R. Austin Freeman qualified as a physician in 1887. In 1907, he wrote the first of many crime mysteries revolving around Dr. Thorndyke, a practitioner of forensic medicine. Freeman may be the father of the CSI genre. In one of his stories, "The Mystery of 31 New Inn," he gives us a clear demonstration of the major challenge with which physicians contend every day: How to keep many balls in the air and yet give each case the care that it deserves. I quote:
"One of the conditions of medical practice is the capability of transferring one's attention at a moment's notice from one set of circumstances to another equally important but entirely unrelated. At each visit on his round, the practitioner finds himself concerned with a particular, self-contained group of phenomena which he must consider at the moment with the utmost concentration, but which he must instantly dismiss from his mind as he moves on to the next case. It is a difficult habit to acquire; for an important, distressing, or obscure case is apt to take possession of the consciousness and hinder the exercise of attention that succeeding cases demand; but experience shows the faculty to be indispensable, and the practitioner learns in time to forget everything but the patient with whose condition he is occupied at the moment."
When physicians today complain about their EHR, they may incriminate the user interface or say it doesn't do what they want it to do or raise myriad similar objections. In many cases, the root cause is that they are trying to multi-task (as Freeman understood) but the EHR forces them to single-task. The work flow imposed by most EHRs is essentially linear and usually obscured because it leads them down many side alleys in the process.
This is why the typical EHR slows down the physician. This is why they end up forgetting to do important things that came to mind but were long forgotten by the time they were freed from the linear straitjacket they had donned when they began a particular task.
The value of windows on a computer is not that they create some sort of metaphor that is supposed to make the computer intuitively obvious, it is that windows enable multi-tasking. Each window can host a separate process or program and all can run (or wait patiently) while the user attends to one window in particular.
With a proper design, a physician could begin seeing and charting on one patient and, if an emergency arose or there was some other immediate need to look at, or annotate, another patient's record, the physician could merely open another window, tend to that business, and return to their prior activities. They could defer the completion of a note while awaiting some point-of-care test results or a call back from a consultant.
The typical EHR uses windows to segregate the display of data about a single patient but their design assumes that the physician's work flow is linear not chaotic; they assume single-tasking rather than multi-tasking and consequently they don't support it.
The reasons for this limitation are pretty obvious. Some of the most "popular" EHRs began life on dumb serial terminals where graphical windows and multi-tasking were simply not possible. Today those EHRs display their applications in windows but the apps continue to function in the same old, obsolete ways. Other EHRs are web-based and whatever advantages the web may offer, effective multi-tasking is not one of them.
EHR vendors and customers both have difficulty recognizing that their existing applications are not a valuable asset. Many are dysfunctional and represent a sunk cost. The value — if any — that was derived from building those applications is the clinical knowledge accumulated, an understanding of the work flow, and of physicians' personalities. The most effective way to get a return on the investment made to gain that understanding would be to build new applications that are freed from 30-year-old conceptual boxes and which fully leverage the latest technology. Remember, the mere fact that a product satisfies a long checklist of criteria doesn't make it useful if the intended users can't or won't use it.