Purpose: Pathological gambling disorder (PGD), characterized by persistent, recurrent, and disruptive maladaptive gambling, affects 1%-2% of adult gamblers. Public health concerns center on the disproportionate representation (5%-63%) and impact of PGD and problem gambling for populations marginalized due to minority status and health disparities.^1-2 These disparities, specifically among African-Americans, may be real or possibly artifacts of how data are collected and how diagnostic criteria are conceptualized and measured, thus requiring closer examination.³ Such attention may facilitate accurate detection and appropriately-targeted interventions.⁴ This presentation assesses whether racial/ethnic differences exist in DSM-IV PGD diagnosis (at least 5 out of 10 criteria), operationalized by the C-GAM^©.⁵

Methods: Using targeted community advertising in St. Louis, Missouri and bordering areas in Illinois, we recruited a diverse sample of Caucasians (n=227) and African-Americans/Other minorities (n=87) (mean age=46.7 yrs.; SD=17.0; range 18-85; 44% female; 55% £ HS/ GED; 43% employed; 36% married; 28% annual household incomes <25,000; 36% city dwellers) who gambled or bet more than 5 times (lifetime) on any one of 11 different gambling activity classifications. Participants completed initial and repeat telephone interviews, held 1-week apart (mean=7.2 days; SD=2.8), about their gambling and associated behaviors. Diagnostic test-retest reliability was estimated using kappa(Κ) indicating agreement that was either poor(<0.40); fair(0.40-0.60), good(0.60-0.75) or excellent(> 0.75).⁶

Results: We found good to excellent reliability for DSM-IV PGD for Caucasians (Κ=0.77) and African-Americans (Κ=0.62), with fair to excellent agreement for activity-specific PGD [Caucasians: lottery (Κ=1.0), own skill/chance games (Κ=1.0); cards (Κ=0.64), African-Americans: slot/fruit machines (Κ=0.62), craps/dice games (Κ=0.54), pari-mutuels (Κ=1.0).] Racial/ethnic variation existed for specific DSM-IV criteria, although kappas were all within the fair to excellent range. We will further examine these findings, controlling for potentially confounders and considering self-reported reasons for discrepant responses using the Discrepancy Interview Protocol⁷ modified for use in this study.

Implications: While racial/ethnic variation was observed, initial results indicate that African-Americans and Caucasians have comparably good to excellent reliability in their self-reports of DSM-IV PGD as measured by the C-GAM^©. Such information may be informative for researchers considering possible criteria revisions for DSM-V. These findings are also both clinically and practically meaningful for social workers requiring additional tools, such as the C-GAM^© for increased accuracy in diagnostic assessment, thus informing the implementation of “best practice” intervention approaches targeted to racially/ethnically diverse populations.

References:

¹NRC. Pathological Gambling: A Critical Review. Washington, DC: NAS Press; 1999.

²Cunningham-Williams R.M. et al. Taking Chances: Problem Gambling and mental health disorders-Results from the St. Louis ECA Study. AJPH 88(7):1093-1096; 1998.

³Horton J et al. Reliability of substance use disorders among African Americans and Caucasians. DAD; 57(3):203-209; 2000.

⁴Gaw AC. Concise Guide to Cross-cultural Psychiatry; Washington, DC: APPI; 2001.

⁵Cunningham-Williams, R.M.., et al. Computerized Gambling Assessment Module (C-GAM), Washington University, St. Louis, MO. September 2003.

⁶Cohen JA, A coefficient of agreement for nominal scales, Educ Psych Meas, 20:37-46;1960.

⁷Cottler, LB et al. The Discrepancy Interview Protocol: A Method for Evaluating and Interpreting Discordant Survey Responses. Inter J Methods Psych Res, 4:173-182, 1994.