Don't Treat Problems via EHR without an Accurate Diagnosis First

If we apply Pareto's 80/20 rule to Medicine, the 20 percent represents what we know or understand with reasonable certainty. For the other 80 percent, we have no definite information, only biases, prejudices, opinions, anecdotal information, or total ignorance. It's unfortunate, but that's life. If anyone believes that this does not describe them, I apologize for projecting my personal feelings - your experience may be part of the 80 percent about which I, personally, am uncertain or ignorant.

For the rest of us, the 80/20 applies to each new complaint or situation that a patient presents, and perhaps even some that are not new. The challenge is: What to do at those times (the 80 percent) when you are uncertain?

The first thing to do is to classify the reasons for uncertainty. They include:
• The complaint, though real to the patient, is not the result of a process that warrants specific treatment;
• Inadequate information about the circumstances;
• Inadequate medical knowledge and theory, either personally or generally;
• The process is evolving and has yet to fully manifest itself; and/or
• The complaint involves anatomy that is difficult to examine or visualize.

The second thing to do is to classify the responses that are possible when faced with uncertainty, which include:
• Follow the crowd and do what everyone else is doing;
• Wait for more/better information;
• Suggest a placebo or other "harmless" intervention such as "Take two aspirin and call me in the morning" and hope that it is truly harmless;
• Assume that the complaint, the true nature of which has yet to determined, is caused by a bacterial infection and give antibiotics;
• Get help or refer the patient to a specialist; and/or
• Admit the patient for a "million dollar" workup.

As we all know, judging from the way some physicians practice, when they are unsure about the nature of a patient’s problem, they assume that the answer is antibiotics. I described such a situation a few weeks ago in which two patients that turned out to have botulism were initially given antibiotics.

For some time now, many healthcare facilities have been suffering from the organizational equivalent of an illness. The economy and healthcare reform have increased both the number that feel afflicted and the severity of their complaints. Like patients that are inappropriately given antibiotics because "Something has to be done, and now," organizations are being told by the government and their doctors (consultants) to "take an EHR and see if that helps." Some organizations don't see a doctor and instead self-prescribe an EHR.

In either case, few take the time and effort to understand the true nature of their complaints or to determine that a particular EHR will actually treat the problems that they have identified without causing other, more serious ones. Considerable time and effort are spent, but the fact that the chosen solution is almost always the same suggests that neither the analysis nor the planned solution is individualized.

The spirit of evidence-based medicine is to limit yourself, as much as possible, to doing things that are well understood and for which there is supporting evidence. It is often more prudent to do nothing than to take randomly selected actions simply to satisfy the urge to be "doing something." There are times, of course, where an unknown condition needs to be approached as a small clinical research problem but these situations call for careful and detailed documentation of both the reasoning behind the decision and the results. Fortunately, in the hands of skilled practitioners, the information that has been developed by years of medical research is sufficient to cover most common situations and yet, not a year goes by that research does not yield a new discovery or a refinement of existing knowledge.

One would expect those skilled in the art of evidence-based medicine to approach other problems, such as choosing an EHR, in a similar manner. Unfortunately, very little is known about the science of medical data and electronic medical records and little basic research is being conducted. That which is known appears to be known only to a few. Fascinated by the potential of healthcare computing, many have simply begun to develop systems in an ad hoc [^1] manner. Few have been willing to take the slower, more deliberate approach of conducting research.

Much of what we know has been learned from a succession of spectacular failures, frustrated expectations, and cost overruns - not from careful study. Failure and disappointment remain all too common. The only aspect of healthcare computing about which there seems to be no great sense of urgency is basic research but, until the fundamentals are discovered and organized into a coherent body of theory that can be used to guide development, the pattern of failure, and frustrated expectations can be expected to continue.

[^1] Wikipedia defines ad hoc as "generally signifying a solution designed for a specific problem or task, non-generalizable, and not intended to be able to be adapted to other purposes."