Clinical Dispositions of Suicidal Patients in an Emergency Room: Neural Networks as Clinicians? University of California, San Francisco, Fresno-Central San Joaquin Valley Medical Education Program, Fresno, California, (Senior Residency Project). ( 0/1991 - 0/1992 )
Patients reporting suicidal ideation seen in the emergency department (ED) were modeled using a backpropagation neural network computer simulation to "train" neural nets to associate demographic/clinical data with clinical disposition from the ED (i.e, whether or not to hospitalize). Once the network was trained (through modification of "synaptic" strengths), demographic/clinical information for a "new" patient was then entered as input. The network was then allowed to "run" and render an output, viz., a recommendation whether to hospitalize the patient. This recommendation was then validated against the disposition that had actually been rendered by the clinician.