Cost Effectiveness of Guanfacine Extended-Release Versus Atomoxetine for the Treatment of Attention-Deficit/Hyperactivity Disorder: Application of a Matching-Adjusted Indirect Comparison

Abstract

Published online: 29 March 2013© Springer International Publishing Switzerland 2013In the November 1, 2012 issue of the Applied Health Economics and Health Policy journal, Erder et al. [1] discussed the cost effectiveness of guanfacine extended- release (GXR) versus atomoxetine (ATX) for the treatment of attention-deficit/hyperactivity disorder using a match- ing-adjusted indirect comparison. We would like to alert you that, after careful review of this article, we noticed a couple of data inaccuracies in the reporting of publicly available ATX data that met the inclusion criteria for trial selection for the indirect comparison.The data that appear to have been reported inaccurately in the Erder et al. [1] article are listed under the ATX column in Table III-'Baseline characteristics before and after matching' (p. 390). Specifically, the mean baseline values for ADHD-RS inattention subscale (16.9) and ADHD-RS hyperactivity/impulsivity subscale (22.2) appear to have been erroneously transposed from the cor- rect original values in the publicly available ATX data. The ATX data were originally presented under the 1.2 mg/kg/ day column in Table 2-'Baseline Symptom Severity' in the article by Michelson et al. [2]. Specifically, the mean baseline value for ADHD-RS inattentive subscale was listed as 22.2 while the mean baseline value for the ADHD- RS hyperactive/impulsive subscale was listed as 16.9. Thus, these correct values appear to have been switched in the article by Erder et al. [1].Importantly, these data inaccuracies may have led to erroneous post-match GXR subpopulations which likely produced results and conclusions that, at a minimum, would be unreliable. Accordingly, we respectfully would like to ask the journal and authors to comment on these inaccuracies and their effects on the research results and conclusions. Should these inaccuracies be confirmed by the journal or authors, we would kindly ask that a revised analysis be conducted and/or that a correction/erratum be published in accordance with ICMJE requirements.