Can Digital Disappearing Ink (An EHR "Glitch") Kill Patients? Part 2

At "Another Health IT "Glitch" - Can Digital Disappearing Ink Kill Patients?" just yesterday, on August 5, 2013, I wrote about a Siemens EHR "glitch" worse than any paper records system problem.  Typed order changes in the medication reconciliation process on patient discharge are disappearing into thin air, unknown to the clinicians typing the orders.  This is likely due to an issue such as some programmer forgetting to put in a statement to write the text to disk, complicated by software testing problems that missed the defect.

I noted:

... "Glitch" is a banal term used by health IT extremists (those who have abandoned a rigorous scientific approach to these medical devices as well as basic patient protections, in favor of unwarranted and inappropriate overconfidence and hyper-enthusiasm).  The term is used to represent potentially injurious and lethal problems with health IT, usually related to inadequate software vetting and perhaps even "sweatshop floor in foreign country directly to production for U.S. hospital floors" development processes (this industry is entirely unregulated).

 
Paper records may have illegible writing that would generally cause the reader to make a phone call or otherwise contact the writer, but those events are one-offs.  EHR defects potentially affect hundreds of installations and thousands of patients, en masse.  (If patients are not dying en masse from such errors, then the whole argument against paper and for IT on the issue of vastly improved safety goes out the windows, but that's an argument for another time.)

Siemens has just released another "glitch" announcement, this time with CPOE (computerized order entry):

(Medication orders "glictch" safety complaint.  Click to enlarge, text below)

Text is as follows:

August 2, 2013

Safety Advisory Notification

Soarian® Clinicals Medication Orders, Safety Complaint ID# EV06643783

Dear Customer:

This notification is to inform you that the Soarian Clinicals Medication Orders may not be operating properly in some cases in Soarian Clinicals 3.3 Service Pack 6 and above.

I note that "glitches" are not uncommon after software patches and upgrades.  See examples at the query link http://hcrenewal.blogspot.com/search/label/glitch.   This reflects inadequate vetting of the patches.

I also note that "medication orders not operating properly" is a very, very serious matter.

Although this may affect only some customers, we are taking a conservative approach and are alerting you to this potential problem. As such, please forward this notification to appropriate personnel as soon as possible.

"May only affect some customers?" (I suspect from this double-indefinite that who is affected is not rigorously known).  "Taking a conservative approach?"  I ask:  what would a non-conservative approach entail?

This letter is being sent as a precautionary measure as there have been no adverse events reported from customers.

Again, they mean "yet."

When does this issue occur and what are the potential risks?

The issue occurs while placing medication orders. In certain cases, when users select orders from predefines or personal favorites and make changes on the order detail forms, the changes are correctly saved and displayed on the forms but the Order As Written (OAW) is not refreshed to reflect the changes. The incorrect OAW is displayed in Siemens Pharmacy in the Order As Written window but the discrete order details are correct. As a result dispensing or administering relying solely on the OAW prior to pharmacy validation may result in error.  [Putting patients directly in harm's way, patients who never consented to the use of these experimental and unvetted medical devices - ed.]  Once the order is validated the OAW in Soarian is updated correctly.

This problem - manually changed data apparently not written to disk - seems similar to the "digital disappearing ink" med reconciliation bug in the aforementioned Aug. 5, 2013 post.

Immediate steps you should take to avoid the potential risk of this issue:

To prevent this issue from occurring at your facility, dispensing or administering of unvalidated order should rely on the order details displayed. Secondly, any deviations from the predefined or personal favorites should be phoned in to pharmacy as a verbal order. During validation, if the pharmacist sees a discrepancy between the order detail and the OAW, verbal follow up with the ordering physician is required.

Again, a workaround.  How many times will this workaround be forgotten, compared to issues of illegibility in a paper record resulting in a phone call to the writer?

Steps that Siemens is taking to correct this complaint:

We are diligently working to develop a correction and will test and deliver it as soon as possible.

Perhaps they should have been working more diligently to detect the "glitch" before it went live.

Also, perhaps the touted power of EHRs to reduce medical errors needs to be re-examined.  Considering bugs like these - creating en-masse problems far worse than possible with paper (another en-masse example at http://hcrenewal.blogspot.com/2011/11/lifespan-rhode-island-yet-another.html) - then, if the EHRs are so essential to safety, one would expect significant morbidity and mortality from these defects.

If one is to believe patients are not being injured by "glitches", then the expenditure of hundreds of billions of dollars for these systems on the basis of "error reduction" compared to paper is likely a waste of money and resources.

--  SS

Another Health IT "Glitch" - Can Digital Disappearing Ink Kill Patients?

Yes, it can.

There's been yet another "glitch" in the world of health IT (see http://hcrenewal.blogspot.com/search/label/glitch for more examples).

"Glitch" is a banal term used by health IT extremists (those who have abandoned a rigorous scientific approach to these medical devices as well as basic patient protections, in favor of unwarranted and inappropriate overconfidence and hyper-enthusiasm).  The term is used to represent potentially injurious and lethal problems with health IT, usually related to inadequate software vetting and perhaps even "sweatshop floor in foreign country directly to production for U.S. hospital floors" development processes (this industry is entirely unregulated).

This from Siemens Healthcare:

Click to enlarge.  Text below.

Text of this "Safety Advisory Notification":

August 1, 2013

Safety Advisory Notification
Soarian® Clinicals Medication Reconciliation EV06736602

Dear Customer:

This notification is to inform you that Soarian Clinicals Medication Reconciliation 3.3 may not be operating properly in some cases.

Although this may affect only some customers, we are taking a conservative approach and are alerting you to this potential problem. As such, please forward this notification to appropriate personnel as soon as possible.

This letter is being sent as a precautionary measure as there have been no adverse events reported from customers.

They mean "no adverse events reported - yet."  And if such events had been reported, Siemens would most certainly not make them public.  (Why should they, when there are no regulations?)

When does this issue occur and what are the potential risks?

This issue occurs when a user moves a free text in-house order from the current and home medications side (left side) to the discharge medication side (right side), and then modifies the continued free text in-house order in discharge reconciliation prior to saving the discharge reconciliation list. After the modification of the continued free text medication order, the changes to the free text medication order are not recorded in the saved discharge medication list.  [In other words, the changes to medication orders the user just typed disappear into thin air.  I note that medication reconciliation failures are among the most common causes of medical error - ed.]

The health IT extremists would invoke the "Leaned Intermediary" doctrine that lays all blame for errors on the user.  It seems the only way to avoid such liability would be after every "enter" or "save" action (or perhaps every keystroke?), users then verify what was saved or entered...

The "fix" to this "glitch" is not too far off from that:

Immediate steps you should take to avoid the potential risk of this issue:

To prevent this issue from occurring at your facility, instruct users not to modify continued free text in-house orders on the discharge medication list. Users may be instructed to enter free text in-house orders manually by selecting the add prescription action button and entering the order.

This is known as a "workaround."  Anyone who believes this edict can and will be 100% reliably followed in often chaotic medical environments, by users from medical students to nurses to physicians, is truly cavalier.

Steps that Siemens is taking to correct this complaint:

We are diligently working to develop a correction and will test and deliver it as soon as possible.

Perhaps FDA and Joint Commission need to inquire about exactly what testing and QC was done on the current code, testing that (if actually performed) did not detect this glaring and Siemens-admitted safety-risk "glitch."

-- SS