Citation

Penick EC, Nickel EJ, Powell BJ, Liskow BI, Campbell J, Dale TM, Hassanein RE, Noble EP. J. Stud. Alcohol 1999; 60(2): 188-202.

Affiliation

Department of Psychiatry and Behavioral Sciences, University of Kansas Medical Center, Kansas City 66160-7341, USA.

Copyright

(Copyright © 1999, Rutgers Center of Alcohol Studies)

DOI

unavailable

PMID

10091957

Abstract

OBJECTIVE: This study directly compared the clinical validity of 11 empirically defined alcoholism typologies to determine whether some typologies are clinically more valid than others. METHOD: A sample of 360 hospitalized alcoholic men were extensively evaluated at entry into the study and again 1 year later. Twenty-three measures of clinical validity were employed; 15 were postdictive and 8 were predictive. Postdictive retrospective measures obtained at entry into the study included family history, age of onset and lifetime course characteristics associated with alcoholism severity, general psychopathology and psychosocial functioning. Predictive outcome measures drawn from information obtained during the 1-year follow-up included: abstinence, alcoholism severity and clinician ratings of outcome. The measures were subjected to various statistical analyses, including factor analysis. RESULTS: We found that all of the alcoholism typologies met at least 7 of the 23 a priori measures of clinical validity. The correlations between these conceptually and methodologically disparate typologies were often striking. Exploratory factor analysis, which explained 35% of the variance, suggested three possible underlying dimensions to account for the overlap among typologies: (1) age and its correlates, including age-of-alcoholism onset; (2) "pure" alcoholism versus psychiatrically heterogeneous alcoholism that encompassed antisocial personality disorder; and (3) current severity of psychiatric distress, impairment and dysfunction. CONCLUSIONS: No single method of subtyping alcoholics clearly emerged as superior. All demonstrated some degree of predictive and postdictive clinical validity. Most methods of subtyping correlated positively with each other at moderate, but typically significant, levels.

Language: en