A comparison of estimators from self-controlled case series, case-crossover design, and sequence symmetry analysis for pharmacoepidemiological studies

Table 3 Results of Simulations for Scenario 1

	Setting	Results of simulations
Methods (Effect measures)	β _X	Mean of estimates (Ratio scale)	Bias (Log scale)	Empirical standard error	Mean squared error	Mean standard error	Coverage (%)
SCCS (IRR)	log(1.0)	0.99	-0.0125	0.1354	0.0185	0.1335	95.1
CCO (OR)		1.00	-0.0005	0.2084	0.0434	0.2116	96.1
SSA (ASR)		0.99	-0.0005	0.2373	0.0563	0.2454	96.3
SCCS (IRR)	log(3.0)	2.99	-0.0030	0.0791	0.0063	0.0777	94.7
CCO (OR)		3.04	0.0121	0.18	0.0309	0.1732	94.8
SSA (ASR)		3.04	0.0119	0.2016	0.0408	0.2003	95.2
SCCS (IRR)	log(10.0)	9.98	-0.0019	0.0438	0.0019	0.0446	95.5
CCO (OR)		10.30	0.0295	0.1663	0.0285	0.1604	94.4
SSA (ASR)		9.97	-0.0026	0.1834	0.0336	0.1817	94.7

Abbreviations ASR adjusted sequence ratio, CCO case-crossover, IRR incident rate ratio, OR odds ratio, SCCS self-controlled case series, SSA sequence symmetry analysis

ISSN: 1471-2288