Monday, March 7, 2011

EHR ED's in New South Wales. Will the Problems Magically "Disappear?"

It occurs that one could look at Prof. Jon Patrick's recent health IT forensic analysis as a kind of "indictment" of the industry.

He can be seen as suggesting the industry needs to be "put on trial" (figuratively) regarding "crimes" (again, figuratively speaking) they've committed with regard to IT robustness and reliability. The latter translate directly to patient safety.

In a lawsuit such as a medical malpractice trial, obvious as well as potential evidence is put under "legal hold." For instance, if an EHR defect is suspected, metadata, audit trails, and patient data are asked (or should be asked) to be frozen or archived in the state they were in at the time of the alleged accident.

It can take a page or three (or more) of specifications simply to define what information, exactly, needs to be put on legal hold. My former staff were frequently required to place myriad materials on legal hold at Merck Research Labs, for example, baed on the lawsuit du jour.

Once frozen, discovery and forensic analysis of these now-static data can then proceed. In fact, cases can be lost on the basis of evidence of an archiving omissions, destruction or tampering when information holds are requested.

It occurs to me that Prof. Patrick has given the industry a detailed look into factors they could start to remediate, without publicity and without telling anyone. While this would be a net plus for patients, it might result in less of a learning experience to the industy than that industry needs, to motivate the industry to avoid future product engineering and quality issues and put quality (not simply margin) as priority #1.

I therefore would believe a "hold" put on the present state of these ED EMR systems, or a "snapshot" of their current state (i.e., an evaluation environment mirrored from the present operational one) would allow a careful evaluation of the impact of the issues noted in the study.

Such an evaluation would be far more difficult with a cybernetic moving target.

The "snapshot" idea would allow evaluation of system risk levels, intermittent "glitches", interference with workflows, etc. in a controlled testing environment, using mock data or data drawn from actual cases.

The "snapshot" approach would also allow incremental remediation of the "live" system that comes out of safe, controlled testing, rather than sticking with what exists now until the studies could be completed on the as-is system, and then applying all the fixes as one or more large "upgrades."

I, for one, would be interested in studying this "built by software professionals" system and comparing it to health IT systems we "academic nerds" were authoring, say, 10-15 years ago.

-- SS