Wednesday, October 03, 2012

Allegheny Health System Computer Crash (Again) and Paper Backups

I reported on a health IT crash in my May 2011 post "Twelve Hour Health IT Glitch at Allegheny General Hospital - But Patients Unaffected, Of Course..."

Now, there's this at the same healthcare system:

Computer system at West Penn Allegheny restored after crash 
Liz Navratil
Pittsburgh Post-Gazette
October 2, 2012


The computer system at West Penn Allegheny Health System crashed about noon today, temporarily leaving doctors and nurses to work off of paper records instead.

Kelly Sorice, vice president of public relations for the health system, said all systems have since been restored. She said the servers crashed about noon today when the system experienced a power surge.

Doctors in the health system keep paper copies of almost all of their records so they can reference them during power outages or scheduled maintenance times, Ms. Sorice said.

Some systems were up eight hours later and others were expected to come online overnight, according to a report at HisTalk.

Assuming the statement about "doctors keep paper copies of almost all their records" was not spin control regarding skeletal paper records, a question arises.

Why, exactly, spend hundreds of millions of dollars on computing if paper records are kept, and are perfectly sufficient to accomplish the following, the usual refrain in health IT crash scenarios?

Ms. Sorice said she did not know of any procedures that had been rescheduled and added that, "Patient care has not been compromised."

As a physician/ham radio enthusiast who did an elective in Biomedical Engineering in medical school, I also want to know:

1)  What caused the “power surge?”
2)  Why were the systems not protected against a “power surge?”
3)  Exactly how did the “power surge” affect the IT?

Note: I've created a new, searchable indexing term for HIT outage stories with the usual refrain along the lines that "patient care has not been compromised." 

See this query link using the new indexing term.

-- SS

Addendum Oct. 3:

Australian EHR reseacher and professor Dr. Jon Patrick opines:

Even if [the paper records are] skeletal they suggest an endemic lack of confidence. I think the hospital spokesperson hasn't seen the implication of their statement.

-- SS

5 comments:

Anonymous said...


We have a new issue, age. Many of us, regardless of our business background, are leaving the work force and with that the ability to work from, or understand, a paper based system.

Steve Lucas

InformaticsMD said...

Steve Lucas writes:

Many of us, regardless of our business background, are leaving the work force and with that the ability to work from, or understand, a paper based system.

In medicine, current electronic systems often hamper understanding of clinical events.

As I've heretically written, a good or even fair paper system is probably better for healthcare than bad health IT (BHIT).

-- SS

Anonymous said...

We tried e-prescribing for 18 months before finally giving up on it (too many errors). The only way we could ever make it "work" was to always give the patient a paper script at the same time.

paulboal said...

I like that you've added the indexing term around "patient care has not been compromised." Have you considered putting together a "database" of sorts that can give us some reporting on all the outages you've been able to document? As someone responsible for implementing and supporting an EMR solution, it would be a great body of "lessons learned" to leverage and build safeguards around. If we have to go in this direction, I want all the data I can get to put forward the best possible solutions.

I would expect to be able to gather statistics like "average outage length" and group by a "category of causes" (fixed list), etc. Maybe a wiki that uses semantic tags? That would also give us the opportunity to collaborate on recommended safe guards to publish and share with the community.

Any response? You do such an amazing job collecting this information. Thank you.

InformaticsMD said...

paulboal said...

Have you considered putting together a "database" of sorts that can give us some reporting on all the outages you've been able to document?

I've thought about it, but wish I had the resources - and information sources - to do so.

These stories appear occasionally in the press, but to see most of them I'd need access to an aggregator such as Current Contents or Dialog which I do not have at present (it's very expensive).

I can say this: I expect the "outages" are (unlike what ONC says) not very rare, and will be getting worse as the HITECH Act increases speed of implementation and reduced caution.

(I note that our healthcare regulators should be performing the function you suggest, of course.)

-- SS