Journal of Personality Disorders, 26(6), 880-889, 2012 © 2012 The Guilford Press 880 This article was accepted under the editorship of Paul S. Links. From Brown Medical School, Rhode Island Hospital, Providence, Rhode Island. Address correspondence to Mark Zimmerman, MD, Bayside Medical Bldg., 235 Plain St., Providence, RI 02905; E-mail: mzimmerman@lifespan.org Revision of the Diagnostic Algorithm FOR BPD Zimmerman et al. 045 AN EMPIRICAL EXAMINATION OF GUNDERSON’S PROPOSED REVISION OF THE DIAGNOSTIC ALGORITHM FOR BORDERLINE PERSONALITY DISORDER Mark Zimmerman, MD, Kristy Dalrymple, PhD, Diane Young, PhD, Iwona Chelminski, PhD, and Jennifer Martinez, BA Gunderson (2010) recently offered a sharp criticism of the draft proposal for diagnosing personality disorders in DSM-5. Based on a review of phe- nomenological, factor analytic, social psychology, family, neurobiological, and treatment studies of borderline personality disorder (BPD), he proposed an alternative revision of the BPD criteria. One of the suggested changes was a modification of the DSM-IV diagnostic algorithm. Gunderson did not, however, provide any data on the impact this new diagnostic algorithm would have on the prevalence of BPD, or the validity of this alternative approach compared to the DSM-IV algorithm. In the present report from the Rhode Island Methods to Improve Diagnostic Assessment and Services (MIDAS) project we administered semi-structured diagnostic interviews to 3,081 psychiatric outpatients and examined diagnostic concordance be- tween DSM-IV and Gunderson’s proposal, and whether there is incremental validity in Gunderson’s diagnostic approach. The results did not indicate that the alternative diagnostic algorithm improved validity, and, depending on the threshold used, could result in false negative diagnoses. In reaction to the proposed changes towards diagnosing personality disor- ders (PDs) by the DSM-5 Work Group for Personality and Personality Disor- ders, Gunderson (2010) recommended an alternative proposal for changing the approach towards diagnosing borderline PD (BPD). Gunderson made several suggestions for modification, one of which was to reorganize the nine DSM-IV BPD criteria into four phenotype subsets (interpersonal hy- persensitivity, affective/emotional dysregulation, behavioral dyscontrol, and interpersonal), and diagnose BPD when five or more criteria were present with at least one criterion from at least three of the subsets. This differs from the DSM-IV algorithm which simply requires that any five criteria be present to make a diagnosis. Gunderson also suggested that requiring one criterion from each of the four subsets might be useful in narrowing the definition of