Volume 11, Number 2—February 2005
Dispatch
Comparing Aberration Detection Methods with Simulated Data
, Timothy Browne*, G. Matthew Seeman*, and Aaron T. Fleischauer*
Table A2
Sensitivity (sen) and time to detection (ttd) by method and outbreak type
| Type of outbreak | Seasonally adjusted CUSUM* |
Historical Limits† |
C1–MILD‡ |
C2–MEDIUM§ |
C3–ULTRA¶ |
|||||
|---|---|---|---|---|---|---|---|---|---|---|
| sen (%) | ttd** | sen (%) | ttd** | sen (%) | ttd** | sen (%) | ttd** | sen (%) | ttd** | |
| Spike†† | 65.55 | 0.0 | 10.0 | 0.0 | 60.2 | 0.0 | 59.9 | 0.0 | 63.3 | 0.0 |
| Spike‡‡ | 48.3 | 0.0 | 7.6 | 0.0 | 43.3 | 0.0 | 43.1 | 0.0 | 47.2 | 0.0 |
| Inverted log normal§§ | 92.2 | 1.3 | 52.6 | 2.4 | 67.2 | 1.2 | 82.6 | 1.4 | 92.8 | 1.3 |
| Inverted log normal¶¶ | 75.9 | 1.4 | 30.4 | 2.3 | 54.4 | 1.3 | 66.0 | 1.5 | 80.8 | 1.5 |
| Inverted log normal## | 98.0 | 2.4 | 77.2 | 4.7 | 66.4 | 2.5 | 83.3 | 2.6 | 95.1 | 2.4 |
| Inverted log normal*** | 88.8 | 2.8 | 50.7 | 4.8 | 59.4 | 3.0 | 71.8 | 3.1 | 88.7 | 3.0 |
| Log normal††† | 92.6 | 0.6 | 52.5 | 1.7 | 70.6 | 0.4 | 82.6 | 0.6 | 92.5 | 0.6 |
| Log normal‡‡‡ | 77.4 | 0.8 | 31.4 | 1.7 | 54.8 | 0.6 | 65.1 | 0.7 | 80.1 | 0.8 |
| Log normal§§§ | 97.2 | 1.0 | 75.8 | 2.6 | 67.9 | 0.8 | 85.4 | 1.0 | 94.3 | 1.0 |
| Log normal¶¶¶ | 88.2 | 1.6 | 49.9 | 2.9 | 56.9 | 1.2 | 71.8 | 1.4 | 87.2 | 1.4 |
*The seasonally adjusted CUSUM method sums the positive differences of the current value from the mean for a period similar to the current value over 5 years.
†The historical limits method compares the current sum of 4 periods to the mean of the sum of 15 sets of 4 periods surrounding the current point of interest over 5 years.
‡The C1–MILD method is based on CUSUM, but the calculations reduce to the current value being greater than the mean plus 3 standard deviations (SD), with the mean and SD based on the past 7 days.
§The C2–MEDIUM method is based on CUSUM, but the calculations reduce to the current value being greater than the mean plus 3 SD, with the mean and SD based on the past 7 days shifted by 2 days.
¶The C3–ULTRA method is based on CUSUM, summing the positive difference of the current value from the mean for 3 days, with the mean and SD based on the past 7 days shifted by 2 days.
**95% confidence values were <0.03. Time to detection must be interpreted with sen because ttd does not include missed outbreaks.
††This type of outbreak represents an increase of 2 SD for 1 day.
‡‡This type of outbreak represents an increase of 3 SD for 1 day.
§§This type of outbreak represents an increase of 2 SD with a fast rise in cases with incubation <7days.
¶¶This type of outbreak represents an increase of 3 SD with a fast rise in cases with incubation <7days.
##This type of outbreak represents an increase of 2 SD with a fast rise in cases with incubation >7days.
***This type of outbreak represents an increase of 3 SD with a fast rise in cases with incubation >7days.
†††This type of outbreak represents an increase of 2 SD with a slow rise in cases with incubation <7days.
‡‡‡This type of outbreak represents an increase of 3 SD with a slow rise in cases with incubation <7days.
§§§This type of outbreak represents an increase of 2 SD with a slow rise in cases with incubation >7days.
¶¶¶This type of outbreak represents an increase of 3 SD with a slow rise in cases with incubation >7days.