Monday, July 20, 2009

Philadelphia VA Hospital Brachytherapy Debacle: For The Want of One Competent and Industrious IT Person

This post should perhaps be subtitled "The Theatre of the Absurd."

At Healthcare Renewal and other sites I've often commented on the remarkable accommodation given to IT personnel in hospitals, even when these personnel make capricious decisions that are contrary to the support of the mission of healthcare organizations, or contrary to the mission itself.

I wrote that many HIT problems observationally appeared due to ill-informed, capricious edicts of overempowered (relative to clinical leadership) MIS leaders, sanctioned by equally ill-informed executive leadership, for which the IT personnel were rarely held accountable.

Endangerment of ICU patients via PC's entirely inappropriate for a biohazards environment, chaos in a critical procedure area from IT complacency and incompetence, payment of millions of dollars for HIT with gross defects rendering it unusable by clinicians without consultation of in-house medical informatics expertise, and denial of access for hundreds of drug discovery scientists to the informatics tools their own leaders said were essential to new drug discovery are just a few of the situations I've personally observed due to IT department whimsy.

Remarkably, none of these situations nor others reported from numerous sources resulted in repercussions against the teflon-coated IT deities (other than, in some cases, generous promotions), thereby obstructing remediation of the attitudinal and competency problems that created the scenarios. If physicians had it this good, they'd be chopping off wrong limbs, removing the wrong organs, and failing to diagnose and treat with impunity.

At my June 2009 posting regarding an unfolding clinical debacle of national import at the Philadelphia VA hospital entitled "Bungled Brachytherapy, Computer Interfaces and Other Mysteries At The Philadelphia Veterans Administration Hospital" I wrote:

... I am uncertain how "computer interface problems" (in the Philadelphia Inquirer, they were referred to as "glitches") prevented medical personnel from determining treatment success over several years. I would be most interested in hearing more about these "interface problems."

That question has now been answered, and I am disappointed (but not surprised) at the results. Having rewritten a poor interface between a gamma scintillation camera and a PDP-11 computer in medical school during a nuclear medicine rotation at Boston City Hospital, allowing endusers to identify ROI's (regions of interest) with a light pen for automated calculation of intensities, I thought perhaps some arcane coding or driver configuration was the culprit at the VAMC Philadelphia. That was not the case:

VA radiation errors laid to offline computer

By Marie McCullough and Josh Goldstein

Philadelphia Inquirer
July 19, 2009

... It is not surprising, then, that NRC and VA investigators spent considerable time delving into why the calculations weren't done for more than a year at the Philadelphia VA.

Their investigative reports blamed a "computer interface problem" - the same terminology Kao used during his testimony last month at a congressional hearing.

The implication was that some intractable technology breakdown was behind the lapse in care [i.e., some cryptic problem requiring magical incantations and byzantine scripts that mere mortals could not understand nor remedy - ed.]

In fact, technology had little to do with the breakdown, as James Bagian discovered when he led an inquiry at the Philadelphia VA and the veterans' health system's 12 other brachytherapy programs.

Bagian, a Philadelphia-born physician and former astronaut who is now the national VA's patient-safety director, discovered that the "interface problem" was nothing more than the disconnected computer.

Here's what else his inquiry found:

The computer was initially unplugged so that another medical device could use the network port. Then, various departments dithered and ducked a request for an additional network port, which was finally installed - after a year. ["Various departments dithered and ducked?" Which departments, exactly? See below - ed.]

Some doctors, physicists, and other professionals at the VA acknowledged it was "clinically inappropriate" to omit the post-implant calculations. Some said they had informed their "chain of command."

When asked why they didn't tell the hospital's patient-safety officer, they said "it had not occurred to them to do so."

["Had not occurred to them to do so?" I've seen situations where physicians and scientists were afraid of IT leaders, due to the latter's often overinflated political influence and proficiency in playing games of political intrigue. Did this occur here, I wonder? - ed.]


So, it appears that for the want of one network connection, installed by one competent, industrious, non-complacent IT person, a national scandal has erupted that has injured many patients.

First, I ask the following questions regarding the "various departments" that "dithered and ducked" the responsibility to install an additional network connection in the brachytherapy suite:

  • Was the job the responsibility of the Department of Medicine? The doctors and nurses? Could they have done so?
  • Perhaps it was the responsibility of the Public Relations Department (who now have to pick up the pieces?)
  • Was it the reponsibility of the Facilities Department?
  • Was it the responsibility of the Housekeeping Department?
  • Or, was it the responsibility of the IT Department and CIO?

(If you need help answering these questions, stop reading now.)

I also ask:

  • Why was the IT department's failing to perform this task kept low profile through use of cryptic "interface problem" language that implied complex IT problems, as opposed to simple people problems?
  • Who originated this language?
  • Did they believe the truth could remain hidden?
  • Why do we use the term "medical malpractice" to describe negligence in medical care, not "provider glitch?" Why the different standards?

I can visualize what went on behind the scenes in the IT department, refrains I have heard before - "we don't have the resources ... we need to hold more meetings to consider the issues ... it's the vendor's responsibility ... it's the network group's job, not the hardware group ... putting a new network outlet in there will cause packet storms and interfere with system XYZ ... we need to get consensus .... you [doctors] can't understand the complexities of the problem, but don't worry, we'll make it better ... hey, the docs don't need it anyway ..."

What would happen to, say, the Facilities Department if a plumbing problem led to failure of a piece of vital equipment in the OR's, and they failed to repair it for a year?

This is not to excuse the multiple layers of complacency among clinicians, safety staff, and others for toleration of this network denial situation and the lack of QC on the procedures themselves, but at the heart of this debacle is this attitude, which seems common in hospital IT departments:

"Doctors and nurses toil in hospitals so IT personnel can have comfy jobs and nifty computers."

Perhaps on this occasion, IT leaders and personnel may end up on the witness stand and be held accountable, and perhaps lose their jobs instead of being promoted. However, even this is doubtful, and it has taken a congressional investigation led by Sen. Arlen Specter to get even this far, to simply find out that the mysterious "interface glitches" were a lack of a network jack due to laziness and complacency.

Finally, I point out that it is hospital IT personnel upon whom clinicians will depend for acquisition and implementation of the HIT tools the President of the United States said are essential to changing the culture of healthcare.

As I have written, before the IT profession can change the culture of healthcare (in a positive manner, that is), its own culture must change.

IT personnel in hospitals must become part of the clinical team and support the mission of clinicians, not the other way around.

July 22 addendum:

I note again that multiple people failed here, including executives, physicians, safety officers, etc. I believe responsibility needs to be fairly assigned. However, IT needs to be included. I've witnessed or heard about too many HIT incidents where IT personnel and leadership remained scot-free when their behavior and attitudes were contributors or root causes.

As per my letter to the editor published in JAMA on July 22, "Health Care Information Technology, Hospital Responsibilities, and Joint Commission Standards", IT privilege and accommodation must stop. HIT is not business IT used for widget inventory or payroll. As the VA incident shows, patient lives and well being are at stake.

July 24 addendum:

More on this from a blog on medical physics, "The Sharp End of the Photon", here:

... the errors in the placement of the radioactive seeds went undiscovered for so long because post-implant dosimetry was not performed. This involves CT scanning the patient, finding the positions of the seeds and calculating the ultimate dose the patient received. In the NRC report, the explanation was that a problem with the interface between the CT scanner and the treatment planning computer prevented transferring the CT images.

A "problem with the interface", indeed.

Also provided is a link to testimonies from witnesses at the house.gov website, which are here.

Notable is the absence of any testimony from IT leadership. The only allusion to IT is in testimony by the doctor who performed the procedures who stated:

"There should be a method of categorizing systematic problems by level of urgency so that serious problems, such as those involving failures of medical equipment or transfer of patient-related data, will receive immediate attention from the proper personnel and be quickly resolved."

Perhaps, but not in this case. A simple phone call to IT should have been adequate to resolve this particular simple problem.

-- SS