Citation

Bates ME. J. Stud. Alcohol 1997; 58(6): 617-621.

Affiliation

Center of Alcohol Studies, Rutgers University, Piscataway, New Jersey 08855-0969, USA.

Erratum On

J Stud Alcohol 1998 Mar;59(2):236

Copyright

(Copyright © 1997, Rutgers Center of Alcohol Studies)

DOI

unavailable

PMID

9391921

Abstract

OBJECTIVE: Current trends in managed mental health care have telescoped the assessment and treatment of individuals diagnosed with an alcohol or other drug use disorder. Yet, there is limited empirical information about the short-term stability of neuropsychological status and other person characteristics that are useful to assess early in treatment. This study examined the stability of neuropsychological test scores within the first 3 weeks following diagnosis of an alcohol use disorder. METHOD: An eclectic neuropsychological battery made up of commonly used, sensitive tests of abstraction, executive functions, memory, visuospatial abilities and verbal ability was administered to female and male alcohol use disordered individuals within 3 days of treatment entry (or following detoxification), 3-5 days later and 21 days later. The three test administrations were completed by 35, 32 and 24 subjects, respectively. RESULTS: Across tests, the average stability coefficient (Pearson correlation) was .82 between Days 3 and 5, .86 between Days 5 and 21, and .79 between Days 3 and 21. Intraclass correlations ranged from .79 to .98 across tests (mean = .92). Clinical stability, defined as the likelihood that a test score fell consistently above or below a standardized impairement cutoff score, was also good. Across tests, percent agreement in impairment diagnoses for the same three time intervals averaged 84%, 92% and 87%, respectively. The chance-corrected kappa (Kappa) coefficients of diagnostic agreement were generally moderate to substantial from Day 3 to Days 5 or 21, and mostly substantial from Day 5 to 21. CONCLUSIONS: Early assessments of neuropsychological status were psychometrically stable, and also provided reasonably stable indicants of clinically significant impairment. It was likely that the data provided lower bound estimates of the stability of impairment classifications due to the repeated measures design and power limitations.

Language: en