Utilization Patterns of Stimulants in ADHD in the Medicaid Population: A Retrospective Analysis of Data from the Texas Medicaid Program

Abstract Background Some previously published research on treatment utilization patterns in patients with attention deficit/hyperactivity disorder (ADHD) has been focused on data from commercial health plans, whereas research in the Medicaid population is lacking. Thus, little is known about these utilization patterns in Medicaid populations, which typically have demographic and clinical characteristics that differ from those of employer-based groups. Objectives The objectives of the present retrospective data analysis were to evaluate the associations of medication groups (categorized by stimulant type [methylphenidate or amphetamine] and duration of action [short-acting (SA) or long-acting (LA)]) with measures of stimulant utilization patterns (adherence, persistence, and switching) in children, adolescents, and adults with ADHD enrolled in the fee-for-service delivery model within the Texas Medicaid Program. Methods Texas Medicaid fee-for-service claims data were analyzed retrospectively. Data from enrollees with ADHD (6–63 years) were included if patients were newly initiated on medication from January 2006 to September 2007, had ≥1 medical claim with a code for ADHD, and had continuous Medicaid eligibility 6 months before and after treatment initiation. Adherence, persistence, and switching were compared by initial stimulant medication group (SA methylphenidate [SA-M], LA-M, SA amphetamine [SA-A], and LA-A). Rates were compared overall and by age group (children, adolescents, and adults). Multivariate models were used to control for demographic, clinical, and utilization covariates. Results Of 15,055 eligible patients, mostly children, 71% were initiated on methylphenidate; 90% received LA formulations (LA-M, 65%; LA-A, 25%). Most children (66%) and adolescents (65%) were initiated on LA-M, followed by LA-A (23% and 29%, respectively). Among adults, 38% each were initiated on LA-M and LA-A. Overall adherence (measured using the days in possession ratio [DPR]) and persistence were significantly greater with the LA versus the SA formulations (mean DPR, 0.497–0.504 vs 0.407–0.418, respectively; mean persistence, 81 vs 66–67 days; both, P < 0.001), and the rates of switching were lower with the LA versus the SA formulations (12.3%–14% vs 27%–28%; P < 0.001). On multivariate analyses, the likelihoods of adherence and persistence were significantly greater with the LA formulations, and the likelihood of being switched was significantly greater with the SA f