Online searches for the term "listeriosis" spiked several weeks before last summer's outbreak of the illness was revealed to the public, an article reviewing Internet-based disease surveillance systems shows.

In fact, in Ontario, Google searches for the term listeriosis — the illness caused by the bacteria Listeria monocytogenes, known as listeria — started to climb as early as mid-April. Inquiries from British Columbia, Saskatchewan, Manitoba and to a lesser degree Quebec started to climb in mid-July.

Searches from all five provinces peaked the week before Maple Leaf Foods announced a recall of listeria-contaminated luncheon meats and federal authorities warned the public of an outbreak.

A disease detection system that monitored searches of health topics on Google and other search engines might have sounded an early warning, the article's authors said. But they noted it remains to be seen whether such a system would be able to tease out true signals of disease against a backdrop of myriad random searches prompted by curiosity, school projects and other motivations.

"Could it have told people sooner there was something going on? Could it have signified to public-health or food-safety people that there are a lot of people in different parts of the country that are looking at this and maybe there is something more national and not just regional? It's hard to know," said Dr. Kumanan Wilson, a scientist with the Ottawa Health Research Institute and first author of the article.

"But it is provocative that something was happening that correlated with the illness and predated the official announcements of the [Maple Leaf meat] recall and the official announcement that there was a real problem."

The article was published electronically by the Canadian Medical Association Journal on Thursday, which happened to be the sixth anniversary of the day the World Health Organization issued a global alert about a new and severe form of pneumonia being seen in parts of Asia. That was the world's introduction to SARS.

Tool already used in business

Wilson and co-author John Brownstein, an assistant professor at Harvard Medical School, decided to look at whether search engine data mining — a tool already used in the area of business intelligence — would be useful for early disease detection.

They looked to see if last summer's listeria outbreak would have been detectable through Google search trends, first looking for the term "listeriosis."

They were surprised to find searching for "listeriosis" peaked before the problem was public knowledge and continued to decline, even after the beginning of what went on to be the largest meat recall in Canadian history.

So they also searched for the better known term "listeria" and found searches for it increased in step with the official announcement of the recalls and outbreak warning.

The authors said they can't tell from the Google data who was doing the searching before the outbreak was publicly acknowledged.

"We don't know who did it. We just know that we saw this pattern," said Wilson, who is also a professor of public health policy at the University of Ottawa.

"It could have been the family members of the people who had been in hospital if the diagnosis had been made at the time. It could have been their physicians. It could have been public health officials. It could have been food safety officials. It could have been industry officials."

The authors also cannot tell how many more searches than usual would have been needed to create the spike they saw. The Google Insights for Search site doesn't quantify searches by number; it shows trends in the form of curves on a graph.

"We don't know the volume accompanied by this search. All we can say is from a baseline level, there was increased activity in searching for the word listeriosis," Wilson said.

Class-action settlements in process of being approved

When federal officials later pieced together the cases linked to the outbreak strain of listeria, the cases traced back to late May.

But it was only late on Aug. 16 that the Canadian Food Inspection Agency notified Maple Leaf Foods that some of its products had been linked to cases of listeriosis. The company announced a recall the next day, and on Aug. 20 federal officials publicly warned of an outbreak linked to the contaminated meat.

By the time the outbreak was announced, one person had died and 16 other cases had been confirmed. At least 20 deaths have been linked to the outbreak.

On Tuesday, Justice Ron Barclay of Court of Queen's Bench in Saskatchewan approved a $25-million settlement in a class action lawsuit against Maple Leaf Foods. Ontario Superior Court Justice Paul Perell will officially certify the lawsuit next week.

Barclay's decision also approves the settlement for Saskatchewan, Manitoba, the Atlantic provinces and the three territories. The agreement still needs approval from a Quebec court where a hearing is scheduled for March 20.

Though both Wilson and Brownstein think web-based information offers potential — Brownstein has already set up an online alert system called HealthMap — they say it remains to be seen if a system built on search engine searches would work.

Brownstein, an epidemiologist who also works at Children's Hospital Boston, said the designers would need to figure out how much additional searching should trigger an alert, and, of the thousands of possible search terms, which ones would be worthwhile to monitor.

"In terms of the practicality of doing this on a perspective basis ... we really have to figure out how this would work," he said from Boston.

Canada has been a leading player in electronic surveillance for infectious diseases. In the 1990s, Health Canada created the Global Public Health Intelligence Network, a program whose analysts monitor news items from around the world looking for hints of disease outbreaks.

GPHIN, as it is called, alerted the WHO to a spike in pneumonia cases in China's Guangdong province in November 2002 in what later was realized to have been the start of the SARS outbreak.