International health officials are scratching their heads over the paucity of positive bird tests for a bird flu, especially given that the human case count is more than double the number of positive bird findings. As of Wednesday, 82 people had been infected and 17 had died.
"Something is happening out there that's not being picked up," says Gregory Hartl, spokesperson for the World Health Organization in Geneva.
"It argues for the fact that we have to continue to cast the investigation net widely."
China's Ministry of Agriculture revealed Wednesday that nearly 48,000 tests samples have been taken from live animal markets, farms and slaughter houses across China.
Of those, a mere 39 tested positive: 38 in poultry from markets in the eastern Chinese provinces where most of the human cases have been found and one in a wild pigeon, in the eastern Chinese province of Jiangsu.
Those findings beg the questions: Where is this bird flu hiding? And is China targeting the right species when it goes looking for H7N9?
Infectious diseases expert Michael Osterholm agrees with the suggestion that at this point, investigators need to keep an open mind about where the virus may be coming from.
But the director of the Center for Infectious Diseases Research and Policy at the University of Minnesota says too little is known about how China is testing for anyone to be confident that all the negative bird tests were true negatives. He suggests more information is needed before people can feel sure that the results being reported are truly as puzzling as they seem.
"There are just a number questions here that we can't answer yet based on the available information," Osterholm said in an interview.
"I think the Chinese have been very forthcoming in providing the results. I think the question now is how to interpret these results… based on how they got the results."
Osterholm says the questions that need to be answered are: What types of tests are being performed on samples taken from animals and how well are those tests doing at detecting the new virus?
To put this in context, one needs to understand that there was no on-the-shelf test for this H7N9 virus, which is a constellation of bird flu genes that hadn't been seen before.
When a new virus is found, specific tests to detect it must be developed and validated so that laboratories can find the virus in all samples where it exists, and not confuse it with other viruses. In the three weeks since China realized it was seeing infections with a new flu virus, it has had to develop and roll out tests to find H7N9 in a variety of species and settings.
Osterholm says another issue that could affect the test results relates to who is taking samples from animals and how well trained are they.
He notes with large outbreaks, authorities sometimes need to draw in people who have few or no skills at grabbing a chicken and swirling a swab around the cloaca, the bird version of an anus. That is where, in poultry, one looks for flu viruses. In people, influenza is a respiratory infection but in birds, the virus invades the gastric system.
"They obviously had to pull people in to do this that likely did not have great experience with this kind of testing," Osterholm says.
"We have seen in the past issues where the testing done by inexperienced individuals actually resulted in a substantial reduction in the number of positives versus those individuals who had much more experience."
Other factors that need to be considered are how quickly birds move through live animal markets and where they come from. If inspectors come to test after infected birds are sold and leave the market, good tests may still produce negative results.
"It wouldn't be an error of the test, it would just be that the sampling didn't detect it," Osterholm says.
"The tests could be giving us exactly the right information, but not testing the right birds. And you can still have a substantial under-reporting."
The complexity of testing for H7N9 is frustrating for public health officials who are used to flu viruses like H5N1 that make their presence know.
That bird flu is no stealth agent. H5N1 has infected at least 622 people in 15 countries since late 2003. In people, it causes mostly devastating illness. In poultry, the result is the same. Whole flocks are wiped out by this highly pathogenic virus.
But in poultry, H7N9 is a virus of low pathogenicity — low path, as flu researchers say. It's not killing birds. It's not even making them sick, so far as anyone can tell.
That adds an unexpected layer of complexity to the task of finding out how people are contracting the virus. And until authorities can figure out how people are coming in contact with H7N9, they cannot take effective measures to try to stop the viruses from infecting people.