Two decades ago, Weed1 suggested that since knowledge was endless and constantly changing, physicians primarily should be skilled observers, rather than repositories of information. After making key observations, physicians should proceed in a disciplined manner, following decision trees to each logical end, leaving no open loops. This approach exposed an entire generation of physicians to the problem-oriented approach and the problem-oriented medical record.
Following that concept, we began a series called "Toward Optimal Laboratory Use" in JAMA in 1975.2 Articles in this series have focused on newly developed formal approaches to clinical decision making, such as using algorithms and decision tables when investigating common clinical problems.3
The status of clinical decision making continues to mature in theory. The Society for Medical Decision Making, now in its 10th year, publishes its own scholarly journal, Medical Decision Making. While decision theory has become more familiar, the diagnostic and therapeutic