OBJECTIVE: To identify patient characteristics that are associated with the incidence of thiadolidinediones (TZDs) or metformin prescnbing in Medicaid managed care plans. RESEARCH DESIGN AND METHODS: We utilized a retrospective cohort study design. Two-and-one-half years of prescription claims of Medicaid managed care organizations (MCOs) patients who were new utilizers of metformin or TZDs were analyzed using univariate, bivariate and multivariate models. Multivariate logistic regression models were built to assess the combined effect of all variables on the likelihood of incident use of TZDs or metformin. RESULTS: Claims for 3,041 patients were analyzed for the period between January 15, 2000 and June 15, 2002. African Americans and urban residents were less likely to be started on TZDs (OR = 0.678, 95% C1 = 0.830-1.206; OR = 0.579, 95% CI = 0.479-0.699, respectively). Advanced age, preexisting comorbidities and diabetes complications, and prior use of other oral diabetes drugs or insulin were predictors of increased likelihood of TZD initiation. CONCLUSIONS: Race, age, residential setting, preexisting comorbidities and diabetes complications, other oral diabetes drug use, and insulin use are statistically significant predictors of initial prescribing of TZD or metformin in a Medicaid MCO population. Findings can potentially inform the management of diabetes in managed care so as to improve outcomes.
OBJECTIVE: To identify patient characteristics that are associated with the incidence of thiadolidinediones (TZDs) or metformin prescnbing in Medicaid managed care plans. RESEARCH DESIGN AND METHODS: We utilized a retrospective cohort study design. Two-and-one-half years of prescription claims of Medicaid managed care organizations (MCOs) patients who were new utilizers of metformin or TZDs were analyzed using univariate, bivariate and multivariate models. Multivariate logistic regression models were built to assess the combined effect of all variables on the likelihood of incident use of TZDs or metformin. RESULTS: Claims for 3,041 patients were analyzed for the period between January 15, 2000 and June 15, 2002. African Americans and urban residents were less likely to be started on TZDs (OR = 0.678, 95% C1 = 0.830-1.206; OR = 0.579, 95% CI = 0.479-0.699, respectively). Advanced age, preexisting comorbidities and diabetes complications, and prior use of other oral diabetes drugs or insulin were predictors of increased likelihood of TZD initiation. CONCLUSIONS: Race, age, residential setting, preexisting comorbidities and diabetes complications, other oral diabetes drug use, and insulin use are statistically significant predictors of initial prescribing of TZD or metformin in a Medicaid MCO population. Findings can potentially inform the management of diabetes in managed care so as to improve outcomes.
Authors: Robert S Sherwin; Robert M Anderson; John B Buse; Marshall H Chin; David Eddy; Judith Fradkin; Theodore G Ganiats; Henry Ginsberg; Richard Kahn; Robin Nwankwo; Marion Rewers; Leonard Schlessinger; Michael Stern; Frank Vinicor; Bernard Zinman Journal: Diabetes Care Date: 2003-01 Impact factor: 19.112
Authors: Junling Wang; Ilene H Zuckerman; Nancy A Miller; Fadia T Shaya; Jason M Noel; C Daniel Mullins Journal: Health Serv Res Date: 2007-08 Impact factor: 3.402
Authors: Rahul A Shenolikar; Rajesh Balkrishnan; Fabian T Camacho; J Timothy Whitmire; Roger T Anderson Journal: J Natl Med Assoc Date: 2006-07 Impact factor: 1.798