Interventions to improve adherence to antipsychotic medication are needed. The aims of the current study were to identify the most common barriers to medication adherence in a cohort of patients receiving outpatient and inpatient treatment for an acute exacerbation of schizophrenia, compare clinical and demographic characteristics of patients with lower versus higher numbers of barriers, and characterize patients most likely to be nonadherent to antipsychotic medication. The present study analyzed data collected during the Schizophrenia Guidelines Project (SGP), a multisite study of strategies to implement practice guidelines that was funded by the U.S. Department of Veterans Affairs and conducted from March 1999 to October 2000. Nurse coordinators had conducted clinical assessments and performed an intervention designed to improve medication adherence by addressing barriers to adherence. Data on patient symptoms, functioning, and side effects had been obtained using the Positive and Negative Syndrome Scale (PANSS), the Schizophrenia Outcomes Module, the Medical Outcomes Study 36-item Short-Form Health Survey, and the Barnes Akathisia Scale (BAS). Administrative data were used to identify patients with an ICD-9 code for schizophrenia. A total of 153 patients who met this criterion and participated in the intervention arm of the SGP had complete data available for analysis in the current study. The most common patient-reported barriers were related to the stigma of taking medications, adverse drug reactions, forgetfulness, and lack of social support. Bivariate analysis showed that patients with high barriers were significantly more likely to be nonadherent (p < or =.02), to have problems with alcohol or drug use (p =.02), to have higher PANSS total scores (p =.03), and to have higher mean BAS scores (p =.02). Logistic regression showed that lower patient education level (odds ratio [OR] = 3.95, p =.02), substance abuse (OR = 3.24, p =.01), high PANSS total scores (OR = 1.02, p =.05), and high barriers (OR = 2.3, p =.05) were significantly associated with the probability of nonadherence. It may be possible to identify patients most likely to benefit from adherence intervention. The data presented here will help to inform future research of clinical interventions to improve medication adherence in schizophrenia and help to stimulate further work in this area.