Premium
Comparative effectiveness research using matching‐adjusted indirect comparison: an application to treatment with guanfacine extended release or atomoxetine in children with attention‐deficit/hyperactivity disorder and comorbid oppositional defiant disorder
Author(s) -
Signorovitch James,
Erder M. Haim,
Xie Jipan,
Sikirica Vanja,
Lu Mei,
Hodgkins Paul S.,
Wu Eric Q.
Publication year - 2012
Publication title -
pharmacoepidemiology and drug safety
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.023
H-Index - 96
eISSN - 1099-1557
pISSN - 1053-8569
DOI - 10.1002/pds.3246
Subject(s) - atomoxetine , guanfacine , medicine , placebo , attention deficit hyperactivity disorder , atomoxetine hydrochloride , confidence interval , randomized controlled trial , rating scale , meta analysis , pediatrics , psychiatry , methylphenidate , psychology , developmental psychology , clonidine , alternative medicine , pathology
Objectives To illustrate a matching‐adjusted indirect comparison by comparing the efficacy of guanfacine extended release (GXR) and atomoxetine (ATX) in reducing oppositional symptoms in children with attention‐deficit/hyperactivity disorder and comorbid oppositional defiant disorder. Methods Individual patient data were used from a GXR trial; only published summary data were used from ATX trials. In a matching‐adjusted indirect comparison, individual patients from the GXR trial were weighted such that their mean baseline characteristics matched those published for ATX trials. Placebo‐arm outcomes were then compared to further assess balance between the matched populations. Changes in the Conners' Parent Rating Scale‐Revised Short Form Oppositional Subscale from baseline to endpoint among GXR‐treated and ATX‐treated patients were then compared. Results Before matching, the GXR ( n = 143) and ATX ( n = 98) trial populations had significant differences in baseline characteristics and placebo‐arm outcomes. After matching, baseline characteristics were well balanced across trials, and placebo‐arm outcomes became nearly identical. Comparing active treatment arms across the matched populations, GXR was associated with a significantly greater reduction in mean Conners' Parent Rating Scale‐Revised Short form oppositional subscale compared with ATX {−5.0 [95% confidence interval (CI): −6.6 to −3.4] vs. −2.4 [CI: −3.7 to −1.1], p = 0.01, effect size = 0.58}. Conclusions In the absence of head‐to‐head randomized trials, matching‐adjusted indirect comparisons can provide timely and reliable comparative evidence for decision makers and can be applied even when very few trials are available for the treatments of interest. Copyright © 2012 John Wiley & Sons, Ltd.