Wednesday, June 24, 2015

We've been swindled

This is a true story.  It is not--repeat, not--about the quality of care received by my friend at two excellent Harvard teaching hospitals.  He was very clear to me that the doctors and nurses and other staff were attentive, considerate, communicative, and expert in their treatment of him.

Rather this is about a utter failure of our government to insist on interoperability between electronic health records, a failure made all the more outrageous in that the government has helped pay for those EHR systems in one of the biggest subsidy programs in modern healthcare history.

The friend was suffering from pain and went to the urgent care facility closest to his home, one operated by Beth Israel Deaconess Medical Center in one of the western suburbs of Boston.  The diagnosis was renal failure. A CT scan located the source of the problem, large kidney stones blocking the functioning of his kidneys. An ambulance was called, and the patient chose to go to the emergency department at Brigham and Women's Hospital, where his primary care doctor practices.

His creatinine level was over 8 mg/dL. (A normal level for men is  0.7 to 1.3.)

Upon arrival, he was whisked through the ED and was being prepped for surgery, but the doctors wanted to have a clearer sense of the location of the stones.  His kidneys were in no condition to have another CT with contrast, and so they wanted to look at the CT scan that had been taken just an hour earlier at the urgent care facility.

There was no way to electronically deliver the image to the BWH team.

The work-around? The BWH ED physician called his counterpart at the BIDMC ED, asked him to download the image from the BIDMC computer system and burn a copy onto a thumb drive, which was then carried three blocks away for hand delivery to the BWH team.

Fortunately, all went well for my friend.

Years ago, my friend and colleague John Halamka, CIO at BIDMC, told me that there was no technical reason that the BIDMC and BWH EHR systems could not be made interoperable.  But they still are not.

Indeed, this is a pattern nationwide, even though over 29 billion dollars of federal investment for EHRs were appropriated in 2009.  Whether you show up at an emergency room around the corner from your hospital or a thousand miles from your hospital, there is more likelihood than not that the receiving institution will not be able to gain access to your health record.  The term "meaningful use," the standard according to which the US government funds these systems, is laughably inadequate when it comes to interoperability.  We need to understand that a large number of healthcare systems, aided and abetted by some EHR providers, view it in their strategic interest to make it difficult for patients to move from one healthcare network (ACO) to another. 

We--the taxpayers and patients of America--have been swindled. Our national interest does not coincide with those corporate strategic interests.

And the fraud is likely to be compounded.  The next step in the process is a forthcoming Department of Defense procurement of an EHR system to serve the military and its dependents, whether being treated at military healthcare facilities or other facilities in the communities in and around our bases and other military installations. As I understand, there is no language in this multi-billion dollar procurement that would require the vendor chosen to achieve interoperability with those EHRs in community facilities where the government will send its patients--or where they might end up for emergent care.

Darius Tahir recently reported on Modern Healthcare:

A defense think tank says the government may regret its plan to lock the U.S. Defense Department into a 10-year contract with an electronic health-record vendor.

The Center for a New American Security released a report that sharply criticizes the department's procurement process for a new EHR system, which is expected to cost $11 billion over the life of the contract and has attracted fierce competition among four bidding teams.

“DOD is about to procure another major electronic (health-record) system that may not be able to stay current with—or even lead—the state-of-the-art, or work well with parallel systems in the public or private sector,” warn authors, who include retired Gen. H. Hugh Shelton and former Veterans Affairs Chief Technology Officer Peter Levin.

“We are concerned that a process that chooses a single commercial 'winner,' closed and proprietary, will inevitably lead to vendor lock and health-data isolation,” they conclude. 

We've been swindled once.  Will it happen again?


Matthew Holt said...

Well at least they're different competing hospital systems with different and competing hospital EMRs (BIDMC's homegrown one now owned by athenahealth and Partners' newly installed $1.3 billion Epic system)

those of you who want a similar version of this WITH CUTE PICTURES OF BABIES might want to read my experience with my kids bilirubin measurement at the SAME hospital system with the SAME EMR in San Francisco.

PS Paul, didnt you at some point have some influence over BIDMC? And are we to suppose that the barrier to interoperability in Boston in this case therefore came from the other side of this divide?

Paul Levy said...

To your PS, yes, indeed. It was an explicit part of the corporate strategy to hold on to their patients--and the associated revenues-- as revealed in one of those municipal bond investor meetings in NYC way back in the mid-2000s.

nonlocal MD said...

I read this story and recall that just last week I read a quote from some medical 'expert' with M.D. after his name, that interoperability was not a critical need, and that the money for it could be better used in other sectors of health IT. I surely wish that person could be made to explain that remark in light of this case, but his attitude might well explain the government's reluctance to throw good money after bad.

The willingness, nay enthusiasm, of the modern doctor to ignore the past medical history just continues to boggle my mind. I have no doubt it bears on the problem of diagnostic error which Wachter and others are concerned with.

Anonymous said...

In the MH article it says the critics are worried about "lock in" apparently due to the $$ and the proposed length of contract... I'd say both of these are irrelevant to "lock-in." Even much smaller organizations take a long time to implement EHR and we're talking about DOD here - it will take them several years to implement and they would be "locked in" (and unable to easily change)by virtue of the time/effort required to change... even if they had no contract


P.S. Now the issue of a 'closed' commercial system vs. open source is a completely different question.

Aditya Pathak said...

Yes, there is indeed no technical reason why the electronic medical data cannot be transported electronically for a specific, uniquely identified patient between two hospitals.

However it cannot be as easily done without manual intervention or without possibly comprising accuracy across a whole set of hospitals that a patient might have visited over their lifetime. The reason for it again is not technical - it has to do with our societal paranoia in creating a national patient identifier number. In the absence of a unique identifier for a patient across the entire healthcare system, we have created labyrinths of algorithms to try to "guess" the patient through a combination of multiple factors. Which undermines the accuracy and reliability of the data received, which in-turn forces any non-native data in EMR to be tagged separately, or converted into PDF and attached separately, and not merged into a core patient record at the hospital.

So, mandating interoperability, while a laudable goal, can only further create more complexity. I would rather have the government mandate national patient identifier, which will make it easier for the vendors and providers to respond to the groundswell of interoperability demand.

Anonymous said...

Healthcare establishment is biggest donor to our political parties. It is not shocking to uncover who wrote meaningful use policy for our government. Here is an excellent article from the past to connect the dots.

Paul Levy said...


There are simple algorithms that can get past the problem you mention. IN 2010 John Halamka employed one to allow Atrius health doctors (using Epic) to view into the BIDMC medical record system several years ago:

More here: He notes:

"We enabled a "Magic Button" inside Epic that automatically matches the patient and logs into BIDMC web-based viewers, so that all Atrius clinicians have one click access to the BIDMC records of Atrius patients. From a security perspective, we record an audit trail of every access using the Epic username of the Atrius provider doing the lookup. Only Atrius patients can be viewed, so we have limited the possibility of privacy breaches. We've also made our ED Dashboard available to those at Atrius with a need to access this information."

Tom B said...

Interoperability would be great but the CT scan should have been sent with the patient to the Emergency Department.

Paul Levy said...

Maybe. That's just another kind of work-around, though. (And if you work in a place where electronic transmission of images is part of the routine, you might not think of doing that.)

Adrian Gropper said...

The situation is very much as described. The problem is sustained, in large part, from the disenfranchisement of patients and physicians form control of their own information tools. It is noteworthy in this anecdote that the transfer between the two hospitals was a digital file and was controlled by either a physician or the patient. In radiology, the digital coding standards issue was settled a decade ago and only patient matching and transfer authorization remain as interoperability problems. Interoperability will not be achieved until patients are given the right to control our own information and to delegate that right to anyone at any time.

Anonymous said...

Yes it will. As long as Judy Falkner runs Epic the way she does, it is going to happen. Business interest for a HUGE profit comes over getting health information to another facility.


The Trout said...

Recently I had a spinal procedure ( nerve block ) at a Partner's outpatient facility. Over the past few years I have had a number of them done. BUT, this time things were different. My physician had no access to his notes from all of my previous procedures, because all records were now electronic and none of the previous records have been transferred to the electronic system. Furthermore, anticipating another procedure in the future, I asked the hospital to supply me with paper copies of all previous procedures. Approximately six weeks later, I received an envelope from a company in Georgia, I'm in Mass. What did it contain? A paper copy of my most recent procedure with no copies of any previous procedures. Apparently, anything before electronic medical records, does not exist in their mind.

Anonymous said...

At issue is data security. If data is not accessible on the network by outsiders, it is secure except by insiders. Sending data to a secured and tokenized, encrypted data bucket limits access. The data can be interoperable, but the full systems themselves must be isolated to prevent breaches.

Paul Levy said...

And fully auditable, as noted by Halamka in his post.

Anonymous said...

Interoperability should be mandated by Law.

If it is tied to one particular contract that just drives up the cost of treating that group, veterans in this case.

if congress mandates that all of these systems must interoperate that should provide a sufficient spur for the industry to do the right thing.

fairhavenhorn said...

You are seeing the failure of a backup for a failure. This is much more than just a governmental failure.

The primary mode for exchanging imaging data has been DVDs. It is standard practice at most hospitals to provide patients with DVDs of all imaging when the patient is discharged or transferred. My cousins all have their DVDs. Even my cat has her DVD.

When I checked BIDMC three years ago, they had already seen a 20% reduction in imaging studies. This was a direct result of receiving imaging studies on DVDs. BIDMC receives thousands every month. There is a history of successful interoperability.

The first failure is the failure to provide that DVD, or to lose it. Experience is that most patients take good care of their medical picture DVDs. This is important to them, and they understand DVDs. It is less common to have an urgent transfer of a patient. In most cases the patient needs to bring the DVD to their next appointment or hospital, often months later and sometimes in another country. DVDs don't usually get lost.

It's also clear that this system is working very well, when it has demonstrated a 20% reduction in imaging studies at BIDMC.

The second failure is the absence of a network transfer backup for cases where the DVD was lost.

This gap was pointed out from the beginning for all of the government meaningful use specifications. I can only speculate about the reasons it was left out. One of them might be the very high success rate demonstrated for DVDs. The network would clearly have been a backup to the DVDs. The DVDs are working well. Perhaps the decision was based on doing something that is not working at all first.

There is a degree of fixation on network rather than media, but the experience with use are less obvious. For starters, when the DVD approach was put in place there were a lot less high capacity networks available. At 25 MBit/s, which is still not available as uplload in many parts of the US, it takes 20 minutes to send a DVD. If BIDMC were receiving those thousands of DVDs over network, they would consume a continuous 25 MBit/sec, and would need more like 150 Mbit/sec to handle surges. Boston can provide that kind of service. Scattered community hospitals have more difficulty with it. The bandwidth of an ambulance carrying DVDs was usually much higher than the available network bandwidth.

When DVD was introduced, only the major centers had the capacity to consider using network transfers.

Mary Wehmeier said...

From Facebook:

Why am I not surprised? Paul, I've got a couple more of these stories. The outcomes were not good-- including Medication Errors. It amazes me this still continues.

Theresa Willett said...

From facebook:

Even within one system, like Epic, the transfer of information can be limited by the systems that implement. When I used Epic at Swedish, and I wanted to check on someone's old records from any other Epic system,including Providence who bought Swedish, I could only access a subset of the data. We were told that to prevent "note bloat" we should not pull PMH, FMH, etc into the note, but those data fields were not actually available to me on a query, so in the interest in a truly "transportable" complete note, but that info in despite the order. In this age to have such cumbersome and archaic interfaces is inexcusable, and the companies have gotten away with it by mandate.

Dave Curran said...

From Facebook:

As an employee of a local hospital IT department with intimate knowledge of the interface capabilities, I can tell you this all comes down to the willingness to share records. Our organization has strong ties with our regional partners as well as competitors, and we share much of the EHR, especially radiology, with both. That being said, I agree it should be mandated.

Alex Beam said...

From Facebook:

This happened to me on a much more mundane level; Spaulding rehab (don't get me started) cdn't access a Harvard-Vanguard x-ray. Don't believe everything you read in the papers ...

Tod L said...

Strong interoperability should have been a requirement from day one of EMR funding in the stimulus. Sadly, the lack of strong interoperability requirements in the definition of "meaningful use" enshrined a generation of client-server EMR systems in what is inherently a cloud-based problem.

Saw this one coming at the time we created the EMR stimulus, but my proposed "you only get the money if you can prove interoperability with 50% of the docs in the region" proposal was shelved instead for the "meaningful use" language which allowed the lobbyists in.

EMR use has still been a boon, but it could have been so much more.

Healthcare IT still remains 10-20 years behind other industries.

jhoger said...

Disagree. Universal identifier is not necessary for interop. The most important thing is agreeing on the data formats and interchange protocol. Matching up formatted data with an identity across systems Ismaily trivial.

Lisa Bond said...

Until there is a truly patient-centric system in place, these stories will be common. Unfortunately, nearly all these situations happen when patients are too stressed or in pain to deal with it.

The CareSync platform allows the patient to have the data and share it at will. 24/7 Care Coordinators actually contact the hospitals, get the needed information, and send it where it is needed.

Paul Levy said...

Right, jhoger. The Magic Button referenced above was designed by Halamka to compare three (I think) identifiers from the two EHR systems. When a match occurred, it had a 99+% chance of accuracy.

Ted said...

It is worse than this. In the old days a person was present in medical records 24/7 to get information of a patient. Now these offices close at 5 pm and no records can be obtained. Even during the day it is arduous and very time consuming. The EHR systems are functioning on a different planet from one another.