Syndromic surveillance for local outbreaks of lower-respiratory infections: would it work?

Abstract

Background: Although syndromic surveillance is increasingly used to detect unusual illness, there is a debate whether it is useful for detecting local outbreaks. We evaluated whether syndromic surveillance detects local outbreaks of lower-respiratory infections (LRIs) without swamping true signals by false alarms.

Methods and findings: Using retrospective hospitalization data, we simulated prospective surveillance for LRI-elevations. Between 1999-2006, a total of 290762 LRIs were included by date of hospitalization and patients place of residence (>80% coverage, 16 million population). Two large outbreaks of Legionnaires disease in the Netherlands were used as positive controls to test whether these outbreaks could have been detected as local LRI elevations. We used a space-time permutation scan statistic to detect LRI clusters. We evaluated how many LRI-clusters were detected in 1999-2006 and assessed likely causes for the cluster-signals by looking for significantly higher proportions of specific hospital discharge diagnoses (e.g. Legionnaires disease) and overlap with regional influenza elevations. We also evaluated whether the number of space-time signals can be reduced by restricting the scan statistic in space or time. In 1999-2006 the scan-statistic detected 35 local LRI clusters, representing on average 5 clusters per year. The known Legionnaires' disease outbreaks in 1999 and 2006 were detected as LRI-clusters, since cluster-signals were generated with an increased proportion of Legionnaires disease patients (p:<0.0001). 21 other clusters coincided with local influenza and/or respiratory syncytial virus activity, and 1 cluster appeared to be a data artifact. For 11 clusters no likely cause was defined, some possibly representing as yet undetected LRI-outbreaks. With restrictions on time and spatial windows the scan statistic still detected the Legionnaires' disease outbreaks, without loss of timeliness and with less signals generated in time (up to 42% decline).

Conclusions: To our knowledge this is the first study that systematically evaluates the performance of space-time syndromic surveillance with nationwide high coverage data over a longer period. The results show that syndromic surveillance can detect local LRI-outbreaks in a timely manner, independent of laboratory-based outbreak detection. Furthermore, since comparatively few new clusters per year were observed that would prompt investigation, syndromic hospital-surveillance could be a valuable tool for detection of local LRI-outbreaks.

Figure 1. Two-step criteria to define (likely) causes for LRI hospitalization clusters detected in 1999–2006.

* As evaluated by the right-sided Fisher's exact test for 2×2 Tables (alpha≤0.01) of hospitalizations within vs hospitalizations outside of the cluster-signal. The proportion of hospitalizations with a specific characteristic (e.g. legionnaires' disease as discharge diagnoses, or age 20–49 yrs) can be significantly higher among hospitalizations within the cluster-signal than the proportion outside of the cluster-signal. ** For the ILI-cluster-signals we could only use 4 major regions as spatial resolution. Overlap in time between LRI and ILI-cluster-signals was defined as occurrence of weekly ILI-cluster-signals within 2 weeks (+/−) around LRI-cluster-signals. ***The annual influenza season was defined as all weeks with a national weekly ILI-incidence ≥3 per 10.000 pop. **** Possibly unreported/undetected local LRI-outbreaks by undetected pathogens.

Figure 2. Clusters and generated cluster-signals on a timescale, including all (likely) causes (by weekly analysis).*

*Clusters are indicated by sets of successive space-time overlapping cluster-signals placed next to each other on the same height on the y-axis. The cluster-signals caused by a data artifact in 2000 are not presented in the graphs. See Figure 1 for the criteria by which the likely causes were defined and see the Figure 2 legend for the graphic indication of likely causes. **In Figure 2a — for the analyses with non-restrictive settings on time and spatial windows — all detected clusters and signals are presented, as well as the (likely) causes according to the criteria in Figure 1. Figure 2b presents the signals and clusters that are still detected with a maximum time window of 7 weeks, and Figure 2c signals and clusters still detected with a maximum radius of 25 km. ***Signals indicated by open symbols (e.g. “○”) have a ≥1 year recurrence interval, coloured symbols (e.g. “•”) have a ≥5 yr recurrence interval. A recurrence interval reflects how often a signal of the observed significance level would be observed by chance . I.e. if the recurrence interval of a signal is say 1 year, 1 signal of the observed significance is expected in 1 year.

Figure 3. The earliest detected Legionnaires' disease outbreak related LRI-cluster-signals (1999 and 2006) as presented on a map of the Netherlands (by daily analysis).

Figure 3a and 3b show the cluster-signals that detected the 1999 and 2006 outbreak respectively. Output of the Satscan scan-statistic software is presented in the legend. On the map the borders of all postal code areas are indicated, the postal code areas of the cluster-signals are marked in dark-grey with the center postal code marked in red.