Regular readers of this blog are already familiar with GPHIN, the Global Public Health Intelligence Network run by the Canadian government's Health agency to electronically monitor infectious diseases outbreaks and which helped detecting SARS in time for the world to respond. I wrote about it when Larry Brilliant made it the subject of his speech at the TED conference last February, and then again in an analysis in March, completed with several updates.
I guess it was apparent from these posts that I consider GPHIN a formidable instrument, and a possible model for other global advanced alert systems. So I was happy to get a chance to meet with Dr. Ron St. John, the director-general of the Center for emergency preparedness and response at the Public Health Agency of Canada in Ottawa. GPHIN - run by a team of 12 people with a current budget of about 1.5 million dollars - is part of his Center.
Dr. St. John pointed out that I got something wrong: I wrote that GPHIN is "basically a sophisticated search engine ... designed to crawl the network, filter the information for relevancy and patterns, detect early signs of a potential problem and flag them". That's how I had understood it. But GPHIN in its current incarnation is not a search engine: it's a 24/7 filtering software that scans some 20'000 sources of information in eight languages, most of them raw newswires and local newspaper reports coming in from all over the world.
As said, GPHIN gave an essential contribution in the early detection of SARS and other epidemics (see examples in the post from March) and has been more recently applied to avian flu, but has mostly remained an obscure operation known only by public health insiders. Then Larry Brilliant mentioned it at TED2006 as the model upon which to build an Internet-based global early warning system in the form, this time, of a specialized search engine. Brilliant is the head of Google.org, the nonprofit/forprofit philanthropic arm of Google. That has attracted attention, but also created some confusion: While Brilliant is pushing for developing GPHIN, that's not a Google.org project. When Brilliant gave that speech he was accepting the TEDPRIZE, and was basically asking the people gathered at the conference (including myself; we call ourselves TEDsters) to help in making his vision happen. Several dozen companies and organizations have pledged support.
I asked Ron St John a few questions:
What is the current status of GPHIN? To what is it been applied?
The current GPHIN system that was launched in 2004 has multi-lingual capacity. It is able to monitor and gather relevant information about potential public health threats in eight languages: Arabic, Chinese (simplified and traditional), English, Farsi, French, Spanish and Russian. The non-English articles are “gisted” into English and vice versa using the system’s unique machine translation engine. The purpose of the gist is to provide the user with the essence of what the article is reporting about. The system is comprised of two components: an automated and a human analysis component. The primary function of the automated component is to process the large volume of relevant information in a timely and efficient manner in order for the team of analysts to review the information for situations that may have serious public health consequences or trends.
What can GPHIN currently do other than flag important information or unusual patterns related to diseases or pharmaceutical problems?
GPHIN’s current capacity was built around the timely dissemination of relevant information to its users. The disseminated information covers a broad scope of public health issues such as animal and human infectious diseases; non-communicable diseases; plant diseases; chemical, radio-nuclear or biological incidents; product safety; food security and safety; and natural and environmental disasters. During a public health emergency, the GPHIN team of analysts provide status reports on the magnitude and the geographical distribution of the incident; the contro and prevention measures that are being considered or have been implemented and any other pertinent information.
What are the current options for developing GPHIN?
To date, the GPHIN system has been enhanced and developed through a Collaborative Research Agreement with Nstein Technologies Inc, a Montreal based company that specializes in content management software based on linguistic artificial intelligence. The ability to continue to enhance the GPHIN system as needed is contingent on the availability of funding. Presently, funding for any enhancements or further developments to the GPHIN system is from users of the GPHIN system.
Long-term objectives also include adding analytical capacity to the system (to offer better context). Dr. St John said the he "personally endorses" Brilliant's vision to expand GPHIN, but added that so far discussions have been "conceptual only".