Errors at the speed of light

Physician Executive, July-August, 2005 by Russell Davignon

Since the well-publicized turn on, turn off of computerized physician order entry (CPOE) at Mount Sinai, the electronic medical record (EMR) has had some general negative press and lost some of its karma.

We have had components of an EMR at Central Vermont Medical Center in Barre, Vt., since 1995 when all laboratory data, diagnostic imaging reports and any dictated/typewritten records were entered into the computer system by various medical center employees.

Except for the ICU and birthing areas, all the medical/surgical nursing area care notes, physical therapy, dietary and similar notes were typewritten by the provider into the record for easy immediate access and review as well as guaranteeing a legible note, immediately accessible at multiple sites throughout the organization.

Admittedly, getting some physicians on board was difficult. I was still in active orthopedic practice and fairly young when the transition occurred but remember the joy at being able to find a complete blood count (CBC) quickly and review the prior CBCs with a keystroke.

Some older physicians were not so enthused and my partner of 25 years, who was nearing 70 when he retired in 2000, still had not learned even the fundamentals of the Meditech system we used when he passed from the practice scene. He circumvented the system and his need to learn by having the nursing staff or ward clerk print out anything he needed to see. So much for a paperless system.

Our physicians had become quite dependent on the electronic record for all but order writing and daily progress note functions when we began the next step of moving toward a truly total EMR in 2000.

We had developed a strategic plan outlining the steps we needed to take before we began the final push for a total EMR. Meditech, our longtime vendor, initially had storage capacity problems that forced the organization to purge data at various points in time. This was not a huge issue but did cause some problems when a record more than three years old was needed.

First step

Our first step was to assure an expandable, essentially infinite, write-protected storage system. One goal was to move to an electronic health record (EHR) that would include other patient-related information including billing, charge capture and other financial data elements that were patient specific.

Additionally, we wanted to move toward a record that would never need to be purged and would have the ability to access, retain and protect other institutional data elements such as policies, protocols and designated order sets--those approved by pharmacy and therapeutics committee as best practices, not the individual physician favorites that most CPOE software allows.

On a parallel track, a PACS (picture archiving and communication software) system for our diagnostic imaging department was also needed for a total EHR. This was an easier fix than might be assumed as most diagnostic imaging departments are already using electronic image capture in MRI and CAT scans. Radiologists are also already used to and engaged in the use of technology for imaging viewing, addition/subtraction techniques and image rotation.

This implementation amounted to little more than finding a vendor and converting current plain film techniques to electronic image capture, easily done with existing radiography equipment by changing the cassette technology.

Optical discs from Valco Data Systems and a FormFast electronic and paper form generating and bar coding system--both Meditech compatible vendors--were key underpinnings to the next steps.

[ILLUSTRATION OMITTED]

These were fairly straightforward determinations that required minimal training and acceptance on the part of medical center staff and none by physicians or nursing staff.

With these keystones in place, the information systems department began what became known as the paper gap analysis (PGA) to determine what paper was still being generated by the hospital for the remaining non-electronic portion of the medical record.

More paper

I had naively assumed that there wasn't much that wasn't already electronic when I began this process but I was very wrong indeed. We counted over 800 forms that were in use somewhere in the organization that needed cataloging and bar coding to be included in the medical record.

Central Vermont's commitment to an EHR was sorely tested by the next step toward CPOE. Finding a vendor for this next portion became somewhat of a search for pixie dust. We found, with great difficulty, a suitable vendor for an electronic medication administration record (e-MAR) and a bedside medication verification system (BMV).

The details of the latter were more difficult than one would initially imagine. Several on-site evaluations revealed significant issues with the wristband bar code readers and nurse workarounds of the drug bar code confirmations that need to be resolved.

Although a multi-disciplinary team was in place, it was key that several of our IS techs were RN's first. They knew the ropes, they knew the nurse workarounds and foresaw the problems. Their insight helped avoid some of the more common pitfalls of the medication administration side of an electronic record.