Implementing your ED Information System: The Reality

Some of you may recall that back in December I briefly mentioned turning on our ED information system on September 11 and then for obvious reasons, turning it off again. Well, on Tuesday, April 9, 2002 we tried it again and right now I’m sitting here with a large smile. Our ED information system (HealthmaticsED™ from A4HealthSystems) is up and running and working pretty well. The staff is tired but mostly upbeat, proud of themselves and their new skills. Justly so. And I’m sitting here, basking in the sense of accomplishment and achievement that I’ve spent nearly seven years at Maimonides reaching for. All is not perfect, there’s still lots to do, problems yet to solve, but we strike more issues off the top of our list every day than we add at the bottom.

One of our greatest challenges, particularly during the first two days we were up was with the creation of two entries on the tracking board. Our triage nurse would create an entry with the first encounter with the patient. This entry would put up an icon of a computer that would alert our registration clerk to the patient then needing registration. The registrar would start registration, but insert the patient’s medical record number into the triage record and then the interface would combine the formal registration in the hospital’s admission-discharge-transfer (ADT) system with the entry in HMED.

Well, we’ve discovered at least five different ways that this plan can fail. I’m not going to describe them for you, but I do think it is worth characterizing the way systems “fail” or at least have troublesome problems in the early days of “go-live.”

First and most commonly system failures occur during implementation because your staff is climbing the learning curve. People make mistakes and even though our staff has been trained (originally for our September 2001 scheduled go-live) and then retrained for this effort in April, people forget or never learned the proper procedures. Some will not truly learn the new process until they are in the midst of it and only then will they grasp the lessons taught in the classroom.

Next, and somewhat overlapping with our first cause, system failures occur because staff fails to adhere to new procedures or because the new procedures themselves are flawed. Distinguishing between these two causes is important, because staff errors are most likely evidence of need for more training and are remediable thereby. Yet, most people will prefer to displace blame onto process, blaming the procedure and possibly obscuring the best means to remediation.

Lastly, technical failures or software shortcomings produce sometimes subtle and sometimes obvious and potentially catastrophic failures depending upon the extent of the problem and whether its impact is restricted to a single workstation or process or more broadly affecting the use of the system.

The late W. Edwards Deming in point six of his fourteen points stated, “Institute Training and Retraining.” During our first few days of running the new system we have learned and relearned this lesson.

All of us who trained in emergency medicine (or any specialty or profession) trained for a fixed period of time. Deming always asserted that training for a fixed period when learning a skill or process was the wrong way to train: That skills should be taught until mastery is demonstrated. He advised “train and test, train and test.” By this he meant that staff should be trained until they demonstrate no further improvement in testing and that the purpose of testing was to evaluate for mastery, not for a grade. Yet in most cases we train for a fixed period, whether in residency or when learning the use of a new ED information system.

This past week we had MIS staff, vendor support staff and our own super users: emergency physicians and nurses, continuously on duty in the ED. As I’ve already mentioned, errors abounded, but we had staff to help train—on the spot—those who erred or were confused. Some detective work and learning about the system in each instance helped persuade the staff that the problems were related to the user, not the system. Corrections to the database and interface errors were the responsibility of other MIS and vendor staff, some on scene and others connected from A4HealthSystem’s headquarters in Cary, NC. We had the personnel on scene around the clock to support the staff that was delivering patient care. We trained and re-trained on the spot. We supported and encouraged staff, helped them learn how to accomplish the tasks for which they were responsible and praised them for what they did well.

Now after one week our MIS staff tells me that we are exceeding their expectations for our low error rate and success in gaining staff confidence. Many of our nurses are already asking to move on to nurse charting rather than waiting for the three months originally scheduled before we implement the next phase. All staff realizes the benefit for patients and for other staff in the hospital. Their enthusiasm, support for one another and persistence in working through their mistakes and identifying problems for our support team has been outstanding. They have performed magnificently and I am incredibly proud of all of them.

Yes, I am delighted. Thirteen years ago I saw an ED information system in a booth at a professional meeting. I knew the moment I saw it that this was a tool that I had to have in my ED. It has taken me since then to successfully bring my vision to reality. But my real lesson of the past week has once again been the realization that it’s the people who make it happen.

People are valuable. Technology is cheap.


Disclosure: I serve as a member of the medical advisory board to A4HealthSystems HealthMatics ED product. I receive no compensation for this service.