Wednesday, May 25, 2011

Transplant Team at UPMC Missed Hepatitis Result - Suspicious for Health IT Failure?

A story suspicious for EMR malfunction appeared in the Pittsburgh Post Gazette:

Transplant team missed hepatitis result

Kidney donor, recipient unaware of virus' presence
Saturday, May 21, 2011

This hepatitis-C fiasco to my eye seems a likely case of IT error of omission or transmission (EOT, see Tables 4 and 5 at the FDA internal memo I cached at

It's inconceivable to me experienced transplant personnel, used to checking for data such as HIV, hepatitis status of donor and recipient, and other blood abnormalities would just miss this.

I would believe the hepatitis status, e.g., Hepatitis A, B, C, would be presented on the same screen. (If not, that's a design flaw itself.) That the Hep-C value could be missed by more than one person [possibly many people; see addendum below -ed.] seems very unlikely - even if the presentation screen(s) were cryptic and poorly organized.

A more plausible explanation to me is that a positive result simply wasn't present at the time when it was needed, e.g., due to a Laboratory Information System (LIS)-Hospital Information System (HIS) interface problem, or a HIS data erasure or misidentification error after the data was correctly received.

Another issue:

For the hundreds of millions of dollars spent on computers at this medical center, I wonder why their "system" in critical areas such as transplantation (as it would appear from the reports) is entirely dependent on two people looking at data.

After all, the IT is "supposed" to reduce error by infallibly calling attention to critical abnormalities such as this. The EMR alerting systems should have been going off like cockpit stall warnings at the data combination of "transplant" and "hepatitis C". That's not rocket science.

This further supports my hypothesis of a health IT malfunction, as does what I read about their "shutting down the living donor transplant program" for now - an extreme measure if this medical error were merely the result of a person or persons simply forgetting to check a data screen or sheet.

More generally, in a nutshell: commercial health IT from the current sellers is a fiasco. It alone among medical devices is unregulated under the FD&C Act even though FDA admits it is a medical device, it is unsafe, the literature shows its benefits grossly exaggerated, it probably cannot be made to work considering the current HIT "ecosystem" and its pathologies and incompetents, and has only captured the market through the purveyors "controlling the channel" of information and memes through their massive trade association HIMSS and political connections to HHS.

Think tobacco circa 1940's-1950's.

Yet UPMC could get away with this debacle scot-free:

Despite the way it occurred, UPMC will probably not be found in violation of any federal guidelines for the failure of the transplant team to review the positive hepatitis C test [or failure of the health IT system to properly transmit it, if that is indeed what occurred - ed.]

That's because while current guidelines from the United Network for Organ Sharing, the organization that oversees the nation's transplant centers and could suspend or shut a program down for deficiencies, do cover deceased donor transplants, they do not specifically cover living donor transplants, except in a limited way.

Emily Blumberg, an infectious disease specialist at the University of Pennsylvania, said that it is because live donations are still a relatively new medical advancement compared to cadaver donations.

"What has happened is that live donations in the United States have really taken off in the last decade, and there's only been a push to standardize live donation processes recently," said Dr. Blumberg.

In other words you're supposed to test dead people and their organs for disease before transplanting them, but live people get a pass? This is the height of absurdity.

One of the factors driving the push to finally address guidelines for living donor transplants is the rapid emergence of so-called paired kidney donation networks. In the networks, a willing living donor, who isn't a match to a friend or relative who needs a kidney, gives their kidney to someone else in the network. That recipient then has a friend or relative give a kidney to someone else, and so on.

Over the last five years, the rise of such networks have allowed hundreds of people to get kidney transplants from people they previously did not know, including a chain involving 32 donors and recipients -- 16 of whom got new kidneys -- that UPMC was part of earlier this year.

"Everybody is [testing for viruses] but it would be helpful if we could all be doing it the same way -- especially with paired donation networks," said Dr. Blumberg, who also chairs UNOS's disease transmission advisory committee.

This seems a smokescreen for this error. Missing a positive hepatitis C test has nothing to do in my mind with "everyone testing for viruses in the same way." The issue is being aware of the results.

Through that committee, UNOS is now beginning to craft new guidelines that are specific to live donation transplants. A conference in July in Baltimore that Dr. Blumberg is helping to organize will act as a public forum on the proposed new guidelines that will address everything from which viruses should be tested for, to which tests should be used, to how the results are communicated and other issues.

Non-communication of results will surely be forbidden.

The guidelines that do exist now don't spell out how information about a test is to be conveyed and confirmed by people on the transplant team, which are protocols that are left to each individual hospital to establish.

This is entirely absurd. "Protocols" are needed to check the most fundamental of tests in a mission critical medical area? Why does this process have to be "spelled out?" In other areas such as blood transfusion, isn't the precess of someone actually looking at test results very straightforward? I did it as a medical student....

Those same guidelines -- which were written for the more common deceased donor transplants -- also don't explicitly require testing of living donor organs for viruses like hepatitis C, though, as a rule, transplant centers like UPMC do so anyway. "I can tell you every transplant center in the country already screens donors for things like hepatitis C or other infections that would affect our use of that organ," Dr. Blumberg said.

That's great, as long as the information doesn't end up in '/dev/null', a.k.a. the bit bucket.

The rest of the Gazette article discusses the obvious regarding a simple issue of testing and reporting prior to sticking an organ into someone else's body.

I believe if this incident was indeed related to a computer error:

  • UPMC needs to reveal this publicly ASAP per Joint Commission Safety Standards, as similar flaws could affect other medical centers (such flaws that cause intermittent data erasure are known to exist; see Dr. Jon Patrick's study here on a Cerner system, for example.)
  • Repeats of episodes like this could reflect criminal negligence (not of clinicians, but of the administration for installing faulty medical devices), or may already represent criminal negligence if the health IT systems in use were known to have bugs causing intermittent data loss.
-- SS

Addendum May 25, 2011:

It occurs to me that hepatitis is a rather routine type of lab test.

Everyone who looked at the electronic chart, usually a cornucopia of people in an academic medical center, had opportunity to review tests like that. They range from medical students, to interns and residents, to fellows to attendings, nurses, and other allied health professionals, not just a few specific people on the transplant team.

In my mind the most likely explanation for so many people missing such a crucial piece of data is that it was simply not there.

-- SS

Addendum May 27, 2011:

A reader reminded me of my 2009 post "UPMC as Proving Ground for IT Tests On Children: Pioneers in Health IT, or Pioneers in Ignoring the Past?", where UPMC made claims it was a "proving ground" for experimental health IT development.

-- SS


Anonymous said...

"Elizabeth Concordia, UPMC's executive vice president, gave a 20-minute presentation to the hospital system's board of directors during its regular meeting and told them that the positive test was missed during 12 steps throughout the process."

Hmmm, the BOD of movers and shakers, including former Pa Supreme Court Judge Ralph J. Cappy, has been hand picked by the $4 million salaried man at the corner suite. Why would she, the $1.5 million salaried heir apparent, need to take 20 minutes to explain why the users are being blamed rather than the EHR.

Yet, no one on the BOD will speak publicly as stewards of this tax payer supported hospital.

This mecca has a new purple color coded advertising program "taking medicine from where it is to where it needs to be".

Anonymous said...

I surely hope they get this figured out soon as there are MANY lives on the line here. A very good friend of mine is on the list there and has myself as a willing paired donor and another mutual friend of ours as a perfect match donor that has one test to be done and then it is a go. With all of the testing that we both have had to go through - and by the way, myself now a second round as it has been more than a year since the original testing and no paired match has been located as of yet, that these things are known as soon as the tests are ran. I would have to agree that this must be beyond human error. My biggest question is, WHY reinvent the wheel? The testing that is done at the donor level for the living vs. cadaver donors is strictly that to be assured that the donor is in tip top health and would not become a patient themselves, otherwise all testing for the purpose of the organ to the recipient should be identical to that of a cadaver donation. Why is it so difficult to use the same protocol and procedure as well as record retention?

This needs to be addressed and corrected like YESTERDAY. We have many out their literally dying on the lists.

Tim said...

This is so full of dysinformation that it will need to be addressed.

InformaticsMD said...

Tim-anonymous wrote:

This is so full of dysinformation that it will need to be addressed.

I have several questions as you have not identified the purpose of your posting.

1. The "this" is a hanging participle. To what does it refer?

2. Dysinformation is a new concept to me; I know what "disinformation" is but not

3. Who, exactly, will need to address it? Will they need to divulge any potential conflicts of interest?

-- SS

Noah Orname said...

This seems like a cultural problem in the medical field.

Just the notion that nurses could complain to administration that a specific doctor had patients dying with 'alarming frequency' AND observed him giving unknown injections could be dismissed is unbelievable to those of us outside the profession.

InformaticsMD said...

Noah Orname said...

This seems like a cultural problem in the medical field.

There are many cultural problems in the medical field, but there are also many cultural problems in the IT field.

Combine the two and the results are as one might expect.

-- SS