A prototype approach to personality disorder diagnosis.

Objective: Virtually no research has tested alternatives to the diagnostic method used since DSM-III, which requires decisions about the presence/absence of individual diagnostic criteria, followed by counting symptoms and applying cutoffs (the count/cutoff method). This study tested an alternative, prototype matching procedure designed to simplify diagnosis. The procedure was applied to personality disorders.

Method: A random national sample of psychiatrists and clinical psychologists (N=291) described a randomly selected patient in their care. Clinician-provided diagnostic data were used to generate categorical and dimensional DSM-IV diagnoses (number of symptoms present per disorder). Clinicians also used one of two prototype matching systems to provide a diagnosis for the selected patient.

Results: Prototype diagnosis led to reduced comorbidity relative to DSM-IV diagnosis, yielded similar estimates of validity in predicting criterion variables (adaptive functioning, treatment response, and etiology), and outperformed DSM-IV diagnosis in ratings of clinical utility and ease of use. Adding a personality health prototype further increased prediction.

Conclusions: A simple prototype matching procedure provides a viable alternative for improving diagnosis of personality disorders in clinical practice. Prototype diagnosis has multiple advantages, including ease of use, minimization of artifactual comorbidity, compatibility with naturally occurring cognitive processes, and ready translation into both categorical and dimensional diagnosis.