Biology Reference
In-Depth Information
At the same time, we must keep in sight the limitations with using media
reports, which can sometimes be misinformed, are often prediagnostic,
and tend to be written by nonexperts—sometimes with inbuilt distribution
biases. There are also the concerns of the public in such areas as privacy. We
do, however, consider that
rumor-based
information
may be of particular value
in helping (a) extend international coverage to areas where traditional sur-
veillance system infrastructure is absent, and (b) strengthen the early detec-
tion capacity of existing systems, particularly at the local level. GHISs should
properly be considered a part of a range of solutions for enabling a timely
and appropriate public health response.
In this chapter, we briefly survey several extant GHISs being utilized by
national agencies before detailing BioCaster, a freely available nongovernmen-
tal system developed by our team at the National Institute of Informatics in
collaboration with a group of partners in Japan, Vietnam, Thailand, and the
United States (Figure 15.2; Collier et al. 2008). Although BioCaster's coverage
of health threats is global, it aims to enhance existing surveillance efforts by
prioritizing languages within the Asia-Pacific region. With the incorporation
of a multilingual ontology of 11 languages, BioCaster contains a single high-
throughput system with advanced text mining and inference capabilities for
t he detec t ion of l i ng u i st ic sig na l s for severe event s of i nter nat iona l sig n i ica nce,
Figure 15.2
Screenshot of BioCaster's Global Health Monitor. News reports that are positively identified as
disease outbreaks are processed to detect country, province, and disease terms before being
plotted on a biogeographic Google map. The map allows users to filter the previous 30 days of
news reports by time, document type, syndrome, or disease, as well as linking to the BioCaster
ontology and external reference sources such as GoPubMed or HighWire.
