Getting an Edge on Disease Outbreaks 05/05/2009

Kumanan Wilson and John Brownstein on using the internet to track outbreaks
http://www.globalhealthmagazine.com/guest_blog/getting_an_edge_on_disease_outbreaks

The SARS outbreak of 2003 resulted in hundreds of deaths and billions of dollars of economic losses. While officially reported in February 2003, there was Internet evidence of the outbreak as early as November 2002. International awareness of the outbreak at that time could have led to measures that could have limited its global impact.

Every year influenza outbreaks affect tens of thousands of Americans. However, recent analyses have shown that Internet searching for terms related to influenza rise prior to the official recognition of these outbreaks. Preliminary data suggests that the current HINI influenza outbreak may have also been forewarned by increase in influenza specific search terms in Mexico. Acting on these signals could potentially allow for measures to be taken to reduce the spread of the infections and, in particular, protect the most vulnerable to their effect.

Last year a food borne Listeria outbreak in Canada killed 22 individuals. An official announcement of the outbreak was made in August 2008. Our analysis of the search term Listeriosis found that it spiked one month prior to the official announcement, and the rise in searching correlated with the retrospective epidemic curve. If measures had been taken at an earlier time perhaps fewer Canadians may have been exposed to the deadly bacteria.

These examples illustrate the potentially powerful tool of using the Internet to aid in outbreak detection. Internet disease surveillance strategies generally fall into 2 categories. The first strategy involves examining the Internet for reports or professional discussion potentially related to outbreaks. The WHO's Global Public Health Intelligence Network searches news feed aggregators for reports that may provide evidence of a disease outbreak. The International Society for Infectious Disease Program for Monitoring Emerging Diseases (Promed) disseminates public health reports through its membership. HealthMap scans 20,000 news sources every hour to aggregate and disseminate information on emerging diseases. The second strategy, one which is more in its infancy, relies on surveillance of changes in Internet search term patterns. These may be potential harbingers, perhaps our earliest warnings, of an emerging outbreak.

Such Internet-based disease surveillance relies upon the idea that there is a collective intelligence consisting of the combined postings and search activities of millions of individual Internet users. However, while promising there are limitations to these approaches. In particular, there is the problem of how many false warnings may be detected for every true warning. These strategies, therefore, need to be complemented by public health vetting of the reports. An interesting question that arises is how many false messages is it worth investigating to find one true message. Would 99 false investigations be worth early identification and prevention of one SARS outbreak? As these tools are further calibrated we will develop a better understanding of the answers to this question as well as to determining the overall effectiveness of these approaches.

Future activities to improve these instruments will likely focus on engaging health-care and infectious disease professional networks and the public health community at-large in these surveillance efforts. Many of these individuals work in the field and have the ability to provide proper verification and context to the alerts. We also expect these surveillance systems to move to a decentralized, open and freely-accessible approach where anyone can contribute, discuss and verify incoming data feed , akin to a Wikipedia view of ongoing disease outbreaks.

A by-product of the ability to use this information is that it should also promote greater transparency in reporting of public health events within nations' borders. According to the revised International Health Regulations, the WHO has the authority to rely upon these forms of unofficial reports to conduct disease surveillance. Efforts by a state or region to hide these outbreaks, because of economic concerns, is challenging when news reports may be available on the Internet and, in particular, when individual search activity may strongly suggest that there is concern at the ground level about a spreading disease.

Public health practice is increasingly governed by the precautionary principle, which argues that complete information of risk should not preclude activities to mitigate the harm from that risk. At least based on this principle, Internet surveillance should be considered as an adjunct to traditional public health surveillance activities to protect populations from rapidly spreading harms.

Kumanan Wilson is the Canada research chair in public health policy at the Ottawa Hospital Research Institute, University of Ottawa

John Brownstein is an assistant professor Children's Hospital Boston, Harvard Medical School and co-founder of HealthMap.