2006). The moving average calculates the mean of the previous values and
compares it to the current value. Data closer in time to the current day can
be given more weight than data farther in the past, as is the case with an
exponentially weighted moving average (EWMA). Moving average charts
are more sensitive than thresholds at detecting small shifts in the process
average (SAS 2005). However, predictions become accustomed to increasing
counts during the early stages of an outbreak, and this increases the time to
alert (Wong and Moore 2006). This problem is the prime motivation for using
the CUSUM approach.
The CUSUM method involves calculating the cumulative sum over time
of the differences between observed counts and a reference value that rep-
resents the in-control mean (O'Brien and Christie 1997). As the algorithm
is based on the accumulation of differences, the CUSUM detects an aberra-
tion very quickly and is able to detect small shifts from the mean (O'Brien
and Christie 1997; Wong and Moore 2006). However, the algorithm will
signal a false alert if deviations in the underlying process that are not asso-
ciated with an outbreak occur, such as a steady rise in the mean (Moore
et al. 2002).
Spatial scan statistics are used to detect geographical disease clusters
of high or low incidence, and evaluate their statistical significance using
likelihood-ratio tests (Kulldorff 1997). This algorithm identifies a significant
excess of cases within a moving cylindrical window over a defined spatial
region and is able to adjust for multiple hypothesis testing (Kulldorff 1999).
It differs from the other algorithms described, as it is able to detect spatial as
well as temporal clusters.
The comparison of alerts generated by aberration detection algorithms
has been used to measure the time difference between two data sources.
An English study applied the peak comparison method and thresholds for
comparing OTC sales of cough/cold remedies to emergency department
admission data (Davies and Finch 2003). Analyses demonstrated that peak
sales both preceded and lagged the peak in admissions over the years inves-
tigated. However, increases above a defined threshold of OTC sales occurred
14 days before the peak in emergency department admissions in all 3 years
of the study (Davies and Finch 2003).
Quenel and colleagues (1994) applied thresholds to determine the date
of alert for various health service-based indicators from hospitals and
absenteeism records. The threshold above which an alert was declared was
defined as the upper limit of the 95% confidence interval of the weekly aver-
age calculated from nonepidemic weeks. An epidemic week occurred when
1% of specimens were positive for influenza A. The health service-based
indicators increased before virological confirmation over the 5-year period.
The emergency visit indicator was the earliest (average 11.2 days, range 7-28
days before virological confirmation) followed by sick-leave reports col-
lected by general practitioners (average 8.4 days, range 7-21 days) (Quenel
et al. 1994).