Waze has turned the boating world upside down by taking the approach of combining basic map data with participatory feedback on best routes, annoying traffic and other road conditions. Today, one of the founding employees of Waze applies a similar concept to the world of healthcare.
StuffThatWorks, an Israeli start-up that has built a platform for people to research various health issues and then bring their own experiences and treatment approaches while researching what others have done, has raised $ 9 million in a seed round from Bessemer Venture Partners, 83North, and Ofek Ventures.
Even though this is the first funding for its announcement and the first time StuffThatWorks has gotten its name, the startup has quietly built its user audience through word of mouth, particularly by tapping into other networks where people come together to talk about health. problems, such as groups on Facebook (about half of which, it turns out, are related to medical conditions). To date, and with relatively little effort, it has gathered nearly 180,000 contributors around 110 conditions, totaling nearly 10 million data points.
The challenge that StuffThatWorks is tackling is the problem and prevalence of chronic disease.
Yael Elish, co-founder and CEO, estimates that around 10,000 chronic conditions have been identified to date in the world of medicine, but only about 40 of them have been the subject of extensive research and established treatments. . This leaves a long tail of rare “orphan” diseases and conditions that simply have not had a critical mass of affected patients, and the subsequent attention and funding of the medical community, to be treated properly.
On top of that, to date there hasn’t been a centralized and organized place where you can look at real-world evidence and patient-reported outcomes related to it.
Elish discovered this problem firsthand when her daughter developed a chronic illness, and it was by exploring a number of community forums and other resources – none of which are truly optimized for data mining, and so much more just to share anecdotal moments – that she began to learn about other people with the same condition and to see the different approaches different doctors have taken to help them.
One day, she managed to find a treatment that she had never seen before, which helped her daughter, but it could just as easily have escaped her.
“It was my a-ha moment,” she says.
She was still working at Waze at the time, which was then taken over by Google. But she realized the opportunity to apply some of the crowdsourcing dynamics that had been used to build the Waze dataset to this conundrum, to build a centralized place, a repository for the experience and knowledge generated by the people. patients who could be shared with others facing the same problems as them.
She eventually quit and founded StuffThatWorks along with two others, CTO Ron Held and Chief Data Scientist, Yossi Synett. Held, a mathematician by training, is a former head of an IDF intelligence team. And Synett is an expert in machine learning, AI, and hands-on analytics, and between the three, they’ve built an artificial intelligence platform that ingests many data points, combines them into actionable data that people can use. and let them – the users – build more boundaries in this effort.
The interface for creating the StuffThatWorks dataset comes in the form of a series of questions that people answer. These are not multiple choices, Elish points out, but intentionally left to people to answer in their own words, so people have the chance to really speak up and be as specific as possible.
“It was important for crowdsourcing to be successful,” she says. Natural language algorithms “read” responses and help begin to categorize data, and eliminate outlier inaccuracies (accidental or otherwise) – as we’ve seen with Waze.
The principle is that people opt in when they provide their data and you can post anonymously – just like with Waze – but in any case, it remains anonymous on the company’s platform and you can delete it yourself at any time. However, the company also says it is collaborating with “a limited number of researchers, medical organizations, and patient advocacy groups on Patient-Reported Outcome (PRO) research.” caveat emptor on how and if this will pose longer term data protection issues.
The data then goes through a series of “stretch goals” so to speak growing: First the condition is established with a profile on StuffThatWorks. Then after 100 contributors you can start to see initial information about the disease including age of onset, symptoms, aggravating factors, and treatments can start to be shared and viewed. After several hundred, the machine learning is starting to be able to classify different treatments by effectiveness. After thousands of contributors, the algorithms are able to predict for you the visitor what might be the most effective treatments for you.
Other startups have started exploring how to leverage the power of AI, crowdsourcing, and simply the internet to better tackle the problem of the large amount of permutations that exist with chronic and unresolved conditions. RDMD, itself founded by Onno Faber, who suffered from his own rare disease, is also using crowdsourcing to help connect people with rare diseases to researchers who are developing drugs and treatments to address them, bypassing some of the same walled gardens that StuffThatWorks treats in its endeavors.
Others, like Paige, are applying AI to “read” and better understand the pathology of cancer, which has its own mine of permutations, defeating many attempts to treat it in the general population.
The biggest challenges remain, of course, but these types of approaches seem like a crucial step in trying to overcome them.
“With more than half of the world’s population suffering from at least one chronic disease, chronic disease is a rapidly growing global epidemic,” Adam Fisher, partner at Bessemer Venture Partners, said in a statement. “We believe that StuffThatWorks can not only help millions of people access valuable knowledge about treatment effectiveness, but also disrupt and innovate the way real-world patient data is collected and analyzed today. “