Until they were replaced by electronic devices in 1986, caged canaries were a fixture in British coal mines, where their high sensitivity to deadly carbon monoxide gas—odorless to humans—made them invaluable sentinels for miners’ safety. If the canaries showed any signs of distress, miners took warning.

The colorfully named Canary Database, a new project overseen by Peter M. Rabinowitz, M.D., M.P.H., associate professor of medicine, also relies on animals as lookouts of a sort, by comprehensively cataloging animal illnesses around the world that might portend outbreaks of disease in humans. “If you have an animal that is sick from an environmental hazard, should human health professionals be concerned?” Rabinowitz asks.

As underscored by recent concerns about an outbreak of bird flu in humans, the question is not hypothetical. In 1999, a veterinarian at the Bronx Zoo reported to public health authorities that crows were dying in unusually high numbers, but it took some time before it was recognized that the birds’ illness was the first sign of the emergence of the West Nile virus in this hemisphere.

In order to overcome gaps between experts in animal health and those in human health, the Canary Database, funded by the National Library of Medicine, makes animal sentinel studies from a variety of biomedical databases easily accessible.

A collaboration of the Yale Occupational and Environmental Medicine program, the Yale Center for Medical Informatics and the U.S. Geological Survey Wildlife Health Center, the database uses an automated and complex research protocol to cull from the veterinary literature papers on animal disease that might have relevance to humans. Five curators, including veterinarians and physicians around the country, review and curate the papers, adding information about epidemiological methods and linkages to human health outcomes.

The information could be used to detect impending disease outbreaks or terror attacks involving chemical or biological weapons. “We want to be a continuing resource,” Rabinowitz says, adding that the database could lead to cross-training between veterinary and medical schools. “We feel that both groups have a lot to learn from each other.”