Recent Articles:

EPtalk by Dr. Jayne 1/24/19

January 24, 2019 Dr. Jayne No Comments

I had a chance to sit in on a Medicare Chronic Care Management (CCM) webinar from Greenway Health while I was on site with a client today. It was nicely done, with a concise agenda and a strong presenter. The sales component was kept to a minimum, with the majority of time being spent educating attendees about chronic care management in general and only a small part was spent on the vendor’s solution. The webinar was also 30 minutes rather than the hour we sometimes see scheduled for these events, which made it easier for a busy practice to squeeze into their day. My client had a response to her follow up request within three hours of the end of the webinar, which was much appreciated.

While the US federal government shutdown rages on, there’s a bill pending in the Senate that would allow for importation of prescription drugs from Canada. When I first started in practice, we used to refer patients to an Ontario-based pharmacy to obtain a particular respiratory medication that had been proven effective around the world but wasn’t available in the US. For some, it’s not just about a potential cost savings, but about helping patients with access. It’s unlikely to get through the Senate, but it’s the thought that counts. Thanks to senators Chuck Grassley (R-IA) and Amy Klobuchar (D-MN) for giving it a bipartisan try.

The Food and Drug Administration is running out of money to fund reviews of new drugs, according to a Fortune piece online. During the shutdown, they’re not collecting user fees paid for drug and device reviews, so they’re trying to use leftover money from 2018 to keep the lights on. FDA Commissioner Scott Gottlieb plans to furlough more employees to keep things moving since drug reviewers aren’t allowed to work unpaid as they’re not considered “essential.” Gottlieb first caught my attention last month when he used Dr. Seuss-style poetic meter to warn against the hazards of eating raw cookie dough.


It’s barely more than two weeks until HIMSS and they’ve released the #HIMSS19 Hashtag Guide. There are more than 20 of them, which is about 19 more than I’m going to be able to keep track of unless they come as part of a snuggly quilt like this one. I’ve already had my quota of below-freezing days this year and there is no end in sight, so it’s definitely time to pile on the blankets.

It’s the last week to comment on ONC’s draft “Strategy on Reducing Regulatory and Administrative Burden Relating to the Use of Health IT and EHRs.” I hope they make the public comments available to all of us, because it will make for some excellent snow-day entertainment. You can comment through 11:59 p.m. on January 28. Get those creative juices flowing. Feel free to share any amusing entries with the rest of us while we wait for the full report.

I’m always excited to see startups or other vendors making it easier to diagnose and treat diseases, particularly ones that either (a) no one wants to talk about; or (b) those who want to talk about it only want to do so to make a political point. Nurx is best known as a telehealth company specializing in birth control pills for women, but now they’re moving into at-home testing for the Human Papilloma Virus (HPV). The test is limited to existing customers, but will be expanded to new customers in 2019. Commercial insurance provides good coverage, although there is a $15 shipping and processing fee. They also offer PrEp, a treatment that reduces the risk of acquiring HIV. These are areas where some of the more mainstream telehealth groups aren’t ready to tread, so hats off to them.

Speaking of telemedicine, the American Academy of Family Physicians (AAFP) blog “In the Trenches” says the Academy is “working to make sure telemedicine is real medicine.” There’s much greater insurance coverage for telehealth services of late, and most states require payment for certain services whether they’re delivered face-to-face or through telehealth technology. It sounds like their focus is on helping traditional family medicine practices to add telemedicine services rather than seeing how telehealth might fit into the broader scheme of things.

The blog author refers to solutions that compete with their ideal traditional practice as “leakage” and makes it clear that he thinks that patients using urgent cares, retail clinics, or direct-to-consumer telehealth services are part of the problem, specifically calling out vendor Teladoc and citing his own “Twitter rant” about its partnership with Aetna. He goes on to demonize telehealth providers as “Snapchat for antibiotics” and “basically Tinder for healthcare.” The gloves are off, apparently.

AAFP has a partnership with Zipnosis for telemedicine, but it is only available for “small, private practices with less than 20 physicians.” It doesn’t support billing when patients have telehealth coverage through insurance and doesn’t integrate with EHRs. They recommend physicians (or their administrative users) download a PDF of the visit (which is documented using a third-party platform) and update EHR medication and diagnosis lists manually. That’s not exactly revolutionary, and if AAFP is serious about having a great solution, I can recommend a physician informaticist who might be able to help.

Regardless of how much AAFP continues to preach about longitudinal relationships with primary care physicians, that scenario just isn’t the reality for many patients in the US. Insurance plans change, networks change, and then there’s the situation where people change jobs and get new insurance and have to switch physicians. Our culture has changed, and many patients don’t value that relationship with a physician because they’ve never experienced it or are turned off by the way that a lot of practices run. Convenience and accessibility trumps that relationship. People are willing to risk counterfeit products via online purchases rather than going to a brick-and-mortar store, and those kinds of sentiments spill over into other facets of our lives.

Not to mention that there are many physicians who have left traditional primary care practices because of lifestyle issues, relatively low compensation for long hours, and burnout. AAFP would do well to encourage physicians to explore practice situations outside the organization’s historical ideal in an effort to keep people working as physicians. As an urgent care physician, I’m part of their “problem,” but it’s gratifying to be the family physician (albeit temporary) for patients who can’t see their own. I’m starting to wish I hadn’t paid my dues for 2019, but I don’t think they give refunds. I haven’t seen that level of vitriol from professional organizations that represent pediatrics or internal medicine.

What do you think about the future of telehealth? Leave a comment or email me.


Email Dr. Jayne.

Morning Headlines 1/24/19

January 23, 2019 Headlines 2 Comments

Nextech Acquires SRS Health, a Specialty EMR Solution for Orthopedic Practices

Health IT company Nextech expands its ambulatory reach with the purchase of orthopedic EHR vendor SRS Health.

With an eye on pioneering the future of Health IT, Greenway Health appoints Kali Durgampudi as chief technology and innovation officer

Greenway Health names former Nuance executive Kali Durgampudi chief technology and innovation officer.

In a sign of its concerns about Atul Gawande’s new venture, Optum sues over trade secrets

Citing a direct threat to its trade secrets and confidential corporate information, Optum sues a former VP now employed by Amazon-Berkshire Hathaway-JPMorgan’s healthcare venture.

NantHealth Could Recover In 2019

Nasdaq gives NantHealth six months to increase its stock price to above $1 for 10 consecutive days, a 43-percent recovery that just might happen if the company’s clinical trials go well, and it receives FDA approval and increased CMS reimbursements for its GPS Cancer Test.

HIStalk Interviews Sean Carroll, CEO, Arcadia

January 23, 2019 Interviews No Comments

Sean Carroll is CEO of Arcadia of Burlington, MA.


Tell me about yourself and the company.

I am a 32-year veteran of healthcare IT over six companies, all focused on some aspect of healthcare data. Some of those companies were service-oriented businesses, some have been technology focused, and one or two have been a combination of both.

I’ve been on an explicit mission over my career to be part of tangible progress in the evolution of healthcare. I firmly believe that progress must be founded in healthcare data as at least as one of the major levers that has to be pulled. I’ve been at Arcadia for five and a half years. I’m glad to be here because Arcadia has the capabilities and positioning in the market to be an instrumental player in that necessary transformation of healthcare. We are seeing that with our customers.

The company is positioned as a population health management company, specializing in delivering population-level analytics in care management to enterprise-class customers. We serve enterprise provider organizations and nationally branded, recognized health plans. We also have a nice partnership business. Since we last talked in mid-2015, the company has experienced significant expansion and growth, elevating itself to 50 million patients and 50,000 providers in our platform. We have continued to maintain top rankings and positioning with all of the analysts who track this market since we emerged in the KLAS report in the end of 2016.

What changes have you seen since 2015 in population health management, value-based care, and the availability and harmonization of EHR data?

It has evolved a lot. One of the positive developments in the ultimate transformation of healthcare is that the commitment to value-based concepts has remained steadfast, even with all of the crosswinds that we all face every day politically and from other influential angles. That’s been a positive trend of staying in that direction.

The market has, without a doubt, matured in many ways in the last three to four years. Organizations that are pursuing technology or services that support value-based care transformation have probably been through one or two tries at implementing something. They have some battle scars and they’re smarter about what they want. Vendors who support them in this effort also have learned a lot about what it takes to be an excellent provider of technology or value-based care services to support these transformations. It’s hard work and the market has come to understand that.

A lot of companies and investors have jumped into population health management technology. What changes have you seen as companies have developed a good or bad performance track record?

The vendor landscape has changed the most in that time frame. Some organizations have fallen down or fallen out of the market. Investors and operators are learning that the economics of providing population health technology require substantial investment and substantial time for that business model to mature, all while simultaneously delivering a high-quality customer outcome. It takes a unique bond among investors, customers, and vendors to make it all work. We have seen that ourselves and with competitors who don’t have those elegant relationships and therefore have struggled or failed to be successful in the marketplace.

KLAS’s recent population health management report commends Arcadia and a couple of other vendors for being good at offering customers strategic guidance as well as technology. How does a health system move from the vague idea of becoming involved in population health management to then developing a strategy and looking at technology?

The organization first typically organizes, within themselves, the right core team to drive that agenda. It becomes a focused strategic element that is driven by a population health executive, title notwithstanding, who has a clear vision of what they’re trying to achieve.

Population health management has a number of entry points. Priorities can be set in a lot of ways. Sometimes we see an absence of priorities or clear priorities, but the organization has good leadership, good governance, and an understanding the foundational elements of successful population health implementation. That’s the starting recipe for how organizations come together. Where we come in on the strategic side is helping them, once they have or are implementing the right set of tools on the right data asset, determine how they might focus. These are very specific, on-the-ground efforts, given the outcomes we’re seeing through the data and the tools that we avail to them to explore that data. It’s a tactical strategy.

Are health systems and insurers interested in improving an individual’s long-term health beyond simply reducing their own short-term cost and risk?

That gets to the heart of what we believe is exemplary population health management. It starts with a clear economic mandate. We’ve shifted our business with all of our capabilities to focus on making our customers economically successful first, in terms of proper execution in their risk contracts, so that they have the opportunity to make other investments in population health management outside of that. We’ve learned in the past five or six years that there’s a lot of things to chase in population health for the good of population health, but sustained financial success is the bridge to sustained successful health management of a population.

Would an insurer approach population health management differently than a large health system?

I would say no at the highest level, but obviously they have different business mandates than providers. We’ve been able to foster the notion between health plans and provider organizations that a properly-positioned, highly-usable, high-quality data asset in the marketplace — where the population at large is visible in every sense of the word from a data perspective — serves all parties, both in their own strategic designs and in the broader design of what we all want for healthcare.

We talked last time about the number of provider organizations that either don’t have a data strategy or are knowingly or unknowingly using bad data to drive their decisions. Has that improved?

It definitely has improved. Everyone understands now that data is important. Not everyone has prioritized it at the top of the list of their value-based care strategy, which we would recommend be the case. We last talked that it was still, in many cases, a challenge to get some organizations to understand why clinical data was an enhancement for them in the context of a data asset. That really is no longer the case.

How do you position and differentiate the company?

A lot of the data work that we do demands a certain quality level that is informed by tenure. We have been working at data harmonization for longer than most organizations in the space, probably by an order of magnitude. Some of that shows up in KLAS reporting in terms of the types of enterprise customers who choose Arcadia. We hear a lot that that’s because of our experience. Outcomes, as in the tangible and referable results from the deployment of our technology over time — improvements in risk coding, reductions in total cost of care, dozens of other outcomes — we focus on seeing that our customers can not only deliver those for themselves, but are eager to talk to other customers about that. Referenceability is a crucial, crucial item for us.

From a market perspective, we’re the only vendor that serves both health plans and providers. That has continued for us since the beginning. That position in the market is important since we believe in this concept of density of market, where providers and health plans in given markets are working together to advance population health. We approach the market a little bit differently in that regard.

Do you have any final thoughts?

We are enthusiastic about where the market is. We feel fortunate to be in the category of leader in advancing progress in managing population health. We’re equally excited about the nature of the market, which has settled out with some really good companies trying to advance this important goal for healthcare.

As a company, we’re excited about our ability to deliver what we believe is the most important thing for the next phase of advancement in this marketplace, which is both speed to value and lasting value for customers. If companies like Arcadia can continue to deliver that to the pioneer organizations that are trying to advance healthcare by better population health management, then everyone will benefit. We’re excited to be in the top category of companies that are delivering against that.

Morning Headlines 1/23/19

January 22, 2019 Headlines No Comments

Hospitals Stopped Readmitting So Many Medicare Patients. Did That Cost Lives?

Medicare patient readmissions within 30 days dropped sharply after CMS started penalizing hospitals financially, but the death rate appears to have increased during the same period.

NantHealth warned over de-listing

NantHealth, whose shares have fallen to $0.70, receives a Nasdaq de-listing notice now that the stock has closed at under $1 for 30 consecutive trading days.

Manhattan’s Largest Automated Parking System Puts More Cars in Less Space

ParkPlus installs an automated parking garage at a hospital in Manhattan, giving patients the ability to retrieve their cars and pay for their space via tablets integrated with the facility’s IT system.

She scammed dozens into giving up their savings so she could stay at the Ritz

A Virginia woman is sentenced to 15 years in federal prison for scamming $5 million from 50 investors in her phony telemedicine company.

News 1/23/19

January 22, 2019 News 5 Comments

Top News


Medicare patient readmissions within 30 days dropped sharply after CMS started penalizing hospitals financially, but the death rate appears to have increased during the same period. That raises questions over whether the lower readmission rate was created by hospitals turning away older patients with a legitimate medical need just to avoid paying penalties.

The 20-percent readmission rate turned sharply down within weeks of the penalty’s implementation (which raises eyebrows of how that happened so quickly), but a Health Affairs study attributes at least half the decline to Medicare’s simultaneous expansion of billing forms to include up to 25 diagnoses instead of the previous nine.

The corrected readmission rate reduction was 1.3 percent, about the same as reported by hospitals that were exempt from the penalties. More patients who had been admitted for heart failure and pneumonia died, however.

Reader Comments


From Associate CIO: “Re: Cerner’s voluntary separation program. Cerner still has many overpaid (by KC standards, anyway) associates it inherited with the Siemens acquisition, Brent Shafer has questioned why Cerner has over 400 VPs, and the company has a lot of topped-out associates who have become stagnant after 20-plus years and it needs some new blood to reignite investor interest.” Unverified. My only issue with VSPs is that they target older employees and the volunteers they attract are often the better-credentialed employees who know they can find jobs elsewhere – the lower-demand employees aren’t about to sign up. There’s also the nagging feeling that the executives are second-guessing their own decisions of who to hire and how much to pay, as layoffs are a management failure. Brent as the new guy may have more energy about correcting whatever undesirable situation exists, although I wonder if, as a first-time CEO, he’s worrying too much about impressing investors by applying the easily wielded but blunt headcount scalpel to give the share price a quick goose.

From Coach Ella: “Re: HIMSS. I notice quite a few bloggers are hitting the exhibit hall floor, participating in and promoting speaker sessions as paid by vendors. Is that allowed under conference rules?” Sure. I wouldn’t do it because it’s kind of a cheap sellout to take vendor money for participating in a booth-hosted “education” session of questionable quality, but I don’t begrudge those who do, even though their presence would give me as an attendee yet another reason to steer clear.


From Savoy Special: “Re: Health:Further. Shutting down after its two-year run of putting on a health ‘festival’ in Nashville. They are refocusing on providing advisory services to larger healthcare firms, odd since it was borne out of the Jumpstart Foundry. The format was unique in offering the opportunity to choose any track within a single large meeting room and for offering collaboration space in the middle of the conference room floor. With the growing competition for attendees from well-funded groups like HLTH, CES, and SXSW, which conference will be the next to fold up?” Health:Further (or maybe it’s now Health:NoFurther) announces that it has pulled the plug on its planned August 2019 event, spinning the news with a subhead of “from meeting to acting” in concluding that a once-a-year event can’t keep up with acquisitions and Silicon Valley’s healthcare interest, so it will pivot into an advisory business. It doesn’t make sense to me, especially since I see little evidence of the organization’s street cred for advising bigger and likely better organizations. My advice to attendees – stay home and do the work your employer is paying you to do and quit wasting company time and money screwing around at low-value conferences (which is most of them) looking for easy answers to your organization’s challenges. Imagine how un-innovative Amazon and Google would be if their executives just copied each other based on seeing what everyone else is doing via conference presentations and networking.

From Tip O’Spear: “Re: email newsletters. You should do an HIStalk one.” I’ll pass, for these reasons:

  • The “open rate” of emailed newsletters – even those that manage to bypass the invincible spam filters – is low. I’ve signed up for several of various kinds and rarely open any of them, having either lost interest or felt misled about the value I would receive.
  • Nobody needs to read industry news every single day. CNN baits you into raising your political blood pressure by making every minor story seem like breaking, relevant news that you need to follow, and other newsletters and sites mimic that reader-unfriendly practice.
  • Newsletters are usually just sterile links to online junk (mostly press releases) as “curated” by faceless, questionably credentialed writers who are willing to work cheap.
  • Newsletters suffer from the same misaligned incentives that exist with most sites – their measure of advertiser success is open rates and clicks (which encourages clickbait headlines), sensationalistic story selection, and articles that run  way too long, all intended to make you spend more time on their site than is necessary. Your time is their product.
  • When is the last time you learned something from a health IT newsletter that allowed you to perform your job better, gave you a career advantage, or led you to feel more fulfilled?

From Robert Lafsky, MD: “Re: NEJM piece on urgent care. A good look from the other side would be a piece by Atul Gawande, where he walks us through a typical day in his surgical practice, accounting for every dollar charged and every dollar collected for every action he takes. The big picture is too big and the problem is at the ground level.” RL references a NEJM piece that I’ll recap below, but in the mean time, this idea is brilliant. I would love to have Atul Gawande describe the downstream effects, especially the financial ones, that emanate like a boat’s wake from his part-time surgery practice. What’s the cost and price of the supplies and drugs that he uses without thinking? How much does he bill and what does insurance actually pay? What’s the hourly cost of running that fully staffed OR? What layers of management are involved in stocking it, cleaning it, scheduling it, and staffing it? How many of his patients experience financial hardship? What documentation is the by-product of a surgical case, how much of that does he input directly, and who benefits? What does he think of competing ambulatory surgery centers? Is he appalled that Partners HealthCare is such an aggressive market bully and pays its executives exorbitantly? (something that he might want to look into on behalf of his new employers).


From DisCerner: “Re: [CIO name omitted], formerly of [a merged health system]. Now the CIO of Banner Health.” Unverified, so I redacted the identifying information and will update if I get confirmation from my media inquiry. UPDATE: Banner Health confirms that Deanna Wise, former EVP/CIO of Dignity Health, has been hired as CIO. She was announced on December 4 as CIO of CommonSpirit Health, the 140-hospital, $30 billion Chicago system being formed by the merger of Dignity Health and Catholic Health Initiatives, but left immediately afterward.

HIStalk Announcements and Requests

image image

HIStalk readers are smarter and more motivated than industry slackers, so I would be remiss if I failed to take advantage of your expertise by asking you to take three minutes to fill out my once-yearly reader survey. You’ll not only be helping me decide how to spend my HIStalk time over the next year, you’ll be entering yourself into a drawing (which will require me to spend half a day stretching my minimal, rusty Excel skills into writing a randomization formula) for a $50 Amazon gift card.

I erroneously (or at least prematurely) predicted at HIMSS18 that voice assistant add-ons would be all the rage by now, as EHR vendors seemed anxious to trade paint in trying to become the first to voice-power their software Alexa style. I haven’t heard much since from the companies that seemed to be leading the charge – EClinicalWorks, Epic, and Nuance come to mind. I’m curious if anyone is actually using those assistants and what their experience has been. I see that Vanderbilt is presenting at HIMSS19 on its work to voice power Epic.


I’m annoyed at the overused, lame pun of “XXX will see you now,” in which “XXX” is a substitute for “the doctor.” When is the last time, other than on black-and-white TV shows, that someone actually said, “The doctor will see you now?” It reminds me of my early days in a small, rural hospital, in which the office nurses who rounded with our poorly skilled but arrogant community-based medical staff reverentially referred to their employer as “the doctor” as though no others existed. I even heard the wives of those mostly foreign-trained, exclusively male physicians – many of whom treated our nurses like lower-caste handmaidens — refer to hubby in social situations as “the doctor,” although maybe that’s a cultural thing. This particular for-profit hospital had a written policy that employees had to clear a path for incoming doctors in patient room hallways, specifically to hug the wall Parting-of-the-Red-Sea style so that the Doctor (administration capitalized it incorrectly to suck up hoping for more bonus-boosting admissions) could proceed unmolested to wreak patient harm more efficiently and then get back to the office to practice their specialty of defrauding Medicare. That sounds bitter and probably is, but our employees had to intervene nearly constantly to keep doctors from killing our patients through their obvious incompetence. Don’t serve on a hospital’s death committee if you want to preserve your healthcare innocence.

Listening: The Budos Band, a nine-piece instrumental Afro-Soul outfit from New York City that is horns-forward in a 1960s “Iron Butterfly meets Chicago and the Ventures somewhere in Africa” sort of way. I was despondent that I couldn’t get Shazam going in time to ID the song playing in the restaurant I was in, but luckily they were tracking just this band and I was able to get the artist on the next song. Their music would be edgier and more interesting than the usual tinkly piano solos for dinner music.


None scheduled soon. Previous webinars are on our YouTube channel. Contact Lorre for information.

Acquisitions, Funding, Business, and Stock

NantHealth, whose shares have fallen to $0.70, receives a Nasdaq de-listing notice now that the stock has closed at under $1 for 30 consecutive trading days. Companies usually skirt that issue by announcing a reverse stock split, which accomplishes nothing other than to confirm their failure and to allow them to keep investors at the table as they hope for an unlikely miracle. 


  • Bergen New Bridge Medical Center (NJ) joins Collective Medical’s network to support opioid monitoring, real-time event notification, and care collaboration.



Quality benchmarking organization URAC hires Shawn Griffin, MD (Premier) as president and CEO.

Announcements and Implementations


Brilliant: a “high-density parking solutions” vendor opens Manhattan’s largest automated parking system for an unnamed hospital. The robotically operated system provides 119 spaces in a 30,000-square-foot, four-floored structure, squeezing in 30 percent more cars than garages that require space for drivers to maneuver. It looks like valet parking to visitors, but the valet just drops their car off at the automated system’s loading bay. The system is integrated with the hospital’s information system so that customers can summon their cars and pay from computer tablets, allowing them to hop in and leave from the valet desk. The company’s other projects allow human-free operation, where drivers retrieve their car by app. I hate navigating through dark, tight parking garages, nearly always getting stuck behind a clueless visitor who will wait forever for someone to vacate a low-level spot when everybody knows that the entire upper level is surely empty. Pro tip: drive up a couple of levels, then go back down on the other side since few will take the risk of not finding a freshly emptied space there. Pro tip 2: it’s better to park higher if that gets you closer to the elevator since you’ll be walking horizontally, not vertically.

Allscripts offers free genetic testing to its employees via its 2bPrecise subsidiary. Free or not, I wouldn’t want my employer – who likely provides my health insurance and makes cost-benefit decisions about each employee – potentially learning my genetic predispositions.

King’s Daughter’s Medical Center (MS) goes live on Meditech’s Expanse Point of Care.



Brendan Reilly, MD pens a cynical yet insightful NEJM recap of his experience taking his wife to an urgent care center in “The Spy Who Came In with a Cold.” His observations, made while keeping mum that he’s a doctor:

  • Everybody wore the same scrubs, white coat, and stethoscope as a corporate standard, leading him to question whether everyone (or even anyone) really knows how to use the stethoscope.
  • He notes that while the clinic’s website boasts, “We’re on your side,” it fails to identify who the opposing “side” is.
  • He describes the eyes of the young doctor working as, “like she slept last night but not the night before. She looks alert but hurt, eager yet ambivalent, like she’s not sure she wants to be here but here we are, so let’s go.”
  • The doctor didn’t use the stethoscope or otoscope correctly and forgot to listen to his wife’s lungs, then diagnosed pneumonia after incorrectly interpreting an x-ray while waiting for a cross-town radiologist’s report.
  • He questions how doctor quality should be measured, noting the unreliability of direct observation and the difficulty in linking care quality to ever-increasing cost.


A Virginia woman is sentenced to 15 years in federal prison for scamming $5 million from 50 investors in her phony telemedicine company, spending the proceeds on luxury world travel. She erred in appearing at her sentencing carrying an $1,800 leather portfolio paid for by one of her investors (a special needs teacher and recent widow), with the angry judge ordering her to surrender it since “you’re not going to be able to take it with you to prison anyway.”


Who’s signing up for this? It’s an award-winning solution (or will be when I announce the HISsies winners).

Sponsor Updates

  • Provider management platform vendor Phynd earns ISO 27001 certification for its information security standards.
  • Pivot Point Consulting enters into a managed services partnership with PDS.
  • Access publishes a new client spotlight, “Norman Regional Health System Takes Next Step on Paperless Journey.”
  • AdvancedMD publishes a new guide, “What NOT to do When Working with a VBC RCM Outsource Supplier.”
  • Aprima’s PRM, Version 18 earns ONC Health IT Certification from the Drummond Group.
  • CarePort publishes a new case study featuring St. Vincent’s Health, “Using Ad Hoc Reporting to Reduce Excess Days and Associated Operational Expenditures.”
  • Arizona HIMSS names CTG Client Partner Patricia Allvin treasurer and VP of finance.

Blog Posts



Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates. Send news or rumors.
Contact us.


Morning Headlines 1/22/19

January 21, 2019 Headlines No Comments

Health IT company consolidates headquarters into new Northwest Austin office

EHR and practice management company EMDs consolidates operations and relocates its headquarters to new office space in Austin, TX.

Kaia Health gets $10M support for AI-powered management of chronic pain

Kaia Health raises $10 million to further scale its chronic pain treatment app.

The VA Wants to use DeepMind’s AI to Prevent Kidney Disease

The VA is working with Alphabet’s DeepMind unit to develop AI software that predicts which patients will suffer from acute kidney injury.

Curbside Consult with Dr. Jayne 1/21/19

January 21, 2019 Dr. Jayne 2 Comments


Being a consultant who is also a physician, I sometimes have the opportunity to work on projects that might not happen if I didn’t wear both hats. This week I worked on a project that I never thought I’d be staffing – a comprehensive analysis of medical marijuana policy for a fairly large healthcare organization. I’m a year into an EHR consulting engagement with them as their independent “physician whisperer” to help physicians who are struggling with adoption, so I was surprised when they asked if I’d be interested in taking on another engagement on a different topic.

I’m pretty sure I was pulled in so that the project could be done somewhat off the books, allowing someone who isn’t a stakeholder to pull the facts together so that the actual stakeholders can have a focused discussion of their options. The organization has a policy that prohibits its employed physicians from participating in the state’s medical marijuana program, but the leadership is under pressure by a subset of physicians to change its position.

It’s similar work to what I did in the CMIO trenches and we wanted to research a new system or solution and determine what kinds of impact it would have on patients and providers. I had to dig into the people, process, and technology issues along with legal and regulatory ones. I am putting together pro/con documentation of the various risks and benefits along with cost modeling for implementation and support of this new service line.

I’ve also done some reference calls with other organizations to see how they are handling and documenting medical cannabis in their practices. It made for some fun voicemails. “Hey there, it’s Jayne. Hope you’re having a good New Year. I’d like to grab 15 minutes on your schedule to talk about marijuana.” Not exactly a sentence I ever thought I’d be saying, but it’s been an interesting project.

Since I’m mainly reaching out to informatics colleagues, they have ready access to data, which has been great. If their physicians participate, they can tell me how many recommendations, certifications, or referrals are being done over the course of a year and what conditions are most commonly being treated. It’s been great to get the objective data and not just talk about the policy piece.

Overall, it seems that a very small number of organizations allow their employed physicians to be part of state medical cannabis programs. Most of the folks I’ve spoken to that work in organizations that permit it have had to customize their EHRs to some degree to support the workflow, which makes sense given the different programs and requirements in the states that allow medicinal use.

One clinical informatics team is playing it both ways. The health system’s physicians aren’t allowed to refer for it, but the independent community physicians that are on the shared EHR platform have the autonomy to do what they want. That makes for some interesting permissions issues in the EHR.

My informatics connections noted that there have also been concerns from patients about privacy issues and whether information about marijuana recommendations become part of the record that is transmitted to other physicians. That can be a grey area. In many states, drug use is considered sensitive information, but if it’s considered a medical treatment, that might fall outside the realm of a sensitive authorization and might be sent without a patient having the ability to restrict such sharing.

Since I cast my research net pretty wide, I was surprised to learn that there is a robust niche for medical cannabis in the telehealth world. California, New York, and Nevada allow such visits to be done via telemedicine. One firm, PrestoDoctor, has conducted more than 50,000 appointments, with patients sending documentation from their local physician prior to an online consultation with one of its physicians. Pricing varies by state, from $49 in the Golden State to $139 in the Empire State. PrestoDoctor also sells various products and accessories on its website, including a bacon-flavored hemp extract for pets.

I also learned that there are cannabis-specific EHRs, along with specialized billing systems since many cannabis practices are cash-only. Other practices use traditional practice management systems and submit regular office visit codes to insurers.

There are plenty of specifics that have to be figured out. Since federal law still considers marijuana an illegal drug, states have developed mechanisms for physicians to “recommend” its use rather than prescribing it, or to “certify” a patient’s qualification under state law. Neither of those approaches fit nicely into the usual EHR’s selection of medication modules, durable medical equipment management, or referral / authorization workflows. States also have varying lists of qualifying conditions, and the small number of clients who might want cannabis content aren’t going to make it sensible for mainstream EHR vendors to support those workflows. I reached out to a couple of vendor friends to find out where medical cannabis might be on their EHR road maps, and the responses ranged from “what did you just say?” to outright hysterical laughter.

Now that the research is largely complete, I will be spending a fair amount of time creating documentation for the discussions and making sure that I have answers to the questions that stakeholders are likely to ask. I will also be putting together the cost modeling to determine whether it makes financial sense to add the service, along with recommendations for technology and workflow changes that might be needed to support an implementation.

I’d like to get permission to independently survey the physicians to determine what percentage would go through the process to be able to recommend medical marijuana, which would significantly impact the decision. I suspect that it’s a small vocal minority and that the majority of physicians aren’t at a place where they would want to do it, but the organization doesn’t want to get people agitated around the issue until it has more of a grasp on what it would entail. There’s clearly enough interest to engage a consultant and I’m not complaining about a bit of unexpected work during an unanticipated slow period.

Does your organization allow physicians to participate in medical cannabis programs, and if so what solutions have you had to employ? Leave a comment or email me.


Email Dr. Jayne.

HIStalk Interviews Eric Widen, CEO, HBI Solutions

January 21, 2019 Interviews No Comments

Eric Widen is co-founder and CEO of HBI Solutions of Palo Alto, CA.


Tell me about yourself and the company.

I have a long background in healthcare, over 25 years now. I’ve worked and focused on solving problems in healthcare, typically using data. We started HBI Solutions with a more sophisticated bent on using data and data science techniques to solve problems in healthcare. That focus has been on predictive insights on patients and populations, so that population managers and even individuals can understand insights into their health that they wouldn’t know otherwise. We use data science methods to do that.

What are the implications of predicting the outcomes of patients who may not even know they have a problem, as you did in applying machine learning to EHR data to identify people likely to eventually have chronic kidney disease?

That’s the million-dollar question. The punch line is, what do you do once you know this information? Our clients are focused on low-hanging fruit from a risk standpoint. They’re working closely on readmission rates for acute settings.

We have two flavors of risk models. One is the acute setting, where our insights predict what could happen before and immediately after discharge. It’s typically an acute team or a post-discharge transition and care team that is focused on things like readmissions, and inside the hospital, sepsis and mortality.

The other models are population-based models. You want to predict what’s going to happen in the future to patients who are healthy at home. The chronic kidney disease, CKD, model that you referenced is one of those. But by and large, organizations are largely focused on utilization and cost as the starting risk models. They target patients that are at risk for ED visits, inpatient visits, or high costs, then proactively enroll them in care programs.

Our more savvy clients are starting to get into disease models. CKD is one of them. But more common use cases involve risk of mortality, which was the subject of a paper we published. We’ve had organizations looking at the risk of death for a patient in a future 12-month period and proactively teeing up discussions about end-of-life and palliative care. We have a heart clinic focusing on getting patients who are at elevated risk for a heart attack into the clinic more frequently than they would otherwise.

We said in our mid-2016 conversation that CMS’s excitement about preventing readmissions would probably end up being more tactical than strategic, not really changing outcomes as much as pushing costs around. What have we learned from trying to fix a problem with what might be a blunt instrument?

A key performance indicator that is put into place by regulatory agency that includes penalties and economic implications always tends to have unintended consequences. With a laser focus on something like readmission rates, you’re looking only at a 30-day window post-discharge. The health of those people continues forever after that.

It’s more appropriate to understand all the risk for an individual. Not only within that 30-day transition and care period, but also for the next year and multiple years after that. You get people into more of a comprehensive risk management program to bend the cost curve longitudinally over time and to take care of all the patient’s risks via care management, not just the 30-day readmission risk. Day 31 and after is as important more important than those 30-day windows.

CMS, Joint Commission, and others have come up with ideas that sound naturally good, but without having data behind them. How should those organizations use the same data science as your customers?

It’s not there yet, obviously. The other unintended consequence is the administrative burden that is bestowed upon clinicians and healthcare organizations to meet these quality markers. It’s a heavy burden and a heavy lift and it doesn’t always lead to high-quality care.

Our philosophy as a company is that we purposely didn’t start with regulatory measurement, or any of those, as a focus of the company. We do those things because organizations have to do them. But our philosophy is around two markers — are we reducing the cost to manage patients and are we getting patients healthier?

As the risk goes down for things like readmission rates, utilization of the ED, mortality, or for having a heart attack, patients are actually getting healthier and moving towards an improved outcome. We’ve had clients use our risk scores as a marker to graduate people from care programs. As their risk goes down, they’re getting healthier and they no longer need an aggressive care management approach. They are closer to self-managed care.

I don’t know how this will affect how a regulatory agency thinks about care, but we have progressive organizations thinking about care in that way. Keeping their eye on the ball on a couple of important markers that are really getting people healthier. As people get healthier, they use fewer resources over time. We’re starting to get into the measurement of the cost effectiveness of this approach.

In healthcare, however, spending money to improve someone’s health today might mean someone else gets the payoff decades letter when the patient has changed jobs or insurers.

There is that kick-the-can-down mentality a little bit in the commercial markets. However, Medicare is the largest payer and people consume the most of their costs and incur most of their diseases as they stay in those programs longer. With managed Medicare, Medicare Advantage, and Medicare HCO plans, you get more of a longitudinal outlook on patients. They stay in the same plan or program. We’re targeting those types of organizations, where those incentives are aligned, because it’s a better fit.

I have seen in the commercial space exactly what you’ve described, the “it’s not our problem” approach. They are more focused on short-term risks, more interested in what could happen in the next 6-12 months instead of the next 5-10 years for that patient.

What is your most impressive customer outcome?

It’s a mix. Organizations that have been with us the longest have shown good longitudinal outcomes in reducing ED visits and readmissions. That was their largest focus, and remains their focus, because they can put their arms around that and put in programs to bend those curves.

We can look to the future for graduating people from care programs, Then we can develop, for example, more mental health-based risks. Getting people to that point of self-help to lower their rate of suicide and opioid abuse. That’s the next wave for us, as the way people think about care becomes more sophisticated. We’re not there yet. That’s our future direction.

Beyond Medicare, the VA has the incentive to pay attention to long-term patient outcomes and to implement mental health programs. Have they expressed interest?

We’ve had several conversations with the VA through one of our partners, InterSystems, which has a long relationship with the VA. We have the ability to deploy into existing technologies as smart engines behind the scenes. We can work with any workflow platform. We plan to work with the VA in the future, but nothing is in place yet.

What data elements do you wish you could get but can’t?

Our approach is to give us whatever data you have and we’ll generate the best insights on your population or on you as an individual. As we look to the future, we’re starting to work with more partners. We’ll announce this as time goes on, but these partners have access to different types of datasets. We have worked with what’s generally available in the electronic medical record and in claims systems. We’ve added outside data sources that we can get from the federal government, like social determinants and things we can add in at the ZIP code and Census Tract level.

We’ve been pitching ourselves as a population health program, but as we look at the future and getting more towards the consumer, we’re starting to work with different datasets that would allow us to develop new diagnostic screening and/or risk tools. Developing tests that don’t exist today. We could, with a mass spectrometry partner, develop analysis of proteins, lipids, and metabolites within the blood itself. It’s new data that hasn’t been harnessed, other than in research settings at a very small population level of tens or hundreds of patients.

We’ve been analyzing millions of patients in a lot of our datasets. As we work with these new partners, they’re looking to secure large population samples in tandem with EHRs in the US, but more globally. Our metabolic makeup changes at a detailed level as we exercise or enter new nutrition programs. You can measure that on a day-to- day or week-to-week basis and pick up these signals. This is really new and interesting to us. It’s the same repeatable process around machine learning that we apply to other datasets. But getting into this, we can start developing new tests in the market.

An example is a newborn screening test, where researchers who work at HBI have, in the academic setting at least, identified for every pregnant woman their likelihood of pre-term. If you’re more likely to deliver pre-term, then you will enter a different care program. Any obstetrician or pregnant woman would want that test. We want to develop tens or hundreds of those types of tests.

Another example of high-throughput analysis of mass spectrometry is diabetics whose glucose and A1C markers are “normal.” We can find that 20, 30, or 40 percent of that “normal” population is actually at risk by measuring the proteins and the metabolites within the blood. This is a new direction for HBI. It doesn’t take away from our current direction, it’s just a new channel that we’re going after. It enhances population risk management, but individual risk markers will be a more meaningful focus for us.

How can we separate the real deal from the posers in the HIMSS19 exhibit hall, where every vendor will suddenly claim that their old products are now powered by machine learning, AI, and analytics?

We pride ourselves on being a no-BS company, rolling up our sleeves and working with clients to get them better technology to make the doctor’s job easier and to get patients healthier. We’ve always taken that as a focus. We can differentiate by giving you a list of physicians who are using our technology, pointing to research papers, and putting you in touch with care managers who are using our product daily to the health and outcomes of patients. A lot of vendors don’t have those three in place.

Do you have any final thoughts?

We have been focused on US problems and issues, which have their own government and regulatory components that drive how you think about entering the market here, But if you take a general view of how to use technology to get patients healthier at a lower cost, it’s relevant to US. There’s no shortage of studies on how upside-down the US is on outcomes and cost.

But it’s relevant across the globe as well. Other countries are getting richer and eating more poorly. People there with more disposable income want access to more types of healthcare. It’s a global problem as well. It’s important to think about how technology applies not only in the US, but globally. That’s our approach in developing solutions that are relevant to people.

Morning Headlines 1/21/19

January 20, 2019 Headlines No Comments

Verily Study Watch Receives FDA 510(k) Clearance for ECG

Alphabet’s Verily announces that its Study Watch has earned FDA clearance as a medical device for performing on-demand ECGs.

Computer virus contained, hospital services returning to normal

Two dozen hospitals in Ontario are nearly done recovering from a Zero Day Virus that impacted administrative and clinical IT systems.

Dignity Health Northridge becomes 1st hospital to acquire earthquake warning system

Twenty-five years after suffering severe damage during the Northridge Earthquake, Dignity Health – Northridge Hospital Medical Center (CA) becomes the first hospital to pilot the government-funded Shake Alert earthquake warning system.

VA needs a contractor to help test its new EHR

The VA issues an RFI for a software testing contractor,  with 80 percent of the job focusing on Cerner.

Kettering luring more than 150 new jobs to business park

Kettering Health Network will this week open a $10 million, 17,000 square-foot command center that will enable staff to view patient status in real time using TeleTracking technology.

Monday Morning Update 1/21/19

January 20, 2019 News 1 Comment

Top News


Alphabet’s Verily announces that its Study Watch has earned FDA clearance as a medical device for performing on-demand ECGs.

Verily’s offering, unlike the Apple Watch, requires a prescription, perhaps because it was developed to capture health data from clinical trials participants. 

Reader Comments

From Stock Picker: “Re: Jim Cramer. What does he know about healthcare?” First, let’s recognize Cramer’s undeniably impressive credentials – Harvard BA and JD, editor-in-chief of the Harvard Crimson, small-town newspaper reporter, successful hedge fund manager (for a short time during a boom market), and co-founder of a stock website, all before he was a TV host. However, it’s always amusing when cheap-seaters (including elected officials) pontificate based on their superficial knowledge. Everybody these days thinks they are an expert on everything they’ve read about from sketchy Internet sources, holding their ground even while emotionally spouting off to an actual expert for whom the topic represents their life’s work. This is a different form of a volume-to-value challenge where loud people talk over more informed ones.


From LinkedOut: “Re: sending an immediate boilerplate sales pitch after connecting on LinkedIn. What if I write a message specific to you and your org? Less annoying?” Only slightly, assuming you are still trying to sell me something. I didn’t join LinkedIn to be pitched, especially by people or companies I’ve never heard of. It also depends on whether you’ve paid any attention to what I actually do, which means that the LinkedIn sales messages I get for software engineering services, cybersecurity consulting, or webinars for non-health IT topics are going to bug me because my profile makes it clear those don’t apply to me. Overzealous Indian dudes in particular seem to love to invade my inbox. It’s like signing up for and having some weirdo immediately start blasting out full-frontal selfies.

From Deadlock Loser: “Re: Cerner. Did indeed announce another ‘voluntary separation plan’ right after Cliff Illig’s retirement and before the long holiday weekend. Unlike the previous two in 2015 and 2016, this one contained a threat that it would be ‘involuntary’ if certain financial goals aren’t met. Combine this with substandard benefits, readjustment in titles that led to up to 15 percent pay cuts, and blunt communications like this one that kill morale. Executives who are treated like royalty prefer to hire young, know-nothings from local Midwest colleges who are cheap and don’t know any better.” Unverified, but DL included what appears to be a copy of Cerner’s internal announcement from the company’s ironically titled chief people officer, which I’ll excerpt as follows:

Today, we are announcing to all US associates a voluntary offering created for a group of US associates who meet specific eligibility criteria … In 2019, we are expecting to meet a financial goal in order to fund important areas of investment to secure our future as a growth company. If we do not meet our corporate financial goal with the voluntary separation program, we may subsequently conduct an involuntary separation program … An associate may be eligible for the VSP if … the associate has at least 10 years of service and the associate’s combined age and years of service equals or exceeds 65.

From Minimal Shrinkage: “Re: healthcare influencers. When the conference circuit gets into high gear, their insulated conversations among themselves, event promotions, meet-ups, chats, etc. grow exponentially. For every smart, successful, succinct, and cogent healthcare person that contributes to the online discourse, there appear to be 30 others who hold almost no bona fides who produce an endless stream of meaningless, pointless, and self-aggrandized commentary. My grandfather used to tell me, ‘Don’t tell me what you’re going to do, show me what you’ve done.’ Those influencers have also reduced my interest in attending the conferences they infest — I’m not interested in seeing their regurgitated ‘takes’ on stage after not being excited to see them on Twitter and LinkedIn.” 

HIStalk Announcements and Requests


About half of provider IT decision-makers look at either KLAS or Black Book before buying software or services, according to last week’s poll. It’s not really a fair comparison since Black Book covers broad trends with big sample sizes while KLAS digs deep into a specific topic (like population health management technology) while interviewing a small number of actual clients, so they are used somewhat differently. I agree with Former Community CIO on the best way to use KLAS, with this being my experience:

  • KLAS is outstanding in describing a market and differentiating vendors within it in ways that vendors themselves would never do for fear of pigeonholing opportunities.
  • A given report tells you who the players are and how much market share they hold.
  • The overall rankings usually jibe with gut feel.
  • Few people would argue with the top- and bottom-ranked products, but KLAS doesn’t help much to distinguish those slugging it out in the middle.
  • Provider executives feel peer pressure to avoid choosing a product that isn’t #1 in KLAS as those peers fail to recognize that one size doesn’t fit all, KLAS isn’t always right, and even top vendors can stumble or get acquired by a crappy parent that ruins the magic.
  • The user comments provide the most interesting information, although they are few in number, you don’t know the commenter’s background and employer to discern relevance, and unknown factors could make them suspect (the client is behind on releases, their IT shop is in shambles, or they merged with someone and had the system foisted on them).
  • Clients who report to KLAS are like Yelp reviewers – they are self-selected with their own agenda, they may not be representative of the entire client base or similar to the reader’s health system demographic, their vendor may influence their participation or rating, and the individual respondent’s background in a particular product may be superficial.

New poll to your right or here: What would you tell Tim Cook if he asked you how Apple should be involved in healthcare?

I invite you to complete my once-yearly reader survey. I have limited ambition, time, and appetite for change just for change’s sake, but nearly everything I do differently today than I did years ago is because I got a spark from someone’s reader survey response. That reminds me that HIStalk will be 16 years old in June, which seems unreasonable given that the initially low readership and conflicting day-job demands should have been strong motivators for me to find another hobby.

I am, as a glass-half-empty curmudgeon, always annoyed by something. My peeve this week, caused by HIMSS19 spam, is people sending an “invite” rather than an “invitation.”Maybe it’s inevitable that verbing nouns (such as “friending,” seeing an Olympian “medal,” or “incentivizing” EHR purchases) will lead to nouning verbs (which along with sending “invites” includes HIMSS listing its federal government “asks.”) It reminds me of nerdy hospital tech guys who worked for me who described their job as “administrating” a system or people (sometimes me included) who say they are “vacationing” given that “vacation” is the noun form of “vacate, ” I say while explanationing.


None scheduled in the coming weeks. Previous webinars are on our YouTube channel. Contact Lorre for information.

Acquisitions, Funding, Business, and Stock


Data analytics and population health technology vendor Innovaccer closes a $35 million Series B funding round, raising its total to $51 million.

Baxter International acquires True Process, whose platform performs medical device integration and analytics.


The soon-to-be-combined Beth Israel Lahey Health names BIDMC CIO John Halamka, MD, MS as executive director of its recently created health technology exploration center.

Announcements and Implementations


Researchers at Northwestern University develop a wireless, battery-free soft skin patch that analyzes sweat for PH and levels of chloride, glucose, and lactate. Possible applications include detecting dehydration (Gatorade is already on board), testing for cystic fibrosis, and potentially screening diabetics. Researchers are are working on sensors to monitor kidney function, assess post-stroke rehabilitation, and to monitor depression.


The latest twice-yearly health IT market review of Healthcare Growth Partners says that the market is full of noise, as private equity investors are anxiously looking for investments, health IT seems attractive, and the PE firms are using CRM and other digital marketing techniques to indiscriminately blast out messages trying to beat the others at finding deals. Venture capital and growth equity players say their health IT investment strategy meets expectations, while buyout companies are less satisfied. Primary deal-killers are poor business models, deal sourcing, competition from strategic investors, and valuation. Interesting respondent observations:

  • Many startups offer gimmicky solutions and are run by inexperienced entrepreneurs.
  • Some startups aren’t ready due to lack of a go-to-market plan, an immature product, or lack of pilot sites or customers.
  • Sellers are sometimes looking for a complete exit instead of a non-controlling equity investment.
  • Funding rounds are getting bigger and occurring earlier, such that even a Series A deal involves a company with no recurring revenue and product still in pilot or even pre-pilot.
  • Seed rounds have reasonable valuations, but those from Series A and beyond are overpriced compared to other sectors.
  • Top sectors are population health management, revenue cycle management, infrastructure, and payer services.
  • The most attractive customer segments are payers / employers and health systems, while the least-attractive is patients and families.

Government and Politics

The VA issues an RFI for a software testing contractor,  with 80 percent of the job focusing on Cerner.


Two patients in Scotland’s Queen Elizabeth University Hospital die from cryptococcal fungal infections that authorities believe originated in the droppings of pigeons that were nesting in a machinery room.


Bizarre: in Scotland, a 28-year-old female hospital nurse is fired and sent to prison for impersonating a rich, male doctor by using photos of handsome men and a voice-changing machine to harass and stalk 10 women she connected with on dating sites.

Sponsor Updates


  • The Patientco team works with humanitarian aid organization MedShare to sort and pack medical supplies for hospitals in developing countries.
  • The Chartis Group publishes a new paper, “2019 Healthcare Outlook: Strategic Imperatives for the Year Ahead.”
  • NextGate will exhibit at the 2019 IHE North American Connectathon January 21-25 in Cleveland.
  • OmniSys will exhibit at the ASAP Annual Conference January 23-25 in Kiawah Island, SC.
  • OnPlan Health publishes “The Changing Landscape of Healthcare Payment Plans” report.
  • Experian Health will exhibit at the HFMA MidSouth meeting January 30-February 1 in Tunica, MS.
  • Flagship Medical becomes the first durable medical equipment provider to join Prepared Health’s EnTouch network.
  • Surescripts will exhibit at North Carolina Epic UGM 2019 January 22-23 in Greensboro, NC.

Blog Posts



Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates. Send news or rumors.
Contact us.


Weekender 1/18/19

January 18, 2019 Weekender No Comments


Weekly News Recap

  • Atul Gawande cancels his HIMSS19 keynote without explanation
  • 24 hospitals in Ontario invoke downtime procedures after a virus brings down their IT systems
  • TV stock pundit Jim Cramer urges Apple to acquire Epic to bolster its sagging business
  • Definitive Healthcare acquires HIMSS Analytics
  • PerfectServe acquires Telmediq
  • Cerner co-founder Cliff Illig retires from the company’s board, which along with Paul Gorup’s 2015 retirement and Neal Patterson’s 2017 death, leaves no founders involved
  • Walgreens signs a technology and software development agreement with Microsoft
  • The first chief digital officer of NHS England resigns to take a job with a vendor that provides telemedicine services to NHS
  • A study of 2018 hospital acquisitions finds that deals are larger, the selling hospitals are usually not experiencing financial distress, and the line between for-profit and not-for-profit health systems is blurring

Best Reader Comments

I love the people who call MUMPS old, but fail to acknowledge that macOS and iOS are based on UNIX, which is also ancient. Yet no one criticizes Apple for building an empire on dinosaur technology. (I know this, it’s a UNIX system!)

If any EHR company is going to be bought, chances are likely it would be Cerner, not Epic. Judy wouldn’t let that happen until she has moved on. On the other hand, now that all three founders of Cerner have moved on (God bless you NP), Cerner is ripe for acquisition, as it has been rumored for years. I would think Apple, Oracle, or Amazon would be the most likely candidates. Not to mention Cerner’s architecture is a bit more compatible, if you will, to those companies. If Epic were ever for sale, which again probably would never happen, would an Apple/Amazon want to purchase a system whose architecture is MUMPS based? (Associate CIO)

HIMSS doesn’t need any help in finding new ways to gouge vendors out of more money. For years now vendors have seen diminishing returns on their massive investments to attend. Having a booth has become fairly pointless as they attract typically current customers and not new. Decision makers do not circulate around the exhibit hall like they used to. It is important for attendees to remember that there is no HIMSS convention without the tens of millions invested by vendors. Do them a favor and visit their booths, otherwise, I see more and more vendors opting for other conferences. (GenesRFree)

There’s also zero evidence in any of the studies cited that government requirements are the factor contributing to burnout. You can speculate, but they just tied together a lot of disparate information in an attempt to sound knowledgeable about industry trends. They also ignore the conclusions in the same surveys which point out the benefits of EHRs. (Boring)

What has always amazed me is that HIMSS was quick to take your money, yet offered little or no help in how to get an ROI out of a very expensive ordeal. What they offered was was some pretty basic guidance (do’s and don’ts) that did not amount to much. What is needed is some in-depth advice on what to do before the show, during the show, and more importantly AFTER. Seems like maybe that could be another revenue avenue for them. They do like revenue, right? (Frank Poggio)

Life was more interesting with Neal at Cerner, Jonathan at Athena, Glen at Allscripts, and Rich at IDX. HIT has lost passionate and interesting leaders. We still have Judy. (xCerner xAllscripts xIDX)

Apple needs to carefully review the history of the healthcare initiatives of IBM, GE, NCR, Honeywell, Burroughs, Martin Marietta, Northrup Grumman, McDonnell Douglas, Siemens, Lockheed, Revlon (yes, the cosmetics company!), Microsoft, Google, American Express, First Data, American Hospital Supply, and about 10 insurance companies and 20 others. Good luck, Mr Cook. You’ll need it. (Frank Poggio)

Watercooler Talk Tidbits

image image

Reader donations funded the DonorsChoose teacher grant request of Ms. S in Los Angeles, who asked for an IPad to run Lego Mindstorms programmable robots in her after-school technology club. She reports, “The IPad helped us link to product videos, building challenges, downloads, and support pages for the kids. Having an additional IPad in the after-school technology club has helped more students access the curriculum for STEAM learning. We are always short on technology and this new additional piece of technology lets me divide the students into smaller groups and allows them to be independent learners and resourceful. I have now become more of the facilitator because the kids have taken on the responsibility of acquiring their own learning. So from the bottom of my heart, I sincerely want to thank you for your generosity.”

The nursing part of Pennsylvania’s PALS online licensing system goes down, leaving healthcare-related graduates unable to take jobs because employers can’t verify their credentials. The state says it has fixed the problem with legacy system EppiccNurse, which it will replace this year. The system was developed by the Pennsylvania State Board of Nursing to link schools with the Board to communicate pre-licensure activity.


A startup that charges older people $8,000 for a transfusion of young-person blood claims to be up and running in five cities and now takes PayPal for online payments. Experts are doubtful and clinical evidence is non-existent, but the company slides under the FDA radar by offering normal blood transfusions with off-label indications. Ambrosia Health Founder and CEO Jesse Karmazin earned an MD from Stanford, joining what seems like every Stanford medical school graduate in doing something sexier and more lucrative than actually seeing patients after taking up a class spot.

ABC says its upcoming documentary on Theranos called “The Dropout” will include footage of the courtroom depositions of founder Elizabeth Holmes.


An executive of White Plains Hospital (NY) dies in the hospital where he had worked for 40 years after collapsing while giving his retirement speech. Ossie Dahl was 64.

A case study in the Irish Medical Journal describes a man with a subcutaneous abscess whose “innovative” method of self-treating his chronic low back pain was to inject himself monthly with his own semen. I’ll admire my own restraint in letting it go at that.

In Case You Missed It

Get Involved



Morning Headlines 1/18/19

January 17, 2019 Headlines No Comments

Consumer Health Tech Company Ciitizen™ Raises $17M in Series A Financing to Accelerate Data Platform Development and Commercialization

Consumer health data sharing company Ciitizen raises $17 million in a Series A funding round led by Andreessen Horowitz.

Gawande, still mum on plans for new health venture, pulls out of major health conference

Atul Gawande, MD cancels his keynote at HIMSS19.

GoFundMe CEO: ‘Gigantic Gaps’ In Health System Showing Up In Crowdfunding

GoFundMe reports that a third of the $5 billion raised on its platform since launching in 2010 has been used for healthcare campaigns.

Epic roll-out creates prescription refill gridlock

After going live on Epic in December, Memorial Hospital (NH) staff struggle to keep up with prescription refill requests as they transfer prescription data to the new system.

Far-reaching impact: Computer virus at HSN affects 24 hospitals across North

A “zero-day virus” forces 24 hospitals in Ontario to revert to downtime procedures while they work to bring their IT systems, including Meditech and Elekta EHRs, back online.

News 1/18/19

January 17, 2019 News 6 Comments

Top News



From Dave N: “Re: CNBC’s ‘Mad Money.’ Jim Cramer said he thinks Epic will be acquired or go public, citing Judy’s age.” The stock-picking record of the highly annoying Jim Cramer is poor, underperforming the S&P 500 index funds while exposing investors to more volatility, and this particular recommendation is especially questionable:

  • Judy Faulkner refuses to sell Epic and has told those whose will control the majority of shares after her — her heirs and the board — not to sell or go public. I highly doubt that drowning in a sea of Apple cash would be her preferred way to go.
  • Cramer thinks Epic needs scale, which they don’t — Epic makes billions with a few hundred boutique customers and doesn’t care about millions of app users paying $2.99, social media platforms counting ad clicks, or illogical fanboys who buy overpriced Apple devices as soon as they’re released.
  • Cramer wants Apple to smooth out its hardware vendor market bumps by buying its away into the software market, a strategy that has failed other desperately growth-seeking companies countless times.
  • Apple’s consumer innovation track record died with Steve Jobs, but acquiring Epic to leap headfirst into healthcare won’t make a dent in the real problem of flagging IPhone sales in the absence of any other compelling offering.
  • Cramer claims that Apple and Cerner practice information blocking (which is a stupid thing to say since it’s their customers that are the problem – providers who want to exchange information do so easily) and proposes that Apple become a universal repository for health data. Apple doesn’t need to buy Epic or anyone else to accomplish that, and it’s a weird thing for Cramer to say since he just got finished saying Epic and Cerner aren’t interoperable.
  • Notice the before-and-after result (above) of the show having to alter Jim’s inane comment that Epic focuses on small hospitals. The top version is the original cached version, while the bottom is a post-surgical shot after Jim’s foot was removed from his enlarged mouth. (I reviewed the show video and he really said that Epic is a small-hospital vendor while Cerner focuses on big ones).
  • “Neither Cerner nor Epic wants to make it easier for their systems to work together because their clients would have an easier time switching to a competitor.” Oh, Jim, how much (and how loudly) you say despite how little you know. And by the way, they actually make shirts whose sleeves don’t have to be rolled up – ask for something called “short-sleeved.”

Reader Comments

From Jelly Doughnut: “Re: population health management. Why isn’t it used more if it really saves money, which it should?” Because it saves money years down the road for someone else when the patient / enrollee has moved on to other insurers and providers. A company spending big dollars today to help a 30-year-old patient get healthy will most likely never see the benefit. We probably need some kind of lifetime patient financial scorecard to fix this, yet another healthcare example in which taking better care of people means making less money. The only real exception is with lifetime insurers such as Medicare and the VA, which get to reap the benefit of whatever savings they create.

From KC1981: “Re: Cerner. Is offering ‘voluntary’ retirement packages and is making it clear that if that doesn’t get to the total numbers needed, the next step is forced retirements and layoffs.” Unverified. I’ll get a call in to Jim Cramer.


From Nicholas Cage-Free Eggs: “Re: HIMSS ‘news’ publications. Down here in Australia, HIMSS dismissed the award-winning contracted editor of its flagship publication after she ran an insightful, critical overview of the problems with the Australian Digital Health Agency, whose CEO is a friend of HIMSS CEO Hal Wolf. That CEO threatened to withdraw sponsorship of the HIMSS AsiaPac conference. Now HIMSS won’t run stories without the agency’s approval.” Unverified, although I’ve heard that before and the editor’s LinkedIn shows that she did indeed leave a few weeks after her story ran. It’s not exactly ProPublica over there – they should have made it clear they weren’t looking for actual news and hired a marketing person instead of a journalist. 


From Acute Malcontent: “Re: social media influencers. There’s something to be said for the HIT crowd that seems to exist only on social media, but meanwhile, I thought you’d get a kick out of this story.” Instagram influencer Caroline Calloway – who in her mid-20s took a $500,000 book advance that she couldn’t pay back when she found out that she can’t actually write – suddenly announces after a two-year hiatus that she’s offering a national “tour” in which she will instruct attendees (at $165 per head) how to create an “Instagram brand.” Twitter roasted her for trying to recruit free helpers, arm-twisting Philadelphia and Boston ticket-holders to travel to New York City instead to make it easier for her, giving attendees packets of flower seeds instead of the promised live orchids to wear in their flowing locks, and letting 20-year-old friends with no work experience run the show. She cancelled the rest of the “tour” and claims in Fyre Festival style that it’s not her fault and she’ll give refunds. Jezebel concludes, “No one is an influencer if everyone is an influencer.” I would add a health IT corollary – you aren’t an influencer, even in its most shallow form, unless you change the behavior (especially the buying kind) of people with real influence. 

HIStalk Announcements and Requests


It’s that just-before-HIMSS time when I ask you to take three minutes to fill out my annual reader survey. I use the results for several things – to plan for the year, to manage my ego up or down as needed by reading the positive and negative comments, and to get a mental picture of who is out there on the other side of that blank screen that I vanquish every day.

Dear overzealous sales types – if you connect with me on LinkedIn and immediately spam me with a boilerplate sales message, I will instantly remove the connection, overriding my initial impulse to report you for spamming. I know people told you to Always Be Selling, but that means you are Often Being Annoying and LinkedIn already offers plenty of that.


None scheduled in the coming weeks. Previous webinars are on our YouTube channel. Contact Lorre for information.

Acquisitions, Funding, Business, and Stock


Consumer health data sharing company Ciitizen raises $17 million in a Series A funding round led by Andreessen Horowitz. CEO Anil Sethi founded the company in 2017 after selling a similar company to Apple for its health records business.


GoFundMe reports that a third of the $5 billion raised on its platform since launching in 2010 has been used for healthcare campaigns. CEO Rob Solomon says he never though the platform’s medical category would become its most popular. “We shouldn’t be the solution to a complex set of systemic problems. They should be solved by the government working properly, and by healthcare companies working with their constituents.”

Consumer healthcare website company Everyday Health Group acquires Castle Connolly Medical, owner of the Top Doctors database.


PerfectServe acquires competitor Telmediq for an undisclosed sum. Telmediq founder and CEO Ben Moore will become PerfectServe’s chief product officer.



Chris McMahon (Turbonomic) joins Kyruus as SVP of people.


Health insurer Oscar names Angela Calman (IBM Watson Health) as VP of communications.


ZeOmega promotes Sandra Hewett, RN to CNO.


Todd Stottlemyer (CNSI) joins the board of Verato.


Nordic hires Tom Kirst (Afton-Armitage) as managing director and special assistant to the CEO.


PatientSafe Solutions names Kathleen Harmon, MS, RN (Sphere3) as chief clinical officer.


  • Partners HealthCare chooses Visage Imaging’s Visage 7 enterprise imaging platform, eliminating multiple legacy PACS and giving the company the largest sale in its history.
  • Williamson Memorial Hospital (WV) will implement Meditech’s Expanse EHR via the company’s subscription service.
  • BridgePoint Healthcare (NJ) will deploy EHR software from HCS at its post-acute facilities in Washington, DC and New Orleans.
  • Arkansas State Hospital selects EHR and RCM technology from Medsphere.
  • Northern California behavioral health services provider Remi Vista selects Cerner Integrated Community Behavioral Health and patient portal.

Announcements and Implementations

Texas Health Resources implements MDsyncNet’s physician on-call scheduling software at eight additional hospitals.

Huggins Hospital and Catholic Medical Center of GraniteOne Health implement Oracle ERP and HCM Cloud.

Managed care company NextLevel Health Partners signs on for PatientPing’s real-time care notifications.

Critical Alert Systems will add real-time patient-generated data from medical devices and EHRs to its critical alarm surveillance solution via Bernoulli Health.

Government and Politics

In Canada, the New Brunswick Department of Health will spend $12.5 million to implement an EHR in three phases over the next several years.

Privacy and Security


In Ontario, Health Sciences North works to recover from a computer virus discovered Wednesday morning that has affected several IT systems. A “major computer system breakdown” was also reported at Hôpital Notre-Dame Hospital in Ontario, though it’s unclear if the two incidents are related.



After going live on Epic last month, Memorial Hospital (NH) staff struggle to keep up with prescription refill requests as they transfer prescription data to the new system. The hospital has set up an “urgent refill” hotline and special walk-up window at the hospital for patients. It has also established a hotline for pharmacists having trouble with the system. Interim CEO Lee Myles admits the new system’s “complexity is amazing,” and that employees are working “ridiculous hours” to get refill workflows back to normal.


CNBC reports that Apple is talking with three Medicare Advantage plans about offering the AppleWatch to members at a subsidized cost. The $400 latest version of the Watch features senior-friendly fall detection. Skeptics have pointed out that more evidence is needed to determine the device’s impact on outcomes, and that seniors may have a hard time manipulating and reading the device’s small controls and display.


Medtronic develops an app that, for the first time, gives users of its pacemakers the ability to pull data from their own devices.


Coursera adds 100 healthcare-related courses to its roster of online classes, emphasizing healthcare informatics, management, and public health.


Atul Gawande, MD cancels his keynote at HIMSS19, which isn’t surprising given his media hiatus since becoming CEO six months ago of the joint venture between Amazon, JPMorgan, and Berkshire Hathaway. Maybe he’ll spend the time he’ll save naming the darned thing. Meanwhile, HIMSS announces that HHS Secretary Alex Azar will join CMS Administrator Seema Verma in a Tuesday evening session on interoperability and patient engagement that, based on their Twitter track records, will turn into a political rally.

A 17-year-old teenager in China who sold a kidney to black marketers for $3,000 in 2012 so he could afford to buy an IPhone 4 and IPad 2 is now a 25-year-old man who will be bedridden and on dialysis for life after his remaining kidney failed due to poor technique in the illegal surgery. Five people, including two moonlighting surgeons, went to jail and Mr. Wang was awarded $300,000 in compensation. It’s probably not this Mr. Wang.

Sponsor Updates

  • Optimum Healthcare IT will host a HIMSS19 lunch and learn session titled “Optimizing Outcomes with Real-Time Data Harmonization” on February 12.
  • Patientco adds financing options from ClearBalance to its SmartFinance patient payment technology for health systems.
  • ZappRx partners with Bayer, giving it insight into the medication access and prescription habits of patients living with Pulmonary Hypertension, and with Relapsing Remitting Multiple Sclerosis.
  • Elsevier Clinical Solutions releases a video summary of its recent Health Dialogue roundtable.
  • Clinical Alert Systems partners with Bernoulli Health for real-time clinical surveillance, advanced analytics, and intelligent alert notification.
  • Healthfinch will exhibit at the Allscripts Global Sales Kickoff January 22-24 in San Antonio.
  • Imat Solutions releases a new podcast, “Advancing Data Confidence at HIMSS 2019.”
  • The InterSystems IRIS Data Platform is now available on the AWS Marketplace.
  • Nordic hires Tom Kirst (Afton-Armitage) as managing director/special assistant to the CEO.

Blog Posts



Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates. Send news or rumors.
Contact us.


EPtalk by Dr. Jayne 1/17/19

January 17, 2019 Dr. Jayne No Comments

All is not quiet on the Congressional front, with the House Committee on Ways and Means sending a letter to CMS Administrator Seema Verma asking for greater transparency in the development of alternative payment models. The Committee notes that “significant policy changes made unilaterally by the executive branch without sufficient transparency could yield unintended negative consequences for beneficiaries and the healthcare community.”

The letter goes on to note that the Center for Medicare and Medicaid Innovation (CMMI) is tasked with consulting experts in medicine and healthcare management, yet the model development process doesn’t always follow the traditional rulemaking cycle that includes public comments. The letter requests that CMS provide a list of “models under active consideration by the agency,” including timelines, comment periods, etc. and gives the agency two weeks to deliver a response. I won’t be holding my breath to see what CMS says.


The US Government Accountability Office has also been busy, releasing a report this week regarding challenges for matching patient records across provider platforms. The 21st Century Cures Act required the GAO to review current matching practices and ways in which federal bodies might improve things. The report highlights work by the Pew Charitable Trusts to improve matching, including standardization of demographic data such as addresses and how EHRs use that data and exploring how biometrics might assist in matching.

It’s clear that we aren’t going to see a universal patient identifier anytime soon, which makes it incumbent for members of the patient care community (including EHR vendors, hospitals, providers, and payers) to come together and create solutions. Having worked with multiple EHRs, I can attest to the fact that some of them don’t even follow US Postal Service formatting standards for addresses, which boggles the mind since they’re readily available and used in many industries. I’ve seen multiple online retailers use the address formatting technology, so I’d think that my medical records should be at least as important as the breadmaking supplies I ordered recently.

The game is afoot for organizations ready to submit 2018 Merit-based Incentive Payment System (MIPS) data. Those eligible clinicians who participated have until April 2 to get the job done at the QPP website. There is a new system in place with a compound acronym – the HCQIS (Health Care Quality Information Systems) Authorization Roles and Profile (HARP) system will be used.

I’d recommend logging in early to make sure your credentials from the previous Enterprise Identity Management (EIDM) system were transitioned correctly and to register with HARP if needed. On a lark, I checked my participation status using the QPP Participation Status Tool and was surprised to see it list a location where I have never seen patients. I’ll be taking the penalty, so will leave it up to my practice administrator to sort it out.

CMS also announced that the new Medicare card mailing process is complete, and that nearly 60 percent of claims are being submitted with the new Medicare Beneficiary Identifier (MBI) numbers instead of Social Security numbers. The final wave of mailings includes Medicare beneficiaries who live in Canada and Mexico. While the final transitions are being worked out, patients can use their old cards through December 31, 2019, although I hope those transactions quickly become few and far between.

Walgreens is uniting with Microsoft to create so-called “digital health corners” for its retail stores. Walgreens plans to begin using the Microsoft cloud and data centers, which means no business for rival Amazon. Walgreens CEO Stefano Pessina is quoted in the article as saying, “Healthcare is too complicated, too big, and if I can say, a little too messy” thus the need to “team up with many, many different, practically all, the players in this industry.” I guess they consider cosmetics service Birchbox to be an industry player since they’ve signed a deal with it as well. The “digital health corners” will be piloted in a dozen stores and are intended to promote management of chronic diseases along with sales of health-related devices. The deal also includes Walgreens signing on for Microsoft 365 services.

In other retail news, Walmart is breaking with CVS after a squabble over prescription pricing. Patients with CVS Caremark pharmacy plans will have to find other places to pick up their medications, although I suspect other providers can’t possibly compete with the loss of the people-watching opportunities at Walmart. In a confusing twist, Walmart-owned Sam’s Club pharmacies remain in-network, so you can still pick up your bulk items at the same time as your pills.

I’m helping some practices make the move to telehealth and have been poring through the literature looking for data on outcomes from virtual care vs. traditional in-person visits. A recent news release from Massachusetts General Hospital caught my eye, as their work was published in the American Journal of Managed Care. The authors found that video visits maintained the same perceived quality of care and communication as in-person visits, while being found more convenient. Mass General has been doing telehealth for more than a decade, with video visits being offered for the last five years for patients requiring follow-up care. Additional key findings included:

  • 79 percent of patients found it easier to find a convenient time for a follow-up video visit compared to an in-person office visit
  • 21 percent of patients thought the overall quality of virtual visits was better
  • 68 percent of patients scored the video visit at 9 or 10 out of 10, and those ranking it low typically cited technical issues as the reason.

Several of my close friends have reached out about HIMSS preparation, especially wanting to know whether I’m in full party-planning mode yet. Although I’ve received a couple of advanced notices from people who want to be sure their event makes my must-attend list, I really haven’t seen much in the way of invitations yet.

One of my usual HIMSS BFFs is likely not going to attend HIMSS this year. The value just isn’t there and it’s time away from work and family that doesn’t seem very productive any more. I have to admit I’ve had second thoughts about attending due to the cost and the general hassle factor, especially since my practice is being very grumpy about allowing the time away in the middle of flu season and the opening of two new locations. I held my ground on the time off, however, since by not going I’d sorely miss the opportunity for those once-a-year catch-ups that I really enjoy.


Email Dr. Jayne.

Morning Headlines 1/17/19

January 16, 2019 Headlines 7 Comments

Definitive Healthcare Acquires HIMSS Analytics

Definitive Healthcare acquires the data services business and assets of HIMSS Analytics.

PerfectServe Acquires Telmediq, #1 KLAS Rated Vendor for Secure Communication Platforms

Clinical communications and collaboration company PerfectServe acquires competitor Telmediq for an unspecified amount.

Apple should buy private digital health records operation Epic Systems, says Jim Cramer

CNBC’s Jim Cramer says Apple needs to make a “big, splashy acquisition” in the EHR space to breathe new life into its downtrodden stock, and back up claims that it will revolutionize healthcare.

Medtronic debuts first apps to let heart patients monitor their pacemakers

Medtronic gives patients the ability to pull data from their pacemakers via its new MyCareLink Heart app.

Definitive Healthcare Acquires HIMSS Analytics

January 16, 2019 News No Comments


Definitive Healthcare announced this morning that it has acquired the data services business and assets of HIMSS Analytics.


Definitive Healthcare founder and CEO Jason Krantz said in the announcement, “HIMSS Analytics has developed an extraordinarily powerful dataset including technology install data and purchasing contracts among other leading intelligence that, when combined with Definitive Healthcare’s proprietary healthcare provider data, will create a truly best-in-class solution for our client base.”

Machine Learning Primer for Clinicians–Part 12

Alexander Scarlat, MD is a physician and data scientist, board-certified in anesthesiology with a degree in computer sciences. He has a keen interest in machine learning applications in healthcare. He welcomes feedback on this series at


Previous articles:

Identify Melanoma in Images

In this last article in the “ML Primer for Clinicians” series, we’ll train a ML model to diagnose melanoma in dermoscopic images.The original data is the HAM10k images dataset, Human Against Machine with 10,000 Training Images,  which is freely available.

The dermoscopic images in the HAM10k dataset have been curated and normalized in terms of luminosity, colors, resolution, etc. The actual diagnosis was validated by histopathology (a.k.a. source of truth) in more than 50 percent of the cases, which is twice more than the previously available skin lesion datasets. The rest of the lesions’ diagnosis was based on a consensus of dermatologists. The 10k images in this dataset belong to the following seven diagnostic categories:

Actinic Keratoses – 327 images solar keratoses, intraepithelial carcinoma, Bowen’s disease (akiec)


Basal Cell Carcinoma – 514 images (bcc)


Benign Keratosis – 1,099 images seborrheic keratoses, senile wart, solar lentigo, lichen-planus like keratosis (bkl)


Dermatofibroma – 115 images (df)


Melanocytic Nevus – 6,705 images (nv)


Melanoma – 1,113 images (mel)


Vascular Lesions – 142 images cherry angiomas, angiokeratomas, and pyogenic granulomas (vasc)


The dataset is imbalanced, as the number of images in each class varies from 115 (dermatofibroma)  to 6,705 (nevus):


Instead of trying to classify seven skin lesions with a highly imbalanced dataset, let’s simplify the task to the diagnosis of Melanoma vs. Not Melanoma. If we summarize the above categories into two groups, the result is still an imbalanced dataset with:

  • 1,113 Melanoma images
  • 8,902 Not Melanoma

We can apply data augmentation to the Melanoma group and bring the number of images to be similar to the Not Melanoma group. Data augmentation applies a random combination of image modifications — such as zoom, angle, shift, horizontal and vertical flips, etc. — and creates synthetic images.


Data augmentation allows the model to be exposed to various modifications of an image of melanoma. This in turn allows the model to learn and later generalize, for example, that a melanoma pointing to the left is still a melanoma if it points to the right. The previously imbalanced 10k images become a 17.8k balanced dataset after the data augmentation of the melanoma group:

  • 8,903 Melanoma images
  • 8,902 Not Melanoma images

With such a balanced dataset, the guessing accuracy should be 50 percent. This is the baseline sanity check before we measure a machine performance on this binary classification task.

Melanoma or Not Melanoma

  • Task: categorize a dermoscopic image as Melanoma or Not melanoma, a supervised binary classification challenge
  • Experience: the 17.8k dermoscopic images dataset detailed above
  • Performance: accuracy, precision, recall, F1 score, and ROC AUC

The newly created 17.8k images dataset is randomly split into three datasets:

  • Train 10,682 images (60 percent)
  • Validate 3,562 (20 percent)
  • Test 3,561 (20 percent)

This image analysis challenge is approached from three angles, as explained in the last article: 

  • Create a convolutional NN (convnet) from scratch
  • Use existing pre-trained models for feature extraction 
  • Fine tune existing pre-trained models

Common to all the following convolutional NN models:

  • Input layer that accepts an image as a tensor: image height (pixels) x width x color (RGB). The models tested accept images from 224x224x3 to 331x331x3.
  • Output layer that predicts a probability between 0 and 1 for the image being Melanoma vs. Not Melanoma. The decision cutoff point is 0.5.
  • In between the above two layers, various architectures of convolutional, pooling, dense, and normalization layers, with millions of trainable parameters (a.k.a. neurons)

Simple Convolutional NN

The Python code and the dataset for this part are publicly available.

This convnet, with 9.7M trainable parameters, achieves an accuracy of 93 percent on the test dataset of images the model has never seen. The learning curves for 100 training epochs of this model, while using dropout, regularizers, and batch normalization techniques:



Transfer Learning

Feature extraction and fine-tuning are two common methods of transfer learning. For the following transfer learning, I’ve cross-validated eight open source, pre-trained image analysis models: VGG16, VGG19, ResNet50, InceptionResNetV2, Xception, InceptionV3, DenseNet201, and NasNetLarge. All these computer vision models are freely available as part of Keras, the ML framework used for this image analysis project.

Feature Extraction

As explained in the last article, with feature extraction, we import one of the above models and freeze it so it won’t be modified during the training process and we add our own trainable layers on top of this pre-trained model. The lowest accuracy results: 86.3 percent with feature extraction on top of VGG16. The highest accuracy was achieved with a model extracting features from ResNet50, 94.5 percent. Note that after 50 epochs, this model was still underfitting on accuracy while perfectly fitting the loss function:



The Python notebook and the data for the feature extraction (ResNet50) are available.

Fine Tuning

Fine tuning is slightly different that feature extraction. While we still import a pre-trained model and freeze it, we unfreeze the model last block of layers so these block weights will be modified during training. Fine tuning is usually done with a slow learning rate since we do not want to modify the pre-learned weights of the model last block too abruptly, as this may destroy the units (a.k.a. neurons) pre-learned “knowledge.”

The worst performance with fine tuning a pre-trained model was 92 percent – NasNetLarge. The best overall performance on this dataset – 94.7 percent accuracy – was achieved by fine tuning ResNet50, the same pre-trained model introduced above. This time the model has 10.8M trainable parameters and it displays the following learning curves:



The Python notebook and the images dataset are available.

Model Performance Evaluation

As previously mentioned, there are other metrics besides accuracy that convey information on the relevance of the results: 

  • Confusion matrix (TP,TN, FP and FN)
  • Precision
  • Recall
  • F1 score 

Considering only the best model in terms of accuracy, fine tuning on top of a pre-trained ResNet50:


This confusion matrix translates into:

  • TN:  1677
  • FP:  104
  • FN:  82
  • TP:  1698

Final metrics for the best ML model:

  • Accuracy: 94.8 percent
  • Recall: 95.4 percent
  • Precision: 94.2 percent
  • F1Score: 0.948
  • ROC AUC: 0.947

Performance Comparison on Similar Melanoma Identification Tasks

  • Dermatologist ROC AUC performance varies between 0.69 and 0.91.
  • Other ML models ROC AUC performance varies between 0.72 and 0.94.


  • Using freely available infrastructure, framework, Python libraries, and a single dataset, one can build a ML model that outperforms both dermatologists and other ML models in the detection of melanoma in dermoscopic images. 
  • By definition, a learning algorithm will improve its performance on a specific task with each and every new experience learned, so the above algorithm is expected to improve beyond 94.8 percent accuracy if it will be exposed to additional relevant images.
  • A fully operational ML model would have to predict in real time each new image it sees, whether it is melanoma or not, with the probability of the decision. Such a ML model will need to fine tune itself, on a daily or weekly basis, by being re-trained on newly labeled images, those with the ground truth reached by consensus or histopathology results. The periodical retraining of the ML model is analogous to the continued medical education of any healthcare professional.
  • A reportedly shortage of dermatologists in the US, combined with the introduction of full-body 3D photography that produces hundreds and thousands of images per patient, may hint towards the future of this type of ML models — not replacing dermatologists, but supporting them with the pre-screening process of the millions of images coming their way.


I hope that the “ML Primer for Clinicians” series has clarified the main terms and concepts in AI, demystified some buzzwords, and in the process, has left you with some wisdom and curiosity about ML. My intention is to use this series of articles as the backbone of my third book, so please feel free to contact me with any issues, concerns, requests to modify, etc. related to this future manuscript. As I am always curious about new ideas in AI/ML in healthcare, please let me know about your interesting ML challenges.

Many thanks to Mr. HIStalk, who was kind enough to offer me a podium for my musing during the last couple of months. Without his initial spark, this “ML Primer for Clinicians” would not exist.

Subscribe to Updates



Text Ads

Report News and Rumors

No title

Anonymous online form
Rumor line: 801.HIT.NEWS



Founding Sponsors


Platinum Sponsors
























































Gold Sponsors

















Reader Comments

  • Billy Bonka: I find it interesting that you pick on Epic for Share Everywhere and Lucy, while completely ignoring the impact of Care ...
  • JeanneC: Much of health care is driven by a "Show me the Business Case" model, not a "What is the best way to employ information ...
  • Fred: Epic does this from time to time; puts out apps it doesn't really care if anyone uses or solves any real problems, so th...
  • MAD: Pointing the finger at Epic Systems is not fair the issue is getting data out of the legacy systems that process is brok...
  • oofda: I wonder how many doctors would buy an EHR that didn't help them bill accurately or allow them to get their mad money fr...
  • Frank Poggio: Yep...get off the billing kick. Epic, Cerner and Meditech all started in the Lab. Did not do billing till they had hundr...
  • Oof: Every time I read the complaints from physicians about how the system was build mainly for billing, I have a "Get Out" h...
  • Yosemite: Yosemite is my favorite national park. The Valley is so stunning that photos taken from the scenic overlook Tunnel View...
  • Mr. HIStalk: They do, but unless diarrhea is involved, we're then talking about intermittent monitoring for men. It's just not fair!...
  • Tarō Gomi: But as you know... everybody poops!...

RSS Industry Events

  • An error has occurred, which probably means the feed is down. Try again later.

Sponsor Quick Links