Table 9.

Odds ratios and relative risks for discontinuation due to AEs (drug versus placebo)

Drug	Odds ratio	95% credible interval for odds ratio	Probability of the treatment being least likely to be discontinued among all	Probability of GXR being less likely to be discontinued compared with each treatment
GXR	4.50	(2.14, 10.31)	<1%	–
LDX	2.95	(1.21, 7.59)	1.67%	22.55%
ATX	2.35	(1.27, 4.37)	<1%	7.74%
MPH-ER	1.29	(0.59, 2.77)	33.50%	1.05%
MPH-IR	1.02	(0.32, 3.18)	63.85%	1.89%

Drug	Relative risk of drug versus placebo	95% credible interval for relative risk	Response rate	Placebo riska (95% confidence interval)
GXR	4.49	(2.10, 8.81)	0.08	0.02 (0.01, 0.02)
LDX	3.11	(1.20, 6.76)	0.06
ATX	2.39	(1.26, 4.11)	0.04
MPH-ER	1.38	(0.60, 2.68)	0.03
MPH-IR	1.20	(0.32, 3.06)	0.02

Core Bayesian analysis: random-effects model, combined doses, excluding short-term studies, and adjusted for age and percent female

Both forced- and optimal-dose studies were included in the network. Among forced-dose studies, randomization arms included GXR 1, 2, 3, and 4 mg/day; LDX 30, 50, and 70 mg/day; ATX 0.5, 1.2, and 1.8 mg/kg/day; and MPH-ER 18, 20, and 36 mg/day. Unlike for the ADHD-RS-IV outcome, MPH-IR trial data with AE-related discontinuation rates were available. Hence, MPH-IR was included in the AE-related discontinuation network. All MPH-IR studies were optimal-dose studies

ADHD-RS-IV Attention-Deficit/Hyperactivity Disorder Rating Scale Version IV, AE adverse event, ATX atomoxetine, GXR guanfacine extended release, LDX lisdexamfetamine dimesylate, MPH-ER methylphenidate extended release, MPH-IR methylphenidate immediate release

^aThe placebo risk is the pooled risk of discontinuation of the placebo arms in the data. The placebo risk uncertainty is measured as the 95% confidence interval around this pooled placebo risk. The odds ratio was converted to a relative risk scale using the pooled placebo rate