Effect of pharmacological treatment of attention-deficit hyperactivity disorder on academic performance: registry-based population study
Ingvild Lyhmann, Tarjei Widding-Havneraas, Simen Markussen, Henrik Daae Zachrisson, Felix Elwert, Ashmita Chaulagain, Ingvar Bjelland, Anne Halmøy, Arnstein MykletunBackground
Attention-deficit hyperactivity disorder (ADHD) is associated with high risks of adverse academic outcomes. A major motivation for initiating pharmacological treatment of ADHD is to mitigate these risks; however, the evidence regarding its long-term effects is inconclusive.
Aims
Our objective was to investigate the long-term causal effects of pharmacological treatment of ADHD on school grades in Norwegian children and adolescents.
Method
We analysed individual-level registry data of all Norwegians aged 5–15 years in 2009–2011 who were diagnosed with ADHD during this period ( n = 6255). Treatment was measured as cumulative ADHD medication prescriptions from time of diagnosis to graduation from mandatory schooling (tenth grade). Outcomes included final grades in nine main school subjects, grade point average and non-assessment (exemption, no basis for assessment). We estimated ordinary least squares (OLS) and instrumental variable models. By exploiting naturally occurring variation in treatment practices, instrumental variable analysis can address residual confounding and estimate causal effects for patients ‘on the margin of treatment’, i.e. those prone to varying treatment practices.
Results
OLS analysis indicated that pharmacological treatment for ADHD was associated with slightly improved grade point average (0.05, 95% CI 0.02–0.09; grade point average s.d. 0.94). Instrumental variable analysis indicated no causal effect of pharmacological treatment on grade point average (−0.01, Anderson−Rubin 95% CI −1.13 to 0.15), individual school subjects or non-assessment.
Conclusions
For children and adolescents on the margin of pharmacological treatment for ADHD, we found no causal effect of pharmacological treatment on any of our measures of academic performance. Although further research is needed, this finding should be considered in the context of clinical decision-making and treatment planning.