EHRs: Think Clinical Process, Not Automation

January 24, 2011

The technology that is readily available today is simply too rudimentary and ... today's practitioners lack the training and experience to make full use of it. If the inventive powers of developers and medical informaticians are not stifled by federal incentives that favor the persistence of obsolete technology, it will take another 10 to 20 years for the next generation of systems to emerge and mature.

In the future, medicine will be practiced in a way that information, encompassing discrete data elements, narrative, audio, video, etc., will be captured, stored and used as an integral part of delivering patient care. This is not possible today for several reasons. 

The technology that is readily available today is simply too rudimentary and, even if new technology suddenly appeared on the scene as Athena sprang forth fully armed from the head of Zeus, today's practitioners lack the training and experience to make full use of it. If the inventive powers of developers and medical informaticians are not stifled by federal incentives that favor the persistence of obsolete technology (like subsidizing corn ethanol), it will take another 10 to 20 years for the next generation of systems to emerge and mature. Similarly, if medical schools began intensive programs of training students to think in terms of data and information (also called computational thinking), a similar 10 to 20 year period will be required before the bulk of the physician community is fluent in this mode of practice.

(An example of the type of innovation I anticipate appeared recently here: "Primary inventor Christian Sax and co-inventor Hannas Lau recently developed a prototype of the keyboard for an iPad. The software is essentially a virtual keyboard that appears underneath a user's fingertips, corresponding to hand size and finger positions. The system works by sensing the position of the user's fingers and creating a fluid, QWERTY keyboard underneath. The keyboard also responds to position and pressure.")

Complicating matters, it is not unusual to discover that many consider that the goal of implementing an EHR is to "automate" the chart, or something to that effect. Having been on the front lines, implementing systems for clinicians to use, I learned quickly that the real objective was to find a way to re-invent the clinical workflow in ways that a computer could be incorporated into the process without causing excessive disruption - and even that was no guarantee that clinicians would accept the changes. A new computerized system is easy to defeat. You simply continue to do your work the old way, on paper, and let the institution and the change agent deal with it. If a few clinicians resist, it won't matter. If the entire medical staff resists, your project is DOA. The most damaging negative response is for the physicians to "go through the motions" of using the product while paying no attention to the accuracy and completeness of the information - a behavior that is actually encouraged by systems that create obstacles to efficient practice.

My first inkling that "automation" was the 600-pound gorilla in the room was during a meeting in 1997 with our CEO du jour. A new physical plant was being planned and one of the objectives was to achieve a virtually paperless operation. The CEO voiced the expectation that this would be accomplished by "automating" the medical record using a Hospital Information System. He knew of a facility in New York that had done just that. In fact, he said, he planned to send us to visit the site, see what they had done, and then return and recreate their success. As you might have guessed, no travel ever took place; there was no success to recreate.

What this episode revealed was the disparity between the view of paperless medical records held by the CEO and those practitioners that already had 15 years experience implementing workable systems. The CEO failed to appreciate that providing non-surgical medical care - and documenting it - are knowledge work. Work of this type bears no relationship to a manufacturing production line. If a worker is drilling holes by hand, a machine can be built that will accept an incoming stream of parts, drill the necessary holes in each and then send the parts along to the next production step. THAT is automation. It's not possible to automate knowledge work.

A few quotes with which I have taken a few liberties provide an excellent conclusion:

"The first rule of any technology used in a business is that [technology] applied to an efficient operation [can] magnify the efficiency. The second is that [technology] applied to an inefficient operation will magnify the inefficiency".
- Bill Gates, Microsoft founder

"There is nothing so useless as doing efficiently that which should not be done at all."
- Peter F. Drucker, management author, educator

"Few companies that installed computers to reduce the employment of clerks have realized their expectations.... They now need more, and more expensive clerks even though they call them 'operators' or [doctors]."
- Drucker