Skip directly to search Skip directly to A to Z list Skip directly to page options Skip directly to site content

Volume 14, Number 10—October 2008

Dispatch

Automatic Outbreak Detection Algorithm versus Electronic Reporting System

Masja Straetemans1Comments to Author , Doris Altmann, Tim Eckmanns, and Gérard Krause
Author affiliations: Robert Koch Institute, Berlin, Germany;

Main Article

Table

Outbreaks January 31, 2005–January 28, 2007, reported and identified by detection algorithm*

Outbreak characteristic Campylobacter spp., no. (%) Norovirus, no. (%)
Total cases 114,176 144,568
Cases as part of a reported outbreak 3,767 (3.3) 103,177 (71.4)
Reported outbreaks with <4 cases 1,453 5,074
Reported outbreaks with >4 cases 118 4,309†
Signal outbreaks generated by detection algorithm
781
2,801
Reported outbreaks with >4 cases identified by detection algorithm signals 52 (100) 2,538 (100)
Reported outbreaks identified by 1 signal 49 (94.0) 1,811 (71.4)
Reported outbreaks identified by >1 signal 3 (6.0) 727 (28.6)
Reported outbreaks identified by 2 signals‡ 3 (6.0) 473 (18.6)
Reported outbreaks identified by >2 signals‡
0
254 (10.0)
Signal outbreaks corresponding to reported outbreak with >4 cases 50 (100) 2,115 (100)
Signal outbreaks corresponding to 1 reported outbreak 46 (92.0) 1,355 (64.1)
Signal outbreaks corresponding to >1 reported outbreak 4 (8.0) 760 (35.9)
Signal outbreaks corresponding to 2 reported outbreaks§ 3 (6.0) 408 (19.3)
Signal outbreaks corresponding to >2 reported outbreaks§ 1 (2.0) 352 (16.7)

*Data through June 1, 2007. Sensitivity detection algorithm 44.1% (52/118) for Campylobacter spp., 58.9% (2,538/4,309) for norovirus; no. reported outbreaks with >4 cases also identified by detection algorithm signal/total no. reported outbreaks with >4 cases. Positive predictive value of detection algorithm 6.4% (50/781) for Campylobacter spp., 75.5 (2,115/2,801) for norovirus. No. signal outbreaks identical to reported outbreak/total number of signal outbreaks.
Excluded are 17 reported norovirus outbreaks of >25 wk and an average of <2 cases/wk because these are likely the result of data entry errors in SurvNet.
‡During the duration of a reported outbreak, the detection algorithm may have triggered multiple signals during several consecutive weeks (Figure 1).
§One signal outbreak may correspond to multiple reported outbreaks if different outbreaks occur in the same municipality during the same period (Figure 2).

Main Article

1Current affiliation: KNCV Tuberculosis Foundation, The Hague, the Netherlands

TOP