Risk for HIV Infection Among Medicaid Beneficiaries with Serious Mental Illness

Background and Purpose: The extent to which serious mental illnesses (SMI) such as schizophrenia, major depressive disorder (MDD), or bipolar disorder increase risk for Human Immunodeficiency Virus (HIV) infection is poorly understood. Some evidence suggests that individuals with SMI are more likely to contract HIV than other persons, perhaps because of unsafe sexual practices (Cournos, Guido, Coomaraswamy, 1994; Goodman & Fallot, 1998; Otto-Salaj, Heckman, Stevenson, 1998) or the poverty-related need to live in impoverished communities with higher rates of HIV infection (Blank & Eisenberg, 2007; Canton, Shrout, Dominguez, 1995; Goldberg, Rollins, & Lehman, 2003). However in the absence of substance abuse (which is often elevated among those with SMI), it remains unclear whether SMI an independent risk factor for HIV infection. In large part, cross-sectional evidence is the basis for the widely held belief that SMI increases the risk of contracting HIV. We examined the longitudinal relationship between SMI and subsequent HIV/AIDS diagnosis.

Method: Using data on 6,417, 676 Medicaid beneficiaries in eight states (CA, FL, GA, IL, NJ, NY, OH, TX), we identified beneficiaries with and without SMI in 2001. We used logistic regression to predict new HIV diagnoses in 2002-2004 among Medicaid beneficiaries who had not been diagnosed with HIV at baseline (2001).

Results: Among Medicaid beneficiaries without SMI in 2001, 0.60% were later diagnosed with HIV in 2002-2004. Among Medicaid beneficiaries with SMI in 2001, 0.70% were later diagnosed with HIV in 2002-2004, a proportion that is slightly higher but that does not support the hypothesis of substantially elevated risk among those with SMI, even on an unadjusted basis. After controlling for substance abuse diagnoses and other selected characteristics, the odds of a new HIV/AIDS diagnosis in 2002-2004 did not differ significantly between beneficiaries with versus without SMI in 2001. Relative to beneficiaries without a substance abuse disorder or SMI in 2001, individuals: (1) with substance abuse but without SMI in 2001 were 3.1 times as likely (OR=3.13, CI=3.03-3.25, p<.001) to receive a new HIV diagnosis in 2002-2004; and (2) with both substance abuse and SMI in 2001 were 2.1 times as likely (OR=2.09, CI=1.96-2.22, p<.001). However, people with SMI and without substance abuse in 2001 were 23% less likely (OR=.77, CI=.73-.80, p<.001) to receive a new HIV diagnosis in 2002-2004 than people without SMI or substance abuse in 2001.

Conclusions and Implications: Findings underscore the link between substance abuse and probability of HIV diagnosis. However in contrast to studies which report associations between SMI and HIV risk, we did not find SMI diagnosis per se (in the absence of substance abuse diagnosis) to be associated with increased risk of HIV/AIDS. Although more research is needed, our study results merit consideration in developing, planning, and implementing HIV prevention services to Medicaid beneficiaries with SMI. While it may be reasonable under some circumstances to design interventions for all persons with SMI, and deliver them to all, greater targeting to those with substance abuse disorder merits serious consideration.