Trigonometric Regression for Analysis of Public Health Surveillance Data
Table 2
Simulation results for E. coli O157:H7 infection outbreaks assuming Poisson model.
E. coli-like outbreak
Probability of detection1
Conditional expected delay2
False-positive rate3
Size ()
Duration (2)
Signature4
Time
Region
State
Region
State
Region
State
1
2
(1, 1): 2
Trough
1.000
0.373
1.000
1.433
0.797
0.531
1
3
(1, 1, 1): 3
Trough
1.000
0.516
1.000
1.878
0.832
0.482
1
4
(1, 1, 1, 1): 4
Trough
1.000
0.637
1.000
2.282
0.811
0.459
2
2
(2, 2): 4
Trough
1.000
1.000
1.000
1.000
0.825
0.521
2
3
(1, 2, 1): 4
Trough
1.000
1.000
1.000
1.798
0.799
0.494
2
4
(1, 2, 2, 1): 6
Trough
1.000
1.000
1.000
1.796
0.803
0.462
2
2
(2, 2): 4
Mid
0.545
0.142
1.451
1.586
0.835
0.553
2
3
(1, 2, 1): 4
Mid
0.421
0.133
2.007
2.107
0.791
0.546
2
4
(1, 2, 2, 1): 6
Mid
0.653
0.207
2.417
2.476
0.762
0.537
4
2
(4, 4): 8
Mid
1.000
0.708
1.000
1.393
0.826
0.558
4
3
(2, 4, 2): 8
Mid
1.000
0.573
1.701
1.979
0.797
0.534
4
4
(2, 4, 4, 2): 12
Mid
1.000
0.789
1.694
2.281
0.784
0.536
2
2
(2, 2): 4
Peak
0.457
0.136
1.432
1.453
0.814
0.548
2
3
(1, 2, 1): 4
Peak
0.371
0.119
1.954
2.004
0.783
0.540
2
4
(1, 2, 2, 1): 6
Peak
0.543
0.183
2.358
2.414
0.740
0.535
4
2
(4, 4): 8
Peak
1.000
0.544
1.000
1.410
0.791
0.529
4
3
(2, 4, 2): 8
Peak
1.000
0.405
1.739
1.931
0.773
0.532
4
4
(2, 4, 4, 2): 12
Peak
1.000
0.596
1.747
2.295
0.757
0.536
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.