Search the KHIT Blog

Saturday, April 11, 2015

"Obama Administration Report Slams Digital Health Records"

Gotta love the framing, "Obama Administration." This is pure ONC, responding to recent congressional rumblings, and looking to stay visible and relevant now that they're essentially out of HITECH money. As reported in The Wall Street Journal,
Report criticizes vendors for making it costly to share patient information

The Obama administration took vendors of electronic health records to task for making it costly and cumbersome to share patient information and frustrating a $30 billion push to use digital records to improve quality and cut costs.

The report, by the Office of the National Coordinator for Health Information Technology, listed a litany of complaints it has received about vendors allegedly charging hefty fees to set up connections and share patient records; requiring customers to use proprietary platforms; and making it prohibitively expensive to switch systems.

The report also cited complaints that some hospital systems make it difficult to transfer patient records to rival systems or physicians as a way to control referrals and enhance their market dominance.

The agency didn’t cite any companies by name, however, and said it couldn’t determine the extent of information-blocking—in part because contracts often forbid customers from discussing prices and other terms.

“This is our first deep dive, where we have taken the opportunity to describe the situation and identify practices that interfere with the flow of health information,” said ONC chief Karen DeSalvo. “We believe it will take an array of solutions and we look forward to working with Congress on this.”...
Absent naming names and citing specific instances, I'm not sure what the impact of this 39 page ONC report (pdf) will be.
Information Blocking and its Potential Impacts

In contrast to these well-known interoperability challenges, the extent to which information blocking is impeding the effective sharing of electronic health information is less clear. While ONC and others are studying the problem, formal research is limited and anecdotal evidence is often difficult to interpret and still more difficult to generalize.

The term “information blocking” presents significant definitional challenges. There are many types of electronic health information and just as many factors that can inhibit its effective exchange and use. Many actions that prevent information from being exchanged may be inadvertent, resulting primarily from economic, technological, and practical challenges that have long prevented widespread and effective information sharing. Further, even conscious decisions that prevent information exchange may be motivated by and advance important interests, such as protecting patient safety, that further the potential to improve health and health care. These interests must be carefully balanced with the potential benefits from sharing of electronic health information. Finally, it is important to acknowledge that certain constraints on the exchange of electronic health information are appropriate and necessary to comply with state and federal privacy laws; this is not considered information blocking. [pg 7]
"[C]onscious decisions that prevent information exchange may be motivated by and advance important interests, such as protecting patient safety..."

The phrase "patient safety" appears ten times in the ONC report. Now, I'm not the sharpest tool in the drawer, but I'm a careful, plodding, thorough reader, and I could find nothing citing where a "patient safety" rationale was proffered (by an EHR vendor or anyone else) as a justification for impeding the otherwise appropriate and necessary flow of ePHI. All we get is ONC's gauzy, oblique, groveling statement that PHI data siloing may "advance important interests, such as protecting patient safety."

Who writes this crap?

A quick dip into Bob Wachter's excellent new book here.
What I’ve come to understand is that computers and medicine are awkward companions. Not to diminish the miracles that are, Google Maps, or the cockpit of an Airbus, but computerizing the healthcare system turns out to be a problem of a wholly different magnitude. The simple narrative of our age— that computers improve the performance of every industry they touch— turns out to have been magical thinking when it comes to healthcare. In our sliver of the world, we’re learning, computers make some things better, some things worse, and they change everything.

Harvard psychiatrist and leadership guru Ronald Heifetz has described two types of problems: technical and adaptive. Technical problems can be solved with new tools, new practices, and conventional leadership. Baking a cake is a technical problem: follow the recipe and the results are likely to be fine. Heifetz contrasts technical problems with adaptive ones: problems that require people themselves to change . In adaptive problems, he he explains, the people are both the problem and the solution. Leadership, he once said, requires mobilizing and engaging people around a problem “rather than trying to anesthetize them so you can go off and solve it on your own.”

The wiring of healthcare has proven to be the Mother of All Adaptive Problems. Yet we’ve mistakenly treated it as a technical problem: simply buy the computer system, went the conventional wisdom, take off the shrink-wrap, and flip the switch. We were so oblivious to the need for adaptive change that when we were faced with failed installations , mangled work flows, and computer-generated mistakes, we usually misdiagnosed the problem; sometimes we even blamed the victims, both clinicians and patients. Of course, our prescription was wrong— that’s what always happens when you start with the wrong diagnosis.

Making this work matters . Talk of interoperability, federal incentives, bar coding, and machine learning can make it seem as if healthcare information technology is about, well, the technology. Of course it is. But from here on out, it is also about the way your baby is delivered; the way your cancer is treated; the way you are diagnosed with lupus or reassured that you aren’t having a heart attack; the way, when it comes down to whether you will live or die, you decide (and tell the medical system ) that you do or you don’t want to be resuscitated. It is also about the way your insurance rates are calculated and the way you figure out whether your doctor is any good—and whether you need to see a doctor at all. Starting now and lasting until forever, your health and healthcare will be determined, to a remarkable and somewhat disquieting degree, by how well the technology works…

We can— in fact, we must— wire the world of medicine, but we need to do it with our eyes open, building on our successes, learning from our mistakes , and mitigating the harms that are emerging.

To do so effectively, we need to recognize that computers in healthcare don’t simply replace my doctor’s scrawl with Helvetica 12. Instead , they transform the work, the people who do it, and their relationships with one another and with patients. Sorting out all these issues will take deep thought and hard work on the part of clinicians, healthcare leaders, policy makers, technology vendors, and patients. Sure, we should have thought of this sooner. But it’s not too late to get it right.

Wachter, Robert (2015-04-01). The Digital Doctor: Hope, Hype, and Harm at the Dawn of Medicine’s Computer Age . McGraw-Hill Education. Kindle Edition, Locations 160-161.
Yeah, it's not a "tech problem," it's an "adaptive problem," one fundamentally -- as it concerns this issue -- a policy/legal problem.

My recommendation, Congress? Make willful ePHI blocking a crime. Later for all this endless, polite episodic white paper consideration of (and deference to) "all stakeholder interests and viewpoints."

Tangentially apropos, more Dr. Watcher:
Since we’re talking about government programs, you can be certain that while many people criticize Meaningful Use and HIPAA for being too much and too rigid, others criticize the ONC for being too lax in certain areas. The two biggies are usability (the ease of use and learnability of the IT system) and interoperability (common standards that allow different EHR systems to connect with each other).

Blumenthal agrees that today’s electronic health records could be far more usable, but when he tried to convert that concern into policy during his tenure as ONC director, he came to believe that judging usability was like Justice Potter Stewart’s “I know it when I see it” take on pornography. “We could not get any experts to say they could measure usability in a reproducible, valid way,” said Blumenthal, “a way we could stand behind from a legal standpoint.” While this issue continues to be debated, Halamka and others believe that the market is now mature enough for usability to take care of itself . With today’s high rate of EHR adoption, Halamka argues, the vendors will compete on usability, and the best products will win. As proof, he is piloting a cloud-based system built by athenahealth in Beth Israel’s extensive ambulatory clinics, a system known for its emphasis on usability (I’ll have more to say about this company later).

On this one, I side with those who believe that government should keep its nose out. Imagine if the government tried to design the look and feel of your smartphone or your computer’s desktop— well, enough said. I do, however, support a proposal to create a federal health IT safety center that can collect and analyze reports of dangerous IT interfaces (like the mg versus mg/ kg Septra confusion that nearly killed Pablo Garcia) and take appropriate action. And federal support for a nonpartisan organization that grades the usability of different systems based on standard criteria— that might work. But as for the government dictating standards for fonts, colors, and clicks, well, thanks, but no thanks.

Interoperability is a different matter. Interoperability is just the kind of thing government intervention was designed for, since it is a powerful public good that the private market will not automatically create. Why is it so important? Think about your cell phone: your iPhone can call your friend’s Samsung without a hiccup, even if your service is with AT& T and your friend’s is with Verizon. The same is true with your Web browser: whether you like Safari, Chrome, or Firefox for its look or features, all of them can get you onto the Internet just fine.

In the EHR world, just think how great it would be to have an interoperable system. We’d no longer need a paper airplane to get crucial patient information from one hospital to another. The costs of switching from a crummy EHR to a better one would plummet, since the patient data stored in the old system would be easily transferrable to the new one. A world of innovators (the folks who are currently writing apps for your iPhone and Android) would be freed up to create programs to help diabetic patients or families of patients with ALS manage their care— and these would sync seamlessly with existing EHRs.

On interoperability, Blumenthal defends the ONC’s early prioritization of adoption over interoperability. “I had the basic feeling that you had to ‘operate’ before you could ‘interoperate.’ And you couldn’t create a business case for interoperability,” at least at first. Today, while the market for interoperability has gotten a bit stronger, he and I both doubt that the same business case that will drive usability will also drive interoperability. This one needs an assist from Uncle Sam.

While it would be great to have total interoperability tomorrow, it will take some time, even if the ONC puts its shoulder behind it, as it now seems to be doing with its October 2014 proposal for standardized APIs. Not only are there political obstacles to overcome (put in place mostly by vendors and healthcare systems that remain reluctant to share, as you’ll see in the next chapter), but the kinds of people involved in the nitty-gritty work of interoperability do not make for a speedy process. Said Halamka, “You know, there are forest people and there are tree people. Standards folks are, by necessity, bark people. They’ll spend a day debating the relative merits of an ampersand or a semicolon.” [ibid, Locations 3613-3614]

Expect yet more of what I've named "Interoperababble"? Ahhh... but, maybe now the current darling HL7® FHIR® APIs will bail us out forthwith.

Color me skeptical, ongoing.


ONC post at THCB

Link here. Some of the comments are scathing. First, the reliably confused conflationist National Privacy Scold Deb Peel, who thinks Epic's EHR killed Dallas ebola patient Duncan:

ONC is finally discovering the lack of interoperability??? It’s been blatantly obvious for years.

The lack of interoperability started in 2002 when HHS cut patients out of data exchange by eliminating the patient’s right to give consent before PHI could be disclosed. Now it’s clear patients are the only ‘stakeholders’ motivated to disclose health data for treatment purposes.

HHS put data holders in control of data exchange, ie of ‘interoperability’. Guess what? The data holders began to treat the nation’s health data as a corporate asset and to use it for their benefits, not ours.

Patient Privacy Rights has pointed out the lack of ‘interoperability’ and the easy, cheap solution for years: restore patient control over data exchange. 

Here’s a blog from 2007:

And another blog from 2012:
HHS abdicated its duty to ensure patients’ rights were built into HIT systems. Why? HHS is protecting corporations, not people. The lack of ‘interoperability’ is a massive technology design flaw engineered by industry & HHS.

Hippocrates swore to keep patients’ sensitive information private. His oath not to disclose patient information without permission is the foundation of trust in physicians. It’s embedded in American law and Medical Ethics so sick people can trust the people who treat them.


  1. Emma Thomas, RN says:

    Pathetic. You should be ashamed. How many $ millions did this report cost? You reinvented the wheel.

    You conveniently failed to address the risks and adverse events from these devices themselves.

    So what? You interconnect a bunch of medical devices that have side effects of injury and death at an incidence that is unknown because there is not any mandatory reporting. Great!

    You are worried about interoperability yet you fail to address the crashes and EHR unavailability and aftermath, as recent;y experienced at all MedStar clinics in your neighborhood and state.

    Why is there not mandatory reporting of any and all EHR crashes and the deaths that arise from the mayhem and delays in care that are part and parcel of these crashes?

  2. Emma Thomas, RN says:


    ONC does not care:

    The systems are not fit for purpose. The adverse events, errors, and near misses go unreported. 

    This week, the doctor had CPOEed and [sic] order for BID potassium with a perameter [sic] to hold it if the serum k+ was > 4.6 mg%. the order on my palate was obfuscated by innumerable lines of instructions such that I missed the perameter [sic] and gave the potassium supplement when the serum potassium was 4.7, and the next day, 4.9.
    The doctor missed it because the MAR is terribly inaccessible and user unfriendly when accessed.

    The next day, the serum potassium level was at life threatening levels and I was beating on myself. 

    Yet, I have no one to report this to, because, no one really cares.
 Emma Thomas, I'm glad you're not my nurse. Learn to spell if you want to be taken seriously -- and, more importantly, do accurate charting.


Interesting comment over at THCB's new post "Why the Market Can’t Solve the EHR Interoperability Problem."
...I disagree strongly with Dr. Yarhagi’s faith in the market to ensure that healthcare information technology (HIT) vendors will be obliged to “develop sustainable revenue stream through reasonable exchange fees negotiated with the medical providers.”  That is, he asserts that if there were a real market without federal subsidies and requirements that all healthcare providers buy the HIT,  then providers and the HIT vendors would agree on reasonable fees for exchanging patient records.

But why would the vendors do this?  Hospitals and doctors are generally locked into their HIT systems for years. A very big hospital EHR can cost $400 million plus 4 times that for implementation (e.g., training, software upgrades, etc.) over several years. Installing an EHR in a small clinic or doctor’s office is of course far less money, but nevertheless costs a few hundred thousand dollars.  The healthcare providers have no bargaining power. Why would the vendors even bother to negotiate over fees for sharing patient data?

The more salient question is: why should any HIT vendor be permitted to charge a penny to help share data that is needed to make medical care safer, more efficient, more informed, better?  A key feature of HIT is that it allows exchanging information on patients.  The government is giving $30 billion to subsidize purchase and use of these technologies; hospitals and other providers are spending trillions of dollars buying and installing them. It is unconscionable that a vendor would even think about charging clinicians to share data on patients’ health—data that clinicians themselves entered in the first place! The government need not threaten vendors who don’t allow sharing of data. There should be no choice. Data exchange must be required through regulation.  We don’t negotiate with car drivers about stopping at red lights, and we don’t compromise on truck weight limits on certain bridges. Some rules are simply necessary for public safety.

Vendors are allowed to keep patients’ data, but they should not be allowed to hold those data hostage until the creators and users of those data pay up.  Anyway, officially, the data are owned by the patients. The last thing patients want is to be treated by doctors who are blind to their medical needs.
See my February 17th post "On ONC's "non-regulatory, market-driven" Health IT fetish, and other related business."


More to come...


No comments:

Post a Comment