Stimulant Adherence and Academic Performance in Urban Youth With Attention-Deficit/Hyperactivity Disorder

Objective: This analysis assessed whether stimulant adherence was associated with improvement in academic grade point average (GPA) among children diagnosed with and treated for attention-deficit/hyperactivity disorder (ADHD). Method: Medicaid claims were merged with academic records from Philadelphia public schools of Medicaid-eligible children in first through eighth grades who were diagnosed with ADHD and who had filled one or more stimulant prescription. Students diagnosed with mental retardation, autism, or speech, hearing, visual, or language impairments were excluded. Marking periods were scored for GPA (range: 0-4.0) based on English, mathematics, social studies, and science grades and for stimulant adherence (medication possession ratio >= 0.70). Random and fixed-effects models estimated the effects of stimulant adherence on GPA, between all adherent and nonadherent marking periods in aggregate and within individual student's marking periods, respectively. Results: A total of 3,543 students contributed 29,992 marking periods, of which 18.6% were adherent. Mean GPA was significantly higher during stimulant-adherent (2.18) than stimulant-nonadherent (1.99) marking periods in aggregate (p < .0001). The regression coefficient representing within-student association between stimulant adherence and GPA over time was 0.108 (p < .0001), indicating that adherence was associated with a 0.108 increase in GPA. In stratified analyses, analogous coefficients were 0.106 for boys, 0.111 for girls, 0.078 for elementary students, and 0.118 for middle school students (all p < .0001). The association was stronger among students with (0.139) than without (0.088) comorbid disruptive behavior disorders (both p < .0001). Conclusions: Stimulant adherence, although found to be low among urban elementary and middle school students diagnosed with ADHD, was associated with a marginal improvement in GPA. J. Am. Acad. Child Adolesc. Psychiatry, 2011;50(5):480-489.