Table 2. Results From the Multilevel Negative Binomial Regression^a.

Variable and model	Median log rate	% in ROPE	IRR (95% HDI)	Posterior probability of IRR >specified threshold, %
				1.0	1.1	1.2	1.3	1.4
Overall
Methylphenidate: baseline vs follow-up	0.34	0	1.41 (1.09-1.88)	99	96	87	72	52
Controls, baseline vs follow-up	0.17	13.76	1.18 (1.02-1.37)	98	82	40	10	1
Methylphenidate vs controls	0.18	28.65	1.19 (0.88-1.64)	87	70	49	30	16
No previous CVD
Methylphenidate: baseline vs follow-up	0.55	0	1.74 (1.24-2.42)	100	99	98	95	88
Controls: baseline vs follow-up	0.35	0	1.42 (1.15-1.72)	100	100	95	81	55
Methylphenidate vs controls	0.20	25.92	1.23 (0.84-1.87)	84	70	54	39	26
Previous CVD
Methylphenidate: baseline vs follow-up	0.01	39	1.00 (0.65-1.54)	48	31	19	10	5
Controls: baseline vs follow-up	–0.10	50.17	0.90 (0.75-1.11)	17	3	0	0	0
Methylphenidate vs controls	0.09	33.77	1.09 (0.68-1.74)	65	49	35	24	15
Difference
Previous CVD vs no history	0.11	24.90	1.11 (0.58-2.13)	62	50	36	23	11

Abbreviations: CVD, cardiovascular disease; HDI, highest density interval; IRR, incidence rate ratio; ROPE, region of practical equivalence.

^a Baseline was used as the reference category.