues, documentation requests, and restrictions on hereditary cancer syndrome testing. 

The letter requests collaboration with stakeholders to address these concerns and ensure access to essential genetic tests for cancer diagnosis and management. Signed by numerous healthcare organizations.

AI Corner #2

A summary by ChatGPT in the style of Lewis Black.  Here.

Brief Blog: Medicare Fiasco News: Patient Bowled Down by Obscure SAD List

BULLET.  Patient trapped with sudden $176,000 drug bill due to an obscure Medicare MAC policy change. +Lawsuit.

####

Background

Medicare policy makes extensive use of "incident to" services.   For example, there is a benefit for durable medical equipment like drug pumps and the drugs they pump are actually secondary tag-alongs to the DME benefit for the pump.  (Weird!).   Drugs that are administered in physician offices, like chemotherapy infusions, are covered "incident to" the service of a physician.

There are some rules around this.   The drug must be NOT self-administered HALF the time or more.   MACs are required to keep and update lists of injectible drugs that are NOT self administered, and thus eligible for office payment.   CMS defines this across all patients; if one patient is quadriplegic, for example, that doesn't matter if 51% of all patients self administer the drug (such as insulin).

MACs don't always agree (SAD lists can differ) and big debates sometimes occur.  See a sample "excluded" list here.

Today's News

MEDPAGE TODAY has a detailed article (by Cheryl Clark) about a patient who faces a major crisis because his drug was switch to "self administered, not payable in office" status by the MAC where he and his doctor live.   (Article may require email registration).  He got a bill for $176,000, whereas up to that point, his 80% payments were covered by CMS and his 20% payments were covered by his medigap plan.  A lawsuit by the Center for Medicare Advocacy is in flight.

Find the article here:

https://www.medpagetoday.com/special-reports/exclusives/106338

###

Another example of a 50% rule separates physician services and other services.   Medicare classifies a physician service IF AND ONLY IF a physician signs performs the service more than 50% of the time.   (42 CFR 415.102 . The rule doesn't state 51% numerically but this is how CMS implements the word "ordinarily" done by a physician.)   

The pathology rule is even tougher, the test/service must REQUIRE (quote - unquote) the service of a pathologist (415.130).  Back in 2012/2013, some stakeholders wanted the then-new genetic test codes to be on the physician fee schedule, but CMS determine that genetic tests did not "require" physician signout (the lab director can be a PhD).  (My 2012 white paper still downloadable here.)

###

Article summary.

Medicare unexpectedly changed its policy on the drug Stelara, classifying it as "self-administered" (SAD) on October 15, 2021. This led to retirees like George Beitzel, suffering from Crohn's and Parkinson's diseases, facing unexpected bills of up to $176,000 for previously covered injections. The Center for Medicare Advocacy filed a class-action lawsuit against this policy shift, arguing for notice, cost waivers, and professional administration options for patients unable to self-administer. Thousands of Medicare beneficiaries may be affected, raising concerns about the impact of such changes on patients' health and financial well-being.

As Haiku:

Medicare's surprise,
Stelara reclassified,
Burdens patients' lives.

##

AI Corner.

I fed a CMS MAC SAD Article to Chat GPT and asked it to figure it out and explain it.  Here.

Very Brief Blog: MAC CAC on BOTULINUM

 While we think of botulinum toxin first for cosmetic uses, it has a range of medical useless as well, such as in neuromuscular conditions.  E.g. blepharospasm.  An example of a MAC LCD is here.

The several MACs, all of them, have a "multi jurisdictional" CAC or public expert advisors meeting on botulinum, coming up on Thusday, October 19.  

FCSO link here (entry point).

NGS MAC runs the show here.

A 17-page background and question list is here:

https://www.ngsmedicare.com/documents/d/ngs/2118_0923_questions_for_botulinum_toxins_sme_panel_508-pdf

Click to enlarge.  Sample.

Very Brief Blog: Interesting Posts from Center for Genomic Interpretation (CGI)

While I don't agree with every position that the CGI takes (Center for Genomic Interpretation), they held my attention for a half hour today looking through their blog posts and Linked In posts from recent months.   

Find their blog posts here:

https://www.genomicinterpretation.org/blog/

And find their Linked In article feed here:

https://www.linkedin.com/company/center-for-genomic-interpretation/posts/?feedView=all

For example, a recent blog here points to a peer reviewed article here.  This links to an August 2023 paper, Patel et al., titled, Genomic Data Heterogeneity across Molecular Diagnostic Laboratories: A Real-World Connect Myeloid Disease Registry Perspective on Variabilities in Genomic Assay Methodology and Reporting.

For another recent example, they linked to the CLIA Advisory Committee CLIAC with regard to an August workgroup on NGS in CLIA labs.   Find the blog here.   You can link through to a 29 page CLIAC summary document here.  Or the direct PDF at CDC here.

See a CGI blog several months ago on what they see as weak points of flaws in recent state level biomarker legislation - here.

Fun fact - the laws vary across the 10 states where they have been passed.  There's a website that details the differences, but not that the table is *very* large and long and hard to read.  (I didn't even realize the table was there for a couple minutes, scroll down and down.)

See a link to Pfeifer et al. 2022 who discuss reference samples for inter lab comparisons of NGS, here.

Multiple Articles: Concerns about Quest LDT Alzheimer Blood Test?

FDA-approved Amyvid (PET Scan) advertised next to LDT article

Multiple news articles this week quote experts as expressing concern about an LDT (lab-developed test) being marketed as an Alzheimer diagnostic.   The test is a blood test; the only FDA-validated tests for Alzheimer's so far have been PET scans and CSF tests.

See an article from Reuters here.  With remarks such as, "Dr. Sarah Kremen, a neurologist at Cedars-Sinai in Los Angeles, was concerned that people who test positive but have no symptoms will come in seeking further testing."

See an article at MedPage here.  With remarks such as, "There are no large-scale, long-term clinical trials that support the idea that the AD-Detect test can predict whether a cognitively unimpaired person will transition to cognitively impaired," said Rebecca Edelmayer, PhD, senior director of scientific engagement at the Alzheimer's Association in Chicago. "As a result, it is unclear what the results of this test may mean about your Alzheimer's risk or your health status/"

The longest article, by Adam Bonislawski, is by subscription at 360Dx.  With remarks such as, "mass spec-based assays for Aβ 42/40 ratio have higher performance ... than do immunoassays, but "you have to be very rigorous about" your measurements, said Suzanne Schindler, associate professor of neurology at the Washington University School of Medicine. "There's only about a 10 percent difference between positives and negatives, and so if you are off by a little bit or your assay drifts, then you can really misclassify a lot of people," she said, adding that Quest has released little data on the analytical and clinical performance of its test. [The article continues, "At the 2022 Alzheimer's Association International Conference, Quest presented a poster on AD-Detect, and said it aims to publish data on the test in a peer-reviewed publication."]

##

The test is NY State approved; search for analyte amyloid and facility Quest, here.

The test is stated in an article above to be 71% specific, which could mean about 1 in 3 would get a false positive, although the negative and positive predictive values are highly dependent on the test population (population spectrum).   One historical problem - not necessarily relevant here - is Alzheimer tests over the decades that were validated on 100 perfect controls and 100 perfect Alzheimer cases, but then perform much worse on the real world and borderline patients who actually need testing.  

##

AI Corner:

ChatGPT reviews the three articles, then describes them in the satirical voice of Louis Black, here.

Journal Club: (1) Scenarios for Whole Genome Seq Adoption, (2) Scenarios for Any MedTech Adoption

This week I ran across two excellent papers on medtech adoption.  

One is recent, by van de Ven 2021, on adoption of whole genome sequencing.  The other (from a citation within Ven) turns out to be a classic, with 1200 citations to it, which is Greenhalgh 2017 on principles for the adoption of any med tech.

van de Ven 2021: 

Whole genome sequencing in oncology: Using scenario drafting to explore future developments.  BMC Cancer 2021.

The Van de Ven paper explores the use of scenario drafting and expert elicitation to anticipate future developments in the implementation of Whole Genome Sequencing (WGS) in clinical oncology. 

It identifies potential barriers and facilitators, highlighting the importance of factors such as price, clinical utility, and turnaround time in determining the likelihood of WGS adoption, offering valuable insights for policymakers and stakeholders in genomics.

Open access here.

For further reading see Ellis 2023 or Fleck 2023.

Greenhalgh 2017:

Beyond Adoption:  A new framework for theorizing and evaluating non adoption, abandonment, and challenges to scale-up, spread, and sustainability of healthcare technologies.  J Med Internet Res 2017.

The Greenhalgh paper introduces the NASSS framework, which offers a comprehensive and nuanced approach to understanding the adoption, scale-up, spread, and sustainability of complex medical technologies. The authors emphasize the importance of recognizing the non-linear and context-dependent nature of technology adoption and highlight the need to address factors related to the technology, the organization, the wider system, and the individual in healthcare settings.

Open access here.

For more about NASSS, see an implementation for a cardio health tracking device, open access, Abimbola 2019, here.  See a shorter overview of NASSS ("Cliff notes"), Greenhalgh & Abimbola 2019, here.  See a 2020 paper that introduces a "NASSS Toolkit" here.

Bonus 1:

There's a new 2023 paper by Greenhalgh et al. on the complexity of figuring out differences in values, among stakeholders in healthcare.   Find it open access at Milbank Quarterly, here.    

Bonus 2:

Greenhalgh et al. 2023 writes a letter to the editor of Annals of Internal Medicine about a review of N95 masks.   It contains this great quote:  

• "Some members of the evidence-based medicine community seem to assume that a randomized controlled trial, however imperfectly and illogically designed, is necessarily superior to other forms of evidence. This is not the case."
• Find it open access here.
• See a lively video by Greenhalgh on flaws of evidence based medicine here.

AI Corner x 3

Find a ChatGPT discussion of Greenhalgh 2017, here.  For example, Chat applies the framework to genomics, and to digital pathology.

Find a ChatGPT discussion of van de Ven 2021, futurism in WGS, here.   

Find a ChatGPT discussion of Greenhalgh 2023, on complex values in healthcare, here.

Stranger Things: USPSTF Launches, Then Halts, Lynch Syndrome Guidance

Currently, at the USPSTF website, there's an announcement that they launched an evidence review plan for Lynch Syndrome in February 2023.  Lynch Syndrome is a range of genes related to colorectal and some other cancers.   USPSTF here:

https://www.uspreventiveservicestaskforce.org/uspstf/announcements/public-comment-draft-research-plan-prevention-lynch-syndrome-related-cancer

FORCE, Facing Our Risk of Cancer Empowered, representing familial cancer stakeholders, just announced the USPSTF Lynch effort is nixed.

https://www.facingourrisk.org/privacy-policy-legal/advocacy/force-comments-of-new-colorectal-cancer-screening-guidelines-asks-for-guidance-on-identifying-high-risk-patients   

At this website, they have an article describing a two-year process of getting USPSTF to take on Lynch as a topic.   

Newly, at the top of this webpage, FORCE has a notice that USPSTF on August 30 notified some stakeholders like FORCE that Lynch review "cannot move directly" forward at this time.  FORCE speculates this could be due to under-staffing but no cause is provided by USPSTF.

Brief Blog: CMS Updates NCD Wait List

 A few weeks ago, I gave a talk CMS's plans for "TCET" Transitional Coverage for Emerging Technology.  I noted that even if you think it sounds good, CMS has a record of underperforming regarding coverage innovations.  I listed three things -

1.  Parallel Review was announced with great fanfare and gusto, but was used ultra rarely.

2.  Coverage with Evidence Development is perennially praised, yet its accomplishments are few (not zero, but few).  (CED is a pillar of the proposed TCET).

3.  CMS some years ago announced a national NCD dashboard, but updated it once or twice and forgot about it (like a child or puppy wandering to a different room and leaving a forlorn toy behind.)  Called the NCD Dashboard (on this page), until a few days ago, its last annual update (so called) was 2019 or so.

It's Updated

CMS has updated the NCD Dashboard on August 23, 2023.  It's here:

https://www.cms.gov/files/document/ncd-dashboard.pdf

Finals

In the past years, 2 NCDs were finalized, for cochlear implantation and for seat elevation (both are related to benefit category issues, too.)   

In Flight

Four NCDs are in process right now.  The Beta Amyloid PET NCD is proposed for deletion.  A USPSTF prevention benefit is being created as an NCD - PREV therapy.  There is an umpteenth update of a carotid stenting NCD.  Then there's the fairly recent opening of the stem cell transplantation NCD, opened June 7, draft decision expected December 7, 2023.  Myelodysplatic syndrome (MDS) is covered only CED, and organizations have asked the CED to be put to rest.  See the nine page request from ASH and others.

Note - MolDx recently covered whole genome sequencing for MDS therapeutics.  Blog.

Backlogged

Unlike finished or in-process NCDs, backlogged future topics are represented by a title phrase and not always penetrable.  Backlog topics include 1) "diaphgram pacemakers for neuromuscular disease [please non cover]", 2) hepatitis C screening [USPSTF endorsed 5/2020], 3) subcu insulin infusion pumps, 4) "power standing systems" (e.g. stand-up?), 5) pulmonary pressure sensor, 6) ventilators for COPD [ DME issue?], and finally 7) HPV for cervical screening.

The USPSTF rec for HPV and cervical screening is a few years old and being revised shortly.  I suspect there may be some revision issue, like coverage every 3 vs every 5 years.  Right now CMS covers annual Pap smears plus 1 HPV per 5 years on top of that.  USPSTF is somewhat laisse faire, allowing any of the following:  1) annual Pap, 2) HPV q5y, or 3) both.

CMS has coverage for insulin pumps now, and they must be subcu (not IV or IM), so there is likely some revision in play on that one, too.  

##

TCET Fun Fact

The comment period for TCET closed in late August.  As of September 7, 153 comments were received but 149 are still on hold and not posted yet. 

Very Brief Blog: Guardant Health Publishes 165 page deck for Investor Day

Guardant Health published a 165 page deck, in connection with its 9/7/2023 investor day.  Find a webcast, and the PDF deck, online at their investor relations page.

https://investors.guardanthealth.com/events-and-presentations/events/event-details/2023/Guardant-Health-Investor-Day-2023/default.aspx 

https://s26.q4cdn.com/594050615/files/doc_presentations/2023/09/GH-2023-Investor-Day-Deck.pdf

AI Corner: Large Study of AI Test Performance, 32 Courses, Compared to Humans

In a large scale study published in the Nature family journal Scientific Reports, essay question grades for ChatGPT were compared to human subjects in 32 classes spread over 8 subject matter areas.   For example, in the Psychology course "Biopsychology," ChatGPT outperformed, while in the Psychology course "Social Psychology," it did slightly worse.   In the chart below, the AI score is shown in green, the human score in blue.

click to enlarge

Find Ibrahim et al. here.

See a news report about the study here.

In each topic area, three real student answers and three ChatGPT answers (total of 6 answers) to each of 10 questions were graded by three graders.   (The inter-rater reliability of grading the 6 answers, varied by subject area).

__

The AI classification programs GPTZero and AI Text Classifier were quite imperfect at detecting which answers were AI generated and which were student-written, making many errors in both directions.  This fell even further when AI answers were "processed" through a rephrasing program called Quillbot. 

___

Some of the computer/math classes required math or coding answers only, not essays. 

Org Chart: Dora Hughes MD as Chief Medical Officer of CMS

For the last several years, Lee Fleischer MD has been Chief Medical Officer of CMS, and head of the offer of Clinical Standards and Quality.  He stepped down on July 31, 2023.  A few of us were lucky to hear him review his time at CMS at the NextGendx conference in DC in mid-August.

The Org Chart shows a new Chief Medical Officer today - Dora Hughes MD.   Dora Hughes had been Chief Medical Officer within the Center for Medicare and Medicaid Innovation (CMMI or "Innovation Center."   The online CMS org chart now shows Dr Hughes at the top of the Clinical Standard and Quality pyramid, with the joint titles of CMO and Director.

https://beta.cms.gov/about-cms/agency-information/cmsleadership/downloads/cms_organizational_chart.pdf

Dr. Hughes was an assistant to Senator Obama prior to his election to President in 2008.  She's had a number of private and non-profit and government roles since then.  

click to enlarge

Note that Tamara Syrek-Jensen, long-time head of the Coverage group, is now acting deputy director of CCSQ.   I've been told that long-time CMS physician Dr. Joe Chin is or will be director of the Coverage group, although this org chart version doesn't show that update yet.   

In addition to her numerous career positions, Dr Hughes holds an MD from Harvard.

New Podcast Today: Perry Dimas Interviews Me on Genomics Policy Trends

New today, lab industry and commercialization expert Perry Dimas interviews me on trends in personalized medicine.

Find it at the podcast, "PROMISE OF PERSONALIZED MEDICINE," available on most podcast apps.   Direct links to Apple and Spotify are provided.  The podcast runs 20 minutes.

Apple       Spotify

I've clipped the episode description below.

Using genomics or other biomarkers to improve health care is the essence of Personalized Medicine according to today’s guest, Dr. Bruce Quinn. Moving forward with novel diagnostic testing could prove challenging for a variety of reasons, one of them being a disruption to the overall healthcare industry. If one test could replace multiple, costly tests, then the ROI for certain payors goes down.

Join us as we also discuss:

• The other challenges facing new tests
• Where test identifiers are heading
• The indirect interest of insurance sales

___

Dr. Quinn is the Principal at Bruce Quinn Associates LLC and is an expert in Medicare policy for innovative technology.  His initial career was as a full-time medical school faculty member.  Armed with an MBA in 2001, he shifted to a career in strategy consulting.  He served as a regional Medicare Part B medical director 2004-2008.   He has worked for a global consulting firm, Accenture, as well as for two DC-based health policy firms.

Since 2016, he has been an independent consulting primarily focused on genomics and digital technologies.  His services include product planning and reimbursement pathways for innovators, as well as due diligence investigations for investors.   His website on health policy and new technology, “Discoveries in Health Policy,” has had over a million views, holds 2000 articles, and has hundreds of subscribers from industry, academia, and government.  Dr. Quinn is based in Los Angeles.

###

AI CORNER

An AI summary of the podcast.

In this two-part podcast discussion, Perry Dimas introduces the topic of personalized medicine and its challenges in the diagnostic industry, particularly regarding its adoption and integration with health plans and insurance systems. The conversation features Dr. Bruce Quinn, who shares insights into the field of personalized medicine and its recent advancements.

##

Dr. Bruce Quinn defines personalized medicine as the use of genomics and biomarkers to enhance healthcare.

He discusses advances in personalized medicine over the last decade, focusing on cancer oncogenes and NIPT (Non-Invasive Prenatal Testing) as impactful areas.

The challenges in the personalized medicine industry are outlined, including its capital-intensive nature, extended timelines for adoption, and the importance of defining clear use cases.

Dr. Quinn introduces the concept of the "valley of death" between Medicare coverage and widespread commercial adoption.

The conversation touches on how health plans should collaborate with diagnostics companies and highlights the need for clear, understandable language.

Dr. Quinn emphasizes the importance of bridging the gap between industry and health plans for successful integration.

##

The discussion continues with a focus on test identifiers and their role in personalized medicine.

Dr. Quinn discusses the limitations of current test identifiers like CPT codes and Z codes, highlighting the need for a comprehensive patient profile.

They mention the challenges health plans face in dealing with the overwhelming influx of new diagnostics and the complexity of making high-quality decisions.

Dr. Quinn emphasizes the unique nature of selling diagnostic tests to health plans compared to traditional B2C or B2B sales.

The conversation explores the importance of educating health plans and building a mutual understanding of their needs and timelines.

Dr. Quinn expresses appreciation for the podcast's role in shedding light on the personalized medicine industry and helping it grow.

The podcast aims to educate listeners about the complexities and challenges of personalized medicine, especially regarding its integration into the healthcare system, health plans, and insurance coverage. It highlights the need for clear communication, collaboration between industry and health plans, and a comprehensive approach to diagnostic test identification and adoption.

AI Corner: Digital Pathology Orlando (October 29-31): AI Reviews the Whole Two Days of Programming

Last week, I highlighted the upcoming Digital Pathology conference in Orlando (October 29-31, 2023), and I included a summary of the conference by Chat GPT based on the presentation titles.

https://www.discoveriesinhealthpolicy.com/2023/08/meeting-note-digital-pathology.html

https://digitalpathologyassociation.org/pathology-visions-conference

https://digitalpathologyassociation.org/pathology-visions-agenda

(See a review of last year's conference that appeared last December in CAP TODAY.)

Today, I revisited the same approach, giving data to Chat GPT, but rather than just use titles, I copied all of the 30 or 40 abstracts of the presentations, and fed that larger data into Chat GPT.  

Here's the resulting Chat article, in the voice of a journalist who's attended the [future] dig path conference.

###

Please write an article about the conference, based on the agenda and abstracts, in the voice of a science and med tech journalist.  

Then, AI summarizes the meeting's AI-driven presentations, and AI summarizes any worries or drawbacks in the field.

Advancing Pathology Through AI: 

Revolutionizing Diagnostic Medicine

[Fictional Summary of October 29-31, 2023, Conference]

By [Science Journalist]

In a groundbreaking conference that brought together the finest minds in pathology and artificial intelligence, the future of diagnostic medicine was unveiled in all its complexity and promise. This prestigious gathering, hosted by renowned institutions, offered a comprehensive exploration of the incredible strides made in digitizing pathology and harnessing the potential of AI to reshape the landscape of healthcare. Over the course of several days, participants engaged in discussions, shared groundbreaking research findings, and exchanged invaluable insights that collectively painted a vivid picture of the current state and exciting prospects of the field.

Unleashing AI's Potential in Pathology

Perhaps the most captivating revelation from the conference was the formidable power of artificial intelligence in transforming pathology. A notable focus was on AI's capacity to elevate the accuracy and efficiency of diagnostics, holding the potential to redefine how pathologists make critical decisions. Through demonstrations and case studies, the integration of AI algorithms into digital pathology platforms was showcased as a means to streamline workflows, reduce human error, and provide insights that could lead to earlier and more precise disease detection.

The conference spotlighted ongoing efforts to develop AI algorithms that can support primary diagnoses, identify specific histological features with remarkable accuracy, and even predict molecular profiles from the intricate details of microscopic images. The profound implications of AI's ability to analyze vast datasets swiftly and uncover patterns that might elude human observation underscored its role in driving advancements in personalized medicine and enhancing patient outcomes.

Embracing Collaborative Data Efforts

A recurring theme throughout the conference was the paramount importance of amassing large and diverse datasets. These datasets, composed of digitized whole slide images (WSIs) coupled with associated metadata, are pivotal in training AI models to perform at their best. The Bigpicture project, a significant initiative highlighted during the event, was a shining example of the collaborative approach adopted to address this challenge. With the ambitious aim of establishing an EU-wide repository of WSIs, the project seeks to foster an ecosystem that encourages cooperation among various stakeholders, including academic institutions, pharmaceutical companies, and regulatory bodies. This collaborative model is set to accelerate the development of AI solutions while navigating complex ethical and legal considerations.

Navigating Challenges and Concerns

While the conference showcased the remarkable advancements AI brings to pathology, it also brought into focus the challenges and concerns intertwined with these transformations. Quality control, the generalizability of AI models, data privacy, and regulatory hurdles all came under scrutiny. Ensuring AI models are trained on comprehensive datasets reflective of the diverse scenarios encountered in clinical practice emerged as a central concern. Ethical questions surrounding the use of patient data for AI training were deliberated extensively, with experts advocating for transparency and stringent patient consent protocols.

The Promise of the Future

As the conference drew to a close, the prevailing sentiment was one of boundless optimism and anticipation. The convergence of digital pathology and AI promises to redefine diagnostic practices, empower healthcare professionals, and ultimately improve patient care. Innovations such as cloud-based digital pathology repositories are poised to reshape the way researchers and clinicians collaborate, unlocking the full potential of AI on a global scale.

This conference offered a tantalizing glimpse into a future where AI algorithms work harmoniously with pathologists, offering rapid and precise diagnoses. From immediate image-based molecular profiling to improved statistical benchmarking of AI models, attendees were treated to a showcase of the remarkable advancements shaping the future of medical practice.

In an era where technology and healthcare intersect more profoundly than ever, the conference underscored the indispensability of collaboration, data sharing, and ethical considerations in ensuring AI-driven diagnostic tools not only enhance medical practice but also uphold the sanctity of patient care and privacy. As the field of pathology embarks on this transformative journey, it stands on the brink of revolutionizing how we approach diagnostics. The dawn of precision medicine is upon us, heralding an era where the synergy between AI and pathology will lead to better patient outcomes and a brighter future for healthcare.

###

###

Highlight some of the presentations that specifically deployed Artificial Intelligence technology into pathology.

ChatGPT

Presentations Paving the Path to AI-Enabled Pathology

Within the symphony of presentations that graced the conference stage, several standout sessions unveiled the cutting-edge applications of artificial intelligence (AI) technology in the realm of pathology. These presentations showcased the exciting convergence of advanced computing power and medical diagnostics, promising to revolutionize the field and redefine patient care.

1. Immediate Image-Based Molecular Profiling

One of the most captivating presentations delved into the realm of immediate image-based molecular profiling. Traditional molecular profiling methods are time-consuming, often taking weeks to yield results. This presentation, however, unveiled an innovative approach where AI algorithms rapidly analyze digitized histopathological images to provide complementary molecular profiling information within minutes. By harnessing the power of AI, pathologists are now armed with vital molecular insights, enabling faster and more informed treatment decisions for cancer patients.

2. Cloud-Based Digital Pathology Repositories

The concept of cloud-based repositories took center stage in a presentation that demonstrated how biorepositories and institutional databanks are being transformed by AI and digital pathology. In this innovative approach, an elastic compute cloud-based infrastructure was created, allowing for the sharing of vast amounts of whole slide images (WSIs) and associated metadata. Researchers gained access to deep search capabilities, enabling them to mine data for relevant information, annotations, and even the deployment of AI modules. This presentation heralded a new era of collaborative research, where AI's analytical prowess is seamlessly integrated into the data-sharing ecosystem.

3. Enhanced Statistical Benchmarking of AI Models

An intricate challenge in AI-enabled pathology is accurately benchmarking the performance of AI models relative to human pathologists. A presentation addressed this challenge head-on by introducing a novel nested pairwise frames evaluation method. This method, designed to ensure fairness in comparing AI models and human pathologists, generated impressive results. By creating an evaluation framework where AI models are treated equivalently to pathologists, the presentation provided a standardized approach to assess AI's capabilities in detecting tissue and cell types within histopathological images.

4. The European Bigpicture Project

The conference also provided a platform to unveil the ambitious European Bigpicture project. This initiative is committed to establishing a comprehensive repository of WSIs and associated metadata from across the European Union. The project aims to overcome challenges in AI development by pooling data resources from diverse sources. The European Bigpicture project exemplifies a collaborative endeavor that endeavors to create an environment where ethical, legal, and technical considerations harmonize, paving the way for the development and adoption of innovative AI-driven pathology diagnostics.

5. Advancements in AI-Powered Diagnostic Algorithms

A presentation highlighted the ongoing strides in developing AI algorithms specifically designed to enhance diagnostic accuracy and efficiency. These algorithms, designed for primary diagnosis and precise feature detection, are capable of outperforming traditional methods and even matching human experts in some cases. By harnessing AI's computational capabilities, pathologists can confidently rely on these algorithms to provide robust insights, streamlining diagnoses and allowing for faster patient care decisions.

Collectively, these presentations offered a panoramic view of AI's remarkable potential in pathology. From expediting molecular profiling to fostering international collaborations through cloud-based repositories, the conference showcased a myriad of ways AI is reshaping how pathology operates. As these pioneering advancements gain momentum, the medical community is poised to witness the birth of a new era in healthcare, where AI and pathology seamlessly intertwine to improve patient outcomes and redefine the very nature of medical diagnosis.

What were the most worrisome concerns of the participants, or also, any skeptical voices amongst the presenters.

ChatGPT

Navigating Concerns and Skepticism in the AI-Enabled Pathology Landscape

Amidst the palpable excitement and groundbreaking innovations at the conference, participants and presenters also voiced their concerns and skepticism, acknowledging the complex challenges that come hand in hand with integrating AI into the field of pathology.

1. Generalizability and Quality of AI Solutions

A recurring concern resonating among participants was the quality and generalizability of AI solutions. While AI has shown remarkable potential in revolutionizing pathology, the worry is that AI algorithms might not consistently perform as desired. The challenge of limited generalizability was brought to the forefront, with participants acknowledging that AI's efficacy might be confined to data it was originally developed with. This sparked discussions on the importance of large, diverse datasets for AI training to ensure robustness across various scenarios.

2. Technical and Ethical Hurdles in Data Collection

Skeptical voices arose when grappling with the technical, ethical, and legal complexities of amassing extensive datasets required for AI development. Presenters highlighted the formidable challenge of collecting diverse and comprehensive datasets from different sources while adhering to ethical guidelines. Concerns about data privacy, security, and patient consent underscored the urgency of establishing transparent practices for data collection, usage, and sharing.

3. The "Black Box" Phenomenon

Another concern that echoed throughout the conference was the "black box" phenomenon inherent in many AI algorithms. While AI can yield remarkable results, the intricate inner workings of these algorithms often remain inscrutable to human understanding. This opacity raises concerns about the transparency and interpretability of AI-driven diagnoses. Participants and presenters alike pondered how to bridge the gap between AI's advanced computational capabilities and the necessity for interpretable results, especially in a medical context where lives are at stake.

4. Integrating AI into Clinical Workflows

Skepticism emerged around the practical integration of AI into clinical workflows. Presenters expressed concerns about how to seamlessly blend AI-driven insights with the expertise of human pathologists. The challenge lies not only in the technological aspects but also in fostering a harmonious collaboration between AI systems and healthcare professionals. Overcoming resistance to change and ensuring a smooth transition were highlighted as pivotal steps towards the successful adoption of AI in pathology.

5. Regulatory and Standardization Hurdles

The complexities of navigating regulatory frameworks and establishing standardized protocols for AI-powered pathology garnered substantial attention. Skepticism surrounded the challenges of aligning AI practices with established medical standards and gaining regulatory approvals. The uncertainty surrounding the regulatory landscape raised questions about how to ensure patient safety, data integrity, and consistent quality across AI-driven diagnoses.

In conclusion, while the conference celebrated the tremendous strides AI has taken in transforming pathology, it also served as a platform to acknowledge the underlying concerns and skepticism. Participants and presenters recognized the imperative to address these challenges head-on to ensure that the integration of AI into pathology is not only groundbreaking but also ethical, transparent, and clinically effective. The symphony of optimism and caution reverberated throughout the conference, signifying a collective commitment to harness AI's potential while navigating the intricate nuances of the healthcare landscape.

##

Summarize in a haiku.

AI's promise gleams,
Pathology's digital dawn,
Challenges embraced.

September 7: Roche Webinar on, Digital Transformation of Patholoogy

 At 12 ET, 9 PT on September 7, 2023: See a Roche webinar on "The Digital Transformation of Pathology - How computational pathology can impact patient care."

Sign up here:

https://lableaders.roche.com/global/en/c/webinar-digital-pathology.html

The organizers write,

"Technological advances in digital and computational pathology open new avenues for personalized healthcare solutions that address the unmet medical needs of patients. By harnessing the power of imaging technology and artificial intelligence (AI), pathology is moving from qualitative descriptions to quantitative measurements and predictions, bringing about a new era of precision medicine. 

Join our panel of experts as they discuss these new advances, why pathologists should embrace AI-based tools, and how by leveraging these technologies, we can improve workflows and add medical value."

The speakers include a panel of 4 leaders for digital pathology at Roche.

Busy Day at MOLDX: New Draft LCD for Prostate GEP Tests; For Lung Nodules; For Thyroid Nodules; New Final LCD For R.A. Biomarkers

Busy day at MOLDX with three new draft LCDs and one new final LCD.

NEW DRAFT LCDS

DRAFT - ADVANCED PROSTATE

See DL39636, Gene expression profile (GEP) tests for decision making in castration resistant and metastatic prostate cancers.  

MolDx has a long history of LCDs for prostate cancer prognostic tests, this appears to be a new version (not style as a revision).   Prior versions focused on early diagnosis and management.  This LCD is based on a February 2022 letter from Dr Davincioni of Decipher/Veracyte.  MolDx publishes only the cover letter of the request.

The LCD notes it does not deal with CGP (e.g Foundation Medicine type tests) or "single biomarker" tests.  Patients must be facing multiple therapy optoins of variable severity and the test will help make that choice.

The LCD does not named covered or non covered tests but the LCD has an associated billing article DA59462.  This lists only 81479 as a potentially covered code, so that's kind of a dead end for policy wonks trying to determine  or understand coverage (e.g. by doing studies on payer coverage and publishing them.)

On a commerical note, the Decipher Veracyte prostate GEP test has been taking a larger share of a market that once was divided in thirds (Veracyte, Genomic Health, Myriad).  A Veracyte Decipher collaboration with Duke was just covered in Genomeweb based on a publication in JCO Precision Oncology by Bitting et al. 

DRAFT LUNG NODULES

DL39654, Molecular biomarkers for risk stratification of indeterminate pulmonary nodules following bronchoscopy. 

Note that this isn't framed as a policy for imaging-based nodules, but for nodules "following bronchoscopy."  While this sounds like, for example, a Veracyte Percepta test, the LCD is framed as "internally initiated."  The rules confirm this is for post bronchoscopy patients.  The LCD separately discusses "first generation" and "second generation" Percepta tests.  

The billing article DA59476 again presents itself as "information that complements the LCD" but simply lists 81479 as a possible billing code. 

For the use case you might have thought of - a mystery small nodule on low dose CT screening or an incidental small lung nodule - and it's before bronchoscopy or PET - Medicare has an LCD for "XL2" test from Biodesix, here. 

DRAFT THYROID NODULES

LCD DL39646 is "Risk Stratification for Thyroid Nodules" and is based on a letter from Veracyte (Dr Gerrard)  from January 2022.  The letter notes that coverage was previously requested for "Veracyte MTC" and that the January 2022 update restyles the request in the format of a foundational (pan-product) LCD.  Coverage is valid for a patient with an indeterminate nodule Bethesda III or IV, or else, a nodule Bethesda V "for which molecular testing may aid in further decision making."

The LCD concludes, Given the cost and risks associated with thyroid surgery, the clinical decision making that goes into the extent of surgery, additional follow up management, and published guidelines, this contractor finds that molecular tests that aid in medical decision making for thyroid nodules of Bethesda Categories III-V are reasonable and necessary. Specifically, this contractor recognizes improved outcomes by safely precluding unnecessary surgical procedures with these services as well as better selecting patients who may benefit from such procedures. 

Reference to specific tests in this document does not imply coverage by this contractor.

Readers unfamiliar with "foundational LCDs" may be puzzled by the remark that evaluation of a specific test in an LCD "does not imply coverage."

A billing article is provided, DA59470.  Potentially covered codes include 81479, again, as well as specific code 81546.  Veracyte GSC Genomic Sequencing Classifier. 

NEW FINAL LCD: FLIP TO COVERAGE

FINAL RHEUMATOID ARTHRITIS

LCD L39424.  This is based on a letter from SCIPHER, making PRISM-RA,   The request letter is undated and the PDF has a document creation of 7/2022, so it might have been the date preceding the October 2022 draft LCD.  While it's an inferenec, SCIPHER's approach for an LCD likely preceded the CAC in December 2021.

The original Draft LCD position on non coverage, in October 2022, was curt. It followed a CAC meeting in December 2021.

• "Current molecular biomarker tests to guide targeted therapy selection in Rheumatoid Arthritis (RA) are non-covered by this contractor."

Champagne corks are popping somewhere because the final LCD provides "limited coverage" based on "new evidence" and following 12 rules.   The patient has a history of failure or contraindication to at least one first line DMARD and is considering a range of alternate therapies (TNF, JAK, etc).  And MolDx has determined the test to be appropriately validated and designed, etc.

The billing article - again - simply points to a possible unlisted code, in this case, 81599, other MAAA test.  See A59529.    

With the flip in coverage status, there is a lengthy "response to comments" document A59519. I extract and highlight the specific rationale for the favorable policy change, and critiques of the field that MolDx still feels are valid.  Sidebar here.  (There, I also ask ChatGPT for its opinion on the policy flip.)

MY COMMENT

IImproved Coordination & Timing for Comment

This crop of LCDs, draft and final, had better coordination of posting, with the same timing at the several MolDx MACs - which will help speed up finalization of those in draft.   Draft LCD have comment from August 31 to October 14.

For two dated LCD request letters, the time from letter to a draft LCD was 6-9 months.  The time from draft to final LCD, for the one final LCD, was 10 months.

Convergent Evolution of LCD Text

The LCDs for indeterminate nodules (lung, thyroid, etc) tend to consolidate toward very similar language in terms of coverage and performance rules.  Eg. rules on the lines of, "the test must be appropriate designed for the population."

Stymies Academics and Publications

There's been excellent work in recent years documenting and publishing the nature of  payor policies, but policy experts with be stymied by the lack of specific coverage details either in LCDs or articles.  To some extent, the DEX database run by Palmetto can back fill this, but includes no coverage dates and no archival form (DEX lives in the eternal present moment only).   This limits the abilities of academic experts, or consultancies, to tabulate coverage issues, even if they are on grants commissioned by HHS.  See eg Trosman et al. 2020 PMID 32389219.

PAMA Thwarted?

The widespread use of "unlisted codes" puts whole fields of genomics outside the elaborately designed CMS and Congressional PAMA pricing system, even as genomics grows as a proportion of lab spending.

AI Corner

Chat GPT 3.5 rewrites this article to 650 words, shortened from 1050.

Here.

I ask GPT 3.5 about the R.A. policy flip [lower part of the linked page]. Here.

GPT 3.5 summarizes the 4600 word R.A. LCD in 600 words.  Here.

MolDx Covers Another Transplant Rejection Test; This One Digital PCR (Oncocyte VitaGraft)

MolDx has issued coverage for the Oncocyte VitaGraft renal rejection test, per a press release here and brief coverage at 360DX here,   

The relevant organ rejection LCD is a broad, or foundational LCD, that allows the potential for future coverage of molecular testing for graft rejection in any organ.   Actual coverage issued has included tests using donor DNA detection, from CareDx and Natera.  There have been some kerfuffles about the LCD, its textual analysis, and revisions.  

Price Pending

As of August 29, the Palmetto "DEX" database did not yet list the Vitagraft test as covered, and, it had no Medicare price list.   Updates will likely appear online over a week or two.

The DEX database refers to the "Allaro" Kidney test, whose description then calls itself the VitaGraft test [?].  One year share price is down from $18 to $3.50.  Market cap is $30M.  For 2022, "full-year 2022 revenues were $5.6 million, down 27 percent from the prior year.  Net loss for the year was $72.9 million." The company has bounced in and out of Nasdaq compliance.  The company has immuno-oncology test assets that are not yet covered by MolDx (Determa IO).  

ddPCR and MolDx

Although examples are still not too common, digital PCR has potential applications in graft rejection and minimal recurrent disease detection in oncology.  

click to enlarge

Large Health System Sues Anthem; Lawsuit PDF Provided

According to a colorful article in MedCityNews here, Bon Secours Mercy Health Virginia (BSMH Virginia) has filed a lawsuit against Anthem Health Plans of Virginia for over $93 million in unpaid and underpaid health claims. 

BSMH Virginia alleges that Anthem engages in delaying and denying payment tactics, impacting patient care. Anthem claims BSMH is using the lawsuit to distract from leaving Anthem's provider network.

MedCity provides a PDF of a 25 page legal filing with more colorful details.  Here.

###

AI Corner.  

ChatGPT provides Haiku Summaries:

Unpaid claims in court,
Anthem and BSMH clash,
Healthcare payments strained.

Lawsuit's bitter fight,
Anthem's tactics under fire,
Patients caught between.

Money owed, care stalled,
Legal clash of health titans,
Healing at a cost.

September 27: Roche Webinar on New Rules for Her2 Staining

 I recently ran a blog about new rules for "low Her2 staining," based in part on a long article in CAP TODAY, and on several Op Eds that were published along with the new rules.

On September 27, 2023, at 11 pacific 2 eastern, attend a webinar sponsored by ROCHE on the new rules.  Find more info here:

https://mailchi.mp/uscap/register-now-her2-ihc-testing-in-the-wake-of-the-2023-her2-breast-guideline-update?e=fdf9fc4ebd

Novitas MAC Publishes Article on "T Codes" for Slides Digitization

Header:  Novitas has published guidelines on the new AMA Category III codes for digitizing medical slides, stipulating special situations where they may be payable.  But I see a few policy bugs in the water.

For subscribers to Genomeweb, see a deep-dive article by Greg Cima on the same topic (including private payers and lab consultant interviews), August 29, 2023, here.

###

Last January, AMA CPT brought out about a dozen Category III codes, unpriced, for digitizing slides for medical purposes (not teaching or team conferences).  And there's more.  For January 2024, AMA has announced a few dozen more.  They are add-on codes that will eventually pair to nearly any code that generates a glass slide.

CMS has put the codes on "bundled/not separately paid" status for the OPPS setting. Ouch!  See my article about the nonpayment edits here.   (To my surprise, there wasn't even a short paragraph on this topic for these new codes in the July 2023 OPPS proposed rule.)

Novitas Article

Novitas publishes the first article about the codes.  It says that they are potentially payable IF REASONABLE AND NECESSARY.  They describe this pretty narrowly, giving the example of a glass slide that is unique and risky to mail, so it needs to be digitized and sent as whole slide imaging.  Another example is an emergency where sending the slide digitally is faster than Fed Ex.

Find the article here, and I'll clip it at the bottom of this blog as well.

https://www.novitas-solutions.com/webcenter/portal/MedicareJL/pagebyid?contentId=00278110   

Two Dilemmas for BQ

I see two "nerd dilemmas."  

One, while I'm an advocate of digital pathology, from a neutral view, it's not entirely clear why, if CMS would not pay $30 for Fed Ex, CMS would potentially pay $30 (or some other fee) for WSI.   One way out of this corner, is to point out that Fed Ex is not an FDA-approved device, whereas a WSI scanner is.

The other dilemma is that both clin chem and pathology have CMS dates of service rules that fix the DOS for 14 or 30 days as the date of specimen collection.  If that's in a hospital (inpatient or outpatient) the bundling rule against these codes may apply.   For example, with a hospital outpatient office visit, routine clin lab tests (but not genetics) are bundled, and they're bundled whether or not the clin chem is done at the hospital or off site.   By analogy, I think the OPPS no-pay bundle rule for the Cat III codes might apply even if the slide is coming from elsewhere as long as the date of service rule ties it to an inpatient or outpatient event.   Novitas' text doesn't dwell on this.  It focuses on a Pt B 1500 claim, not a hospital 1450 claim, but the bundle rules are established by specimen origin location and timing, not by the choice of claim form (wouldn't that be nice).   Avoiding the OPPS non payment policy, the Novitas article would be a good fit (for example) for an private office-based skin biopsy processed in a Part B lab. 

Actually, a bonus, third dilemma, is that these are styled as "add on codes," so they have to be on the same claim form on same day of service as the parent code (e.g. 88305).  That limits the usefulness, even under Novitas' own rules, when the special consult might come up a couple weeks after the 88305 event.

NOVITAS - June 29, 2023

Billing for digitization of glass microscope slides, CPT codes 0751T-0763T

Digital pathology is a practice in which glass microscope slides are digitally scanned by clinical staff. The digitized images are saved to a computer or cloud-based system, where they can further be viewed, analyzed, and shared with other providers. Providers can remotely access digitized pathological images and data for review and diagnostics. Digital pathology can enhance diagnostic efficiency, prognosis, and prediction of illness. These codes are used to report additional staff work and resource needs associated with the process of scanning and digitizing pathology content.

Effective for claims received on or after August 5, 2023

Providers shall provide a statement on Item 19 of the CMS-1500 claim form, or the electronic equivalent, to document how the digitization of glass microscope slides is medically reasonable and necessary. An example of a potentially reasonable and necessary service includes if a slide requires an outside consultation that would normally require mailing of the slide, digitization could potentially be medically reasonable and necessary if either the slide represents the only slide demonstrating the pathology of interest (and loss of the slide would mean loss of irreplaceable material) or the consultation is required urgently and mailing a slide would take too much time and the delay would put the patient's health at risk.

Failure to prove a statement on Item 19 of the CMS-1500 claim to document the need for the digitization of glass microscope slides may result in claims rejected as unprocessable.

##

Some MACs publish pricing for codes not priced by CMS.   I spotchecked this at Novitas and the Dig Path code I looked up was still $0, at least today.