May 7, 2012
Online crowdsourcing has a proven record of delivering good answers in evaluating consumer products and providing ideas in research and development challenges. A problem is presented to an online audience, members respond with suggestions and observations, and they do so for free, motivated by their own interest.
Researchers at UCLA worked from the assumption that large groups of non-experts could be trained to recognize infectious diseases with the accuracy of trained pathologists. Microscopic views of malaria-infected red blood cells are distributed for viewing by the online crowd through digital images. The UCLA team found that a group of non-experts they recruited to play the game delivered accurate diagnoses within 1.25 percent of the results that medical professionals would produce.
The power of the group helps deliver accuracy, researchers say. “If you look at one person’s response, it may be okay, but that one person will inevitably make some mistakes,” said Aydogan Ozcan, an associate professor of bioengineering and a co-author of the crowdsourcing research. “But if you combine 10 to 20, maybe 50 non-expert gamers together, you improve your accuracy greatly.”
If hundreds, even thousands, are taught to play the UCLA game, many more cases of the disease could be diagnosed more quickly, at no cost, with a high degree of collective accuracy, according to a UCLA press release.
The World Health Organization (WHO) estimates that about 216 million malaria cases occurred in 2010 worldwide, resulting in more than 650,000 deaths. Ninety percent of those deaths were in Africa, WHO reports, among children under 5.
Early diagnosis of malaria is important because untreated the disease can escalate very rapidly — going from mild symptoms, similar to those caused by other viral infections, to severe symptoms of coma and renal failure within hours. Currently, diagnosis must be done by a trained pathologist equipped with a microscope who views images of cells from a suspected malarial patient and counts the disease-causing parasite. It’s a time-consuming task, and the capacity to put the personnel and equipment in place where they are needed is difficult in underdeveloped countries.
“Scaling up accurate, automated and remote diagnosis of malaria through a crowdsourced gaming platform may lead to significant changes for developing countries,” Ozcan said.
But getting an early diagnosis is only part of the problem. Cases of other viral infections are frequently mistaken for malaria, and drugs are prescribed inappropriately. Better diagnosis through the crowdsourcing game could address that problem too.
“It could eliminate the current overuse and misuse of anti-malarial drugs, improve management of nonmalaria fevers by ruling malaria out, lead to better use of existing funds, and reduce risks of long-term side effects of anti-malarial drugs on patients who don’t need treatment,” said Sam Mavandadi, a postdoctoral scholar in Ozcan’s research group and the study’s first author.
As an introduction to the game, each player is briefed on malaria and views sample images of malaria-infected red blood cells. The player then moves into the actual game, where he or she views multiple frames of red blood cell images and use a syringe-like tool to kill the infected cells.
“It is diagnostics outside the box — that is, the study introduces an entirely new concept in diagnostics with the use of games for this purpose,” said Karin Nielsen, a professor of infectious disease in UCLA’s Geffen School of Medicine. “The potential applications of this new approach are immense.”
The UCLA team is working to set up field trials of the diagnostic tool. Funding for the research came from several agencies in the U.S. government, including the U.S. Army Research Office, the Office of Naval Research, the National Science Foundation and the National Institutes of Health.