On criterion weighting in the DSM-IV

Differential weighting of illness signs and symptoms has surfaced recurrently in psychiatric nosology. Six alternately weighted algorithms for diagnosing alcohol dependence in accordance with the Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association, 1994), based on statistical, unit, rational and random criterion weighting systems, were used to predict an array of concurrent validators and 6-month drinking outcomes in a regional clinical sample of 365 participants. Comparable predictive efficiency across all algorithms, including the randomly weighted versus statistical best-fit model, was observed. Further analyses and geometric modeling suggested that this was due to the extremely high internal consistency of the DSM-IV criteria. An alternative strategy that favors factorially complex, less homogeneous criteria was used to develop an experimental DSM-IV algorithm from an array of 39 candidate criteria. This algorithm had extremely low internal consistency, high difficulty, and complex factor loadings. Differential weighting of its criteria produced a good range of efficiencies, predictive power for rational models exceeding the random weight model, and a best-fit algorithm with substantial surplus predictive power. These results illustrate an emerging conflict in nosology between 2 opposing trends: a press for the promulgation of criterion arrays with high internal consistency and a clear desire to assign some criteria extra weight for prognosis or decision making. Both cannot be had in the same algorithm. An alternative approach emphasizing diagnostic criteria with complex structures can satisfy the multiple demands of brevity, validity, and weighting performance.