Refining personality disorder subtypes and classification using finite mixture modeling

• Published in: Personality disorders Vol. 4; no. 2; p. 121
• Main Authors: Yun, Rebecca J, Stern, Barry L, Lenzenweger, Mark F, Tiersky, Lana A
• Format: Journal Article
• Language: English
• Published: United States 01.04.2013
• Subjects: Adult; Aggression; Antisocial Personality Disorder - diagnosis; Antisocial Personality Disorder - psychology; Diagnostic and Statistical Manual of Mental Disorders; Female; Humans; Male; Models, Psychological; Paranoid Personality Disorder - diagnosis; Paranoid Personality Disorder - psychology; Personality Assessment; Personality Disorders - classification; Personality Disorders - diagnosis; Personality Disorders - psychology; Reproducibility of Results
• ISSN: 1949-2723
• DOI: 10.1037/a0029944

The current Diagnostic and Statistical Manual of Mental Disorders (DSM) diagnostic system for Axis II disorders continues to be characterized by considerable heterogeneity and poor discriminant validity. Such problems impede accurate personality disorder (PD) diagnosis. As a result, alternative assessment tools are often used in conjunction with the DSM. One popular framework is the object relational model developed by Kernberg and his colleagues (J. F. Clarkin, M. F. Lenzenweger, F. Yeomans, K. N. Levy, & O. F. Kernberg, 2007, An object relations model of borderline pathology, Journal of Personality Disorders, Vol. 21, pp. 474-499; O. F. Kernberg, 1984, Severe Personality Disorders, New Haven, CT: Yale University Press; O. F. Kernberg & E. Caligor, 2005, A psychoanalytic theory of personality disorders, in M. F. Lenzenweger & J. F. Clarkin, Eds., Major Theories of Personality Disorder, New York, NY: Guilford Press). Drawing on this model and empirical studies thereof, the current study attempted to clarify Kernberg's (1984) PD taxonomy and identify subtypes within a sample with varying levels of personality pathology using finite mixture modeling. Subjects (N = 141) were recruited to represent a wide range of pathology. The finite mixture modeling results indicated that 3 components were harbored within the variables analyzed. Group 1 was characterized by low levels of antisocial, paranoid, and aggressive features, and Group 2 was characterized by elevated paranoid features. Group 3 revealed the highest levels across the 3 variables. The validity of the obtained solution was then evaluated by reference to a variety of external measures that supported the validity of the identified grouping structure. Findings generally appear congruent with previous research, which argued that a PD taxonomy based on paranoid, aggressive, and antisocial features is a viable supplement to current diagnostic systems. Our study suggests that Kernberg's object relational model offers a plausible substantive aid in refining PD classification.