Improving the accuracy of clinical judgment.

Diagnostic reasoning in psychiatry is as intricate as it is risky (Elstein et al. 1978; Fitzhenry-Coor 1986; Fitzhenry-Coor and Nurcombe 1983; Gauron and Dickinson 1965; Sandifer et al. 1970; Kendell 1973; Nurcombe and Fitzhenry-Coor 1982). During the clinical encounter, the clinician elicits, notes, and verifies pertinent clinical cues. Early in the encounter, from an assembled but incomplete pattern of data, an array of diagnostic hypotheses is generated. Subsequently, in the light of evidence systematically gathered, each diagnostic possibility is progressively refined or deleted until the diagnostic conclusion is reached. The validity of the clinical endeavor, thus, depends upon the pertinence of the hypotheses generated, for it is upon them that deductive reasoning will hinge. It also depends upon the reliability of the history, signs, and investigations that constitute the evidence gathered for or against each of the hypotheses. The purpose of this paper is to propose systematic strategies based on principles of probability and empirical research, with the aim of sharpening clinical reasoning.