Rates and correlates of pathological gambling among VA mental health service users

Abstract

The rate and correlates of diagnosed pathological gambling (PG) among mental health patients in the Veterans Health Administration, the only national system of mental health care, have not been studied. Using fiscal year 2009 (FY2009) VA administrative data, a case-control study compared those with an ICD code of 312.31 (PG) versus those without. The analytic group was limited to 1,102,846 Veterans Affairs (VA) specialty mental health (MH) services users because 94.5% of all those diagnosed with PG in the U.S. VA health system received such services. Chi-square tests and logistic regression assessed associations between demographic and clinical factors and PG diagnosis. The past-year rate of PG diagnosis among veterans treated in specialty MH program was 0.2%, significantly lower than prevalence rates in other treatment samples and the general U.S population, suggesting under-diagnosis and/or a low-income sample. Being female, ages 40-74, and higher income increased the risk of PG diagnosis, as did past-year homelessness (Odds Ratio (OR) = 2.2), alcohol use disorders (OR = 2.8), bipolar disorder (OR = 2.1) and personality disorders (OR = 2.1). Depression, schizophrenia, and anxiety disorders other than PTSD, were also positively associated with PG diagnosis. Drug use disorder had no significant independent association with PG. PTSD, dementia, and living in isolated rural areas conferred reduced risk. More systematic screening and surveillance of PG among MH service users generally, and veterans with heavy alcohol use, severe mental illness, and homelessness specifically, appears warranted.