TY - JOUR
T1 - The relation between ethnicity and cognistat performance in males seeking substance use disorder treatment
AU - Schrimsher, Gregory W.
AU - O'Bryant, Sid E.
AU - Parker, Jefferson D.
AU - Burke, Randy S.
PY - 2005/10
Y1 - 2005/10
N2 - This study examined the relation between ethnicity and results obtained during standard administration of the Cognistat, a screening measure of cognitive functioning, in a sample of male veterans seeking substance use disorder treatment (n = 312). Results indicated that screening items for the Naming, Calculation, and Similarities subtests were missed significantly more frequently by African American compared to Caucasian veterans, although the frequency of identification as impaired on the full metric portion of this scales did not differ based on ethnicity. The mean scores on the metric items for these three scales were very similar between groups. Additionally, African American veterans diverted to the Constructional Ability metric items had significantly lower scores on items from that subtest. Logistic regression analyses of impairment identification on the various subtests indicated that ethnicity was a significant (p < .005) predictor beyond the covariates of age and years of education for the Calculation subtest, but only accounted for 4.8% of the variance. Based on the results of the current study, it is recommended that 1) the entire Cognistat be administered regardless of screening item performance to minimize the risk of potential ethnic or cultural based performance bias and 2) a prospective study of potential demographic bias based on comparing Cognistat screening results to a battery of specific neuropsychological assessments of the same constructs be performed to maximize the potential specificity and sensitivity of this assessment for all demographic groups.
AB - This study examined the relation between ethnicity and results obtained during standard administration of the Cognistat, a screening measure of cognitive functioning, in a sample of male veterans seeking substance use disorder treatment (n = 312). Results indicated that screening items for the Naming, Calculation, and Similarities subtests were missed significantly more frequently by African American compared to Caucasian veterans, although the frequency of identification as impaired on the full metric portion of this scales did not differ based on ethnicity. The mean scores on the metric items for these three scales were very similar between groups. Additionally, African American veterans diverted to the Constructional Ability metric items had significantly lower scores on items from that subtest. Logistic regression analyses of impairment identification on the various subtests indicated that ethnicity was a significant (p < .005) predictor beyond the covariates of age and years of education for the Calculation subtest, but only accounted for 4.8% of the variance. Based on the results of the current study, it is recommended that 1) the entire Cognistat be administered regardless of screening item performance to minimize the risk of potential ethnic or cultural based performance bias and 2) a prospective study of potential demographic bias based on comparing Cognistat screening results to a battery of specific neuropsychological assessments of the same constructs be performed to maximize the potential specificity and sensitivity of this assessment for all demographic groups.
UR - http://www.scopus.com/inward/record.url?scp=25644447573&partnerID=8YFLogxK
U2 - 10.1080/13803390490919173
DO - 10.1080/13803390490919173
M3 - Article
C2 - 16183620
AN - SCOPUS:25644447573
VL - 27
SP - 873
EP - 885
JO - Journal of Clinical and Experimental Neuropsychology
JF - Journal of Clinical and Experimental Neuropsychology
SN - 1380-3395
IS - 7
ER -