Table 3.

Adjusted^* Empiric Bayes Geometric Mean (EBGM) Ratios comparing Relative Risks for Outcome Events for Current Rofecoxib Users Versus Current Non-Selective Non Steroidal Anti-Inflammatory Drug (NS-NSAID) Users, Sorted by Significance based on Overlap of EB05/95 Confidence Intervals

	Event (n) in Rofecoxib Exposed	Events (n) in NS-NSAID Exposed	EBGM Ratio of Rofecoxib to NS-NSAID**	Ratio of Rofecoxib EB05 to NS-NSAID EB95 CIs†
Rehabilitation care; fitting of prostheses; and adjustment of devices (254)	40	35	2.10	1.30
Complication of device; implant or graft (237)	30	30	2.00	1.17
Osteoarthritis (203)	45	53	1.72	1.12
Acute cerebrovascular disease (109)	21	22	1.85	1.01
Coronary atherosclerosis and other heart disease (101)	34	48	1.60	0.99
Spondylosis; intervertebral disc disorders; other back problems (205)	28	38	1.52	0.91
Nonspecific chest pain (102)	23	32	1.57	0.90
Cardiac dysrhythmias (106)	23	28	1.57	0.90
Diverticulosis and diverticulitis (146)	14	13	1.76	0.87

Numbers in parentheses () in the first column refer to the Agency for Health Research and Quality Clinical Classifications Software (CCS) event group

^*adjusted for age, gender, calendar year, and Charlson comorbidity index

^**estimates relative risk

^†ratios greater than 1.00 indicate that the confidence intervals of the two EGBM ratios did not overlap. All events with ratios > 0.85 are displayed.