Diagnostic Support in Electronic Patient Records

Given that American medicine now ranks at some 35th or 36th in longevity and infant mortality, and at best 15% of diagnoses are wrong, some form of clinical diagnostic decision support seems warranted. Autopsy went out of fashion for many reasons. It was once the final arbitrator of quality medicine and arguably lead to both modern scientific medicine and the high quality of our medical schools. Electronic records offer some hope of restoring a measure of that quality support.


Within the electronic patient record, a differential diagnostic listing covering all of the possibilities might give the patient greater assurance that: over confidence, snap diagnosis or more conveniently reimbursable diagnosis, will not lead to some unfortunate outcome. With a sufficient differential diagnostic listing, the physician will likely consider the person’s true condition, even the rare ones.

Problem oriented charting went a long way to meet the need for considering all of the patient’s problems. It introduced a level of broader consideration of both subjective and objective findings before offering an assessment and finally a diagnosis. However, this list of problems, symptoms and findings with a considered assessment may point to many underlying possibilities.

Listing all of these possibilities in a statistically weighted manner supports a considerably higher confidence and probability of accuracy in the final diagnosis. Treatment protocols offer little, if the clinician makes the wrong diagnosis.

The best of physicians realize that medicine is an art and diagnosis often allusive. They will welcome a diagnostic tool if they find it accurate and useful. Sir William Osler at Hopkins challenged his students to look deeply for underlying diagnosis when considering a number of superficial problems. He did the same in his classical textbook, Principals and Practice of Medicine in 1892.