++
Key Points
Medical knowledge has traditionally been memorized and the clinician is expected to keep that knowledge current.
New medical knowledge and “best practices” are developing at a rapid pace.
Decision support tools can be used to provide medical information and “best practice standards” at the point of care.
++
Traditional medical training requires the memorization of vast bodies of information and regurgitation or application more or less on demand. The best medical students are often those with the most prodigious memories: “roundsmanship” is prized. However, medical information has become so complicated and changes so rapidly that it is impossible to stay current. Additionally, the concept of “best-practice” is increasingly replacing that of “the way I do it,” and the definition of best practice is, in and of itself, difficult to keep up with. Many estimate that much of the information that a clinician graduates from school with is obsolete within a decade.
++
Computerized decision support (as contrasted with decision making) is the obvious answer. Decision support tools can be kept current by people or technologies dedicated to that task, and clinicians can tap in on that current information on demand, rather than needing to maintain their own, memory-based version of current medical knowledge (Fig. 11-1).
++
++
Key Points
Early decision support tools were developed at several places including the University of Pittsburgh and Stanford.
These systems were designed to codify bodies of medical information and act as consultants.
++
The first systems developed to assist in medical decision making date back to the early days of artificial intelligence in the 1970s and 80s, because the practice of medicine was recognized early on to be a logical application for smart technology.
++
Two systems designed to assist medical decision making deserve mention. The first was developed at the University of Pittsburgh, by Dr. Jack Myers and a computer engineer named Harry Pople, as well as Dr. Randall Miller and called INTERNIST-1. This system was developed in 1974 and designed to address diagnostic dilemmas in the fields of internal medicine and neurology. It used what was been characterized as the “hypotheco-deductive” approach, wherein observations about a patient, such as signs and symptoms, were used to deduce a set of compatible diseases. By prioritizing the possibilities and asking additional questions, the system narrowed the possibility list to a few possibilities or a single best choice. When compared to experts using cases from the New England Journal, the system performed competently, but it was never widely integrated into medical practice.
++
Another medical decision support system (DSS), MYCIN, was designed by Dr. Edward Shortliffe at Stanford University in the mid-1970s (Fig. ...