Trigonometric Regression for Analysis of Public Health Surveillance Data

Table 4

Simulation results for E. coli O157:H7 infection outbreaks assuming a negative binomial model for the entire state with overdispersion parameter , , or .

E. coli-like outbreak				Overdispersion			Overdispersion			Over dispersion
Size (b)	Duration (2c)	Signature^4	Time	POD^1	CED^2	FPR^3	POD^1	CED^2	FPR^3	POD^1	CED^2	FPR^3
5	4	(1, 1): 2	Trough	1.000	1.000	0.932	1.000	1.000	1.067	1.000	1.000	1.321
5	5	(1, 1, 1): 3	Trough	1.000	1.984	0.939	1.000	1.985	1.069	1.000	1.985	1.306
5	6	(1, 1, 1, 1): 4	Trough	1.000	1.984	0.952	1.000	1.988	1.073	1.000	1.983	1.289
10	4	(2, 2): 4	Trough	1.000	1.000	0.910	1.000	1.000	1.060	1.000	1.000	1.305
10	5	(1, 2, 1): 4	Trough	1.000	1.000	0.930	1.000	1.000	1.060	1.000	1.000	1.303
10	6	(1, 2, 2, 1): 6	Trough	1.000	1.000	0.951	1.000	1.000	1.046	1.000	1.000	1.319
10	4	(2, 2): 4	Mid	0.533	1.446	0.934	0.523	1.444	1.070	0.518	1.431	1.271
10	5	(1, 2, 1): 4	Mid	0.426	2.009	0.922	0.429	2.000	1.057	0.432	2.009	1.266
10	6	(1, 2, 2, 1): 6	Mid	0.632	2.422	0.926	0.636	2.414	1.023	0.637	2.411	1.228
20	4	(4, 4): 8	Mid	1.000	1.000	0.938	1.000	1.000	1.068	1.000	1.000	1.299
20	5	(2, 4, 2): 8	Mid	1.000	1.701	0.929	1.000	1.704	1.037	1.000	1.709	1.268
20	6	(2, 4, 4, 2): 12	Mid	1.000	1.709	0.910	1.000	1.703	1.031	1.000	1.718	1.245
10	4	(2, 2): 4	Peak	0.452	1.422	0.901	0.441	1.433	1.041	0.447	1.428	1.219
10	5	(1, 2, 1): 4	Peak	0.385	1.936	0.906	0.386	1.942	1.012	0.414	1.921	1.203
10	6	(1, 2, 2, 1): 6	Peak	0.545	2.305	0.857	0.545	2.323	0.975	0.552	2.325	1.167
20	4	(4, 4): 8	Peak	1.000	1.000	0.917	1.000	1.000	1.030	1.000	1.000	1.244
20	5	(2, 4, 2): 8	Peak	1.000	1.745	0.890	1.000	1.748	1.003	1.000	1.747	1.208
20	6	(2, 4, 4, 2): 12	Peak	1.000	2.751	0.846	1.000	2.748	0.952	1.000	1.747	1.176

Probability of detection (POD) per “outbreak.”
²Conditional expected delay (CED) per “detected outbreaks.”
³False-positive rate in terms of false-positive signals per year.
⁴The signature is the number of cases in each week during the outbreak, along with the total number of cases.