Algorithmic disease surveillance: Staying ahead of an epidemic

20170306120000, Rachel Seah
Using algorithms to monitor early signs of disease outbreak via sources on the web can allow epidemics to be contained more quickly, saving more lives.
In a connected world, infectious diseases such as Zika can spread across the globe in mere hours, causing a global emergency. However, traditional methods of disease surveillance are slow, involving lengthy processes followed by multiple layers of state bureaucracy before information is reported and action is taken.

In addition, not all countries are able or willing to give out information on new epidemic outbreaks for fear of negative economic effects.

These methods need to be replaced with faster alternatives, and the internet has emerged as a powerful tool for real-time epidemic detection. Algorithms are used to help filter and scan the vast amount of available information for signs of a new outbreak. World Health Organisation (WHO) has estimated that about 60% of its information on disease outbreaks is obtained from informal internet-based surveillance systems.

Researchers from the Centre for Global Health Policy at the University of Sussex studied three such systems, observing a movement towards algorithmic-based surveillance.

1. Program for Monitoring Emerging Diseases (ProMED-mail)

ProMED-mail first emerged in 1994 as a new platform to exchange unofficial information about new infectious disease outbreaks on a daily basis.

The pioneering system publishes information of initial outbreak reports in real-time and sends email notifications to subscribers. The subscription is free, and currently, it boasts a membership which extends to more than 150 countries.

ProMED-mail actively encourages subscribers to contribute data, and collaborates with authorities in outbreak investigations and prevention efforts. They also respond to requests for information. Each report is screened and edited by a group of 45 moderators to ensure its validity.

As it is an independent programme of the International Society for Infectious Diseases, a non-governmental organisation, it circumvents the problem of delay or suppression of disease reporting by governments.

2. The Global Public Health Information Network (GPHIN)

Where ProMED-mail relies on human moderation, GPHIN relies on algorithms for disease surveillance. GPHIN focuses on the monitoring of news reports in multiple languages around the world, potentially identifying early warning signs of a disease outbreak. Highlighting the relevant reports then allows experts to further analyse the data.

This method was possible as the media tends to pick up on suspicious events and signs of epidemic outbreaks more quickly than official government channels nowadays.

With algorithms to monitor and filter vast amounts of data available in sources of media, GPHIN was able to detect the signs of the SARS epidemic in late 2002 even before it was reported to the WHO in 2003.

3. HealthMap

HealthMap represents an even more drastic shift towards algorithmic disease surveillance. Its algorithms enable the collation of data from both official and unofficial sources, which are then broken down by precise geographic location displayed using a Google Maps plugin.

The algorithm proved its efficacy when it detected the recent Ebola outbreak in West Africa one week before it was confirmed by the Guinean Ministry of Health.

Its use also extends to the prediction of future outbreak patterns. For example, researchers behind HealthMap produced a Zika map to outline the geographic history, global range, and relative distribution of the virus.

“Our goal is to provide epidemiologists and governments the most accurate and exact information as early as possible, so governments can respond better to […] Zika,” said HealthMap Program Coordinator, Colleen Nguyen.

“For the public, this would be one central place they could go to get all the information they would need regarding Zika’s latest developments and updates,” she added. MIMS

Read more:
A strange hurdle: Finding a Zika vaccine could require a new epidemic
7 promising medical developments to advance the future of healthcare
Using big data analytics in healthcare research: Google Flu Trends and other projects

Sources:
https://theconversation.com/algorithms-could-save-your-life-when-the-next-pandemic-arrives-73210
https://medtechboston.medstro.com/blog/2016/08/16/healthmap-develops-live-zika-tracker/