Detecting cancer clusters in a regional population with local cluster tests and Bayesian smoothing methods: a simulation study

Table 1 Summary results of 1000 realizations from an inhomogeneous Poisson process

Lung cancer cases	Cluster	Expected¹(mean)	Observed (mean) RR = 2.0²	SIR (mean) RR = 2.0	SIR 95% Poisson CI³RR = 2.0	Observed (mean) RR = 4.0²	SIR (mean) RR = 4.0	SIR 95% Poisson CI³RR = 4.0
Census tracts level
Males	Urban cluster	18	35	1.92	1.38-2.67	69	3.78	3.03-4.85
Females	Urban cluster	6	13	2.13	1.26-3.72	25	4.24	2.27-7.04
Males	Rural cluster	3	6	2.08	0.90-4.45	12	4.11	2.82-6.17
Females	Rural cluster	1	2	2.2	0.50-7.99	4	4.32	1.5-10.66
Males	No cluster, urban	1000	996	1	0.94-1.06	998	1	0.94-1.10
Males	No cluster, rural	1000	1004	1	0.94-1.06	1010	1	0.95-1.07
Females	No cluster, urban	358	369	1	0.94-1.06	358	1	0.9-1.1
Females	No cluster, rural	358	371	1.03	0.94-1.15	363	1	0.91-1.12
Community level
Males	Urban cluster	49	66	1.35	1.10-1.71	99	2.03	1.66-2.46
Females	Urban cluster	18	25	1.36	0.94-2.10	37	2.07	1.49-2.84
Males	Rural cluster	30	33	1.11	0.78-1.55	39	1.32	0.95-1.78
Females	Rural cluster	11	12	1.06	0.62-1.92	13	1.25	0.69-2.04
Males	No cluster, urban	942	934	1	0.93-1.06	966	1.03	0.99-1.13
Males	No cluster, rural	942	967	1.03	0.96-1.09	983	1.04	0.98-1.11
Females	No cluster, urban	336	333	1	0.9-1.1	346	1.03	0.93-1.14
	No cluster, rural		346	1.03	0.93-1.1	354	1.05	0.95-1.17

ISSN: 1476-072X