Simple algorithms for predicting psychotropic drugs assigned to psychiatric inpatients.

The ability of simple algorithms, primarily diagnostic, to assign classes of psychotropic drugs to psychiatric inpatients in agreement with actual clinical decisions was investigated. Overall agreement (hit-rates) ranging from 48 to 65 percent were found across four major drug groups: major tranquilizer, antidepressant, minor tranquilizer, and no psychotropic drug. These hit-rates may be compared with multivariate formulae developed in earlier studies that achieved hit-rates ranging from 62 to 77 percent. Agreement among clinicians (about 70 percent) appears to be the limiting factor in most studies of this type.