Short Communication MedISys: An early-warning system for the detection of (re-)emerging food- and feed-borne hazards Agnès Rortais a, * , Jenya Belyaeva b , Monica Gemo b , Erik van der Goot b , Jens P. Linge b a Emerging Risks Unit, European Food Safety Authority (EFSA), Largo N. Palli 5/A, 43121 Parma, Italy b Joint Research Centre (JRC), European Commission, Ispra, Italy article info Article history: Received 17 November 2009 Accepted 8 April 2010 Keywords: Emerging risks Media monitoring MedISys ProMED-mail RASFF Food Feed abstract We evaluated the Medical Information System (MedISys) as an early-warning system for the detection of food- and feed-borne hazards. Nine hazards were selected in the period from January 2007 to March 2009 from the Rapid Alert System for Food and Feed (RASFF) and traced back on MedISys and ProMED-mail. In addition, from January to March 2009, food- and feed-borne (re-)emerging hazards were monitored on MedISys and traced back on ProMED-mail and RASFF. MedISys has demonstrated to be an effective early-warning system for food- and feed-borne hazards. However, further customization is required to improve its sensitivity, in particular by increasing the number of multi-lingual categories related to food and feed items. MedISys tended to detect food- and feed-borne hazards earlier and more frequently than ProMED-mail, but the information from both systems was often complementary. Ó 2010 Elsevier Ltd. All rights reserved. 1. Introduction Recent food safety incidents such as the fraudulent use of mel- amine in feed and food (Ingelfinger, 2008) have demonstrated the increasingly globalized nature of the food chain (Kuehn, 2009). As part of a comprehensive programme to improve EU food safety, the European Food Safety Authority (EFSA) was established as an inde- pendent source of scientific advice and communication on risks associated with the food chain (European Commission, 2002). In 2008, an EFSA Emerging Risks Unit (EMRISK) was set up and initi- ated the development of automated monitoring tools to collect such data. Among these tools, the EFSA Scientific Cooperation Working Group requested the evaluation (European Food Safety Authority, 2009) of the Medical Information System (MedISys, http://medusa.jrc.it), an automated system that monitors the med- ia. Since its inception in 2004, MedISys has extended its media cov- erage, improved information aggregation across documents and languages, and has gained a broad user base in the process. MedI- Sys enables users to moderate incoming news items and to dissem- inate the collected information (Linge et al., 2009; Steinberger, Fuart, Pouliquen, & van der Goot, 2008). The efficiency of media monitoring systems in detecting early global health events and in delivering information at reduced cost and with increased transparency has been largely documented (Wilson & Brownstein, 2009). Yet, to date, there is no data to show to which degree these systems can detect emerging food- and feed- borne hazards. Therefore, the objective of this study was to present data to evaluate the performance of MedISys as an early-warning system for the detection of food- and feed-borne hazards using two complementary approaches. First, a retrospective approach based on several case studies from the Rapid Alert System for Food and Feed (RASFF) was used in the period from January 2007 till March 2009, applying mainly English keyword searches. These hazards were traced back on MedISys and compared to ProMED- mail, which disseminates information on outbreaks of infectious diseases and acute exposures to toxins that affect human health (Madoff & Woodall, 2005). Second, food- and feed-borne hazards detected on MedISys by monitoring all available languages with multi-lingual category definitions were traced back on RASFF and ProMED-mail in the period from January to March 2009. 2. Materials and methods 2.1. Materials The current figures (i.e. as of August 2009) show that MedISys 1 processes approximately 90,000 news articles from more than 2200 news sites in 50 languages per day. All news items are categorized into hundreds of categories which include geolocation, entities 0963-9969/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodres.2010.04.009 * Corresponding author. Fax: +39 05210360243. E-mail addresses: agnes.RORTAIS@efsa.europa.eu (A. Rortais), jenya.belyaeva @ext.jrc.ec.europa.eu (J. Belyaeva), monica.gemo@jrc.ec.europa.eu (M. Gemo), erik.van-der-goot@jrc.ec.europa.eu (E. van der Goot), jens.linge@jrc.ec.europa.eu (J.P. Linge). 1 MedISys, Medical Information System, http://medusa.jrc.it. Food Research International 43 (2010) 1553–1556 Contents lists available at ScienceDirect Food Research International journal homepage: www.elsevier.com/locate/foodres