Acting on the idea that millions of heads are better than one, a British research firm is crowdsourcing cancer, and has just invited everyone on the Internet who has perhaps five extra minutes to spare to help it cure the potentially deadly disease.
Cancer Research UK has asked web surfers to comb through tens of thousands of pictures of breast cancer cells pulled from studies to help identify tumor cells. When volunteers spot cells that are stained yellow, they send them on to researchers who then use them to analyze trends between the cells and patient treatment.
Anyone who signs on must first complete a short tutorial, and scientists do check the work to minimize mistakes.
This is not a complete first. Citizen scientists have already helped accelerate discoveries in disciplines such as biology, astronomy and archaeology. And business and marketing companies have used Internet crowdsourcing for years to promote products and services. So, too, have a number of other health care websites.
One celebrated instance of medical crowdsourcing was initiated by Patientslikeme.com, a health data-sharing website based in Cambridge, Mass., that aggregates patient information for research. The site has more than 100,000 people representing more than 500 conditions.
Upon the publication of an Italian study in 2008 that found the drug lithium could delay the progression of ALS – also known as Lou Gehrig’s disease – about 450 members of Patients Like Me’s ALS community decided to test the theory out. Within nine months, they showed the treatment was a bust.
The results were published in Nature Biotechnology in 2011, and may have been the first time a social network was used to monitor patient treatment in real time.
”Medical crowdsourcing can really move along a disease by engaging patients and getting them to share their experience and data,” said Ben Heywood, founder and president of Patients Like Me. “In this sense, it’s helpful for both the collective and the individual.”
But the community approach to medical research is not without potential downsides.
”Crowdsourcing sites provide a novel way for those with less common conditions and researchers to find each other, and that’s a great thing” said Kathryn Schmitz, an associate professor of epidemiology at the University of Pennsylvania’s Perelman School of Medicine. “However, a question that must be addressed is the bias of those who sign up to be connected to these websites versus those who don’t. We know these groups are different.”
But problems of bias are not insurmountable, said Schmitz, noting that bias exists in trials run by scientific experts too.
Self-reported data, she said, has been increasingly embraced by the scientific community as more scientists turn to health social networking sites to further their research. Schmitz said she herself has used Patients Like Me to connect with people who have lymphedema, a condition in which the lymph nodes swell as a result of cancer treatment.
But she urged caution in interpreting crowdsourced results and said she wouldn’t trust the job of a trained pathologist to anyone with a laptop.
“That’s really specialty training,” she said. “I certainly wouldn’t want my mother treated based on a crowdsourced diagnosis.”