OBJECTIVES: This study aimed to examine type 2 diabetic patients’ resource use and costs associated with changes in dosing regimen (monotherapy (MONO), dual therapy (DUAL), and fixed-dose combination therapy (FDCT)). METHODS: This study utilized a retrospective cohort of Medicaid enrollees from eight states with type 2 diabetes newly starting antidiabetic therapy. Patients were followed for 12 months after initiation of index dosing regimen (MONO, DUAL, or FDCT) and 12 months after early change in dosing regimen (DUAL or FDCT). Demographic characteristics, pharmacy records, economic and utilization-related variables were extracted for eligible patients. Multiple log-linear regression analysis was employed to model health care costs while multiple logistic regression analysis was utilized to study likelihoods of change in therapy, hospitalizations, and ER visits. Zero-inflated negative binomial and negative binomial regressions were employed to model counts of hospitalization and ER visits respectively.                       RESULTS: There were 10,749 patients in the study cohort including 8,528 (79.4%) MONO, 1558 (14.5%) DUAL, and 663 (6.2%) FDCT patients. No significant different in distribution of age was found across the three groups. Among the 1318 patients who changed their therapy, 809 (61.4%) had an augmentation to DUAL and 509 (38.6%) switched to FDCT. Patients with higher MPR had lower odd of switching to FDCT (p <0.001). FDCT patients had 9.2% lower costs than the augmentation to DUAL patients (p <0.05). Patients who switched to FDCT had 16.5% lower ER visits than those patients who shifted to DUAL (p <0.01). Age, race, gender, comorbidities, and diabetes-related complications were also significant predictors of the outcomes. Medication adherence was mainly associated with hospitalization and ER visit CONCLUSIONS: By considering clinical profile and therapy-related issues of the patients, disease management programs can target patients at-risk for poor outcomes and help them gradually reduce the gap between actual adherence and optimal outcomes.