All healthcare startups strive to offer innovative, applicable solutions to problems that plague patient care and reduce efficiency. But not all of those solutions require conscripting medical detectives.
CrowdMed, a San Francisco-based startup, aims to provide patients with elusive, complex diagnoses the opportunity to have large, interdisciplinary teams of specialists and experts take a look at their difficult cases for a flat fee, an opportunity they may not otherwise have.
"I found medicine to be a sample size of n=1 sort of industry," says Jared Heyman, founder and CEO of CrowdMed. "Patients are given the opinion of one medical expert looking at them in isolation."
CrowdMed is Mr. Heyman's third company. He said he knew from his previous experience with starting companies in the market research space that crowds or large groups of people with their collective intelligence properly aggregated tend to be much smarter than even the smartest individual experts in the world. "I wanted to bring crowdsourcing to medicine in a scalable and accessible way," he says.
The company has raised over $3 million in capital to date from Silicon Valley venture funds, has six employees and plans to add a seventh in the new year.
Mr. Heyman took time to speak with Becker's Hospital Review about why a consensus-based diagnosis works in medicine, the company's success rates solving mysterious cases and how hospitals can insulate themselves from risk by crowdsourcing.
Question: What inspired you to start CrowdMed?
Jared Heyman: My sister spent three years with a very difficult undiagnosed medical condition. It took a very long time, almost two dozen doctors and six-figure medical bills before she finally arrived at a correct diagnosis. Fortunately, she got it eventually and is now doing great, but the struggle she had with the medical system was awful. I really got to see first hand the deficiencies that exist in today's medical system when it comes to helping patients like her with complex conditions that either aren't easily diagnosed and treated or that doctors don't see often.
Q: CrowdMed's model seems to be based in both a tech-driven solution and a people-driven solution, is this accurate?
JH: My previous two companies were market research firms, and in market research you're always working with people, whether it's a survey or focus group. You're always tapping into the collective intelligence of people, so I really came to appreciate the wisdom of crowds and how intelligent crowds can be the right mechanisms in place to incentivize them and aggregate their collective knowledge. I guess it's my bias to come up with solutions to problems that involve crowds and tap into the power of productive intelligence, that's perhaps unsurprising given my background. CrowdMed is still very much a technology company, even though the intelligence of our community is what ultimately fuels the answers we provide to our patients, we have very sophisticated technology in place to aggregate their collective intelligence and basically get the disparate opinions of 20 or more people to combine into one coherent answer for patients that's properly filtered and vetted. We have two U.S.-issued patents and have spent a couple million dollars and countless man hours developing our core tech, so that's definitely a significant component.
Q: How does the service work?
JH: A patient goes to our site to submit a case. After they register, they answer an in-depth questionnaire and upload diagnostics and test results. Some even upload their entire medical records. We then use all of the data to automatically prepare their case, which goes live on our site. Once that happens we have a very structured process in place for our case solvers and medical detectives to find a case that matches their unique expertise, then communicate with the patient and with one another to suggest innovative or insightful diagnoses and solutions.
We've resolved over 1,300 cases on our site to date. We define "resolved" as a case having been through our entire process from submission to going live for two to three months to ultimately receiving a list of diagnostic and treatment suggestions from our community with explanations as to why and a probability for each one. A subset of those are successful, which we define as the patient having told us after a case was concluded that they got insights that led them closer to a correct diagnosis or cure. Our success rate by that definition is more than 60 percent.
Another way of looking at success is whether one of our top diagnoses is medically confirmed to be correct by a patient's doctor, which is a more conservative definition, but if you look at it that way, our success rate is still high, more than 50 percent. Especially considering that none of these patients have had their patients successfully resolved through the medical system.
Q: What's on the horizon for CrowdMed?
JH: We're now beginning to partner with at-risk provider groups because we can save them incredible amounts of money. There's a subset of their patient population that has complex conditions and those patients are most likely bouncing from doctor to doctor, having a bunch of tests run and significantly driving up costs. This used to not be a problem for hospitals as they were paid just for procedures and tests, so from a financial perspective they wouldn't be hurt by a patient who bounced around and needed to get a lot of tests done. But now as they begin to bear more risk, it's more of a problem for them.
Patients that used to be profitable are now driving up costs and hospitals are losing a lot of money on them. In those cases, what we do is dramatically reduce the costs of bringing those patients to a correct diagnosis or cure and also reduce the time that's required to get that diagnosis. We've had third parties run studies on us that found we provide an immediate 40 percent cost reduction for these patients and an immediate 75 percent reduction in their provider visit frequency, which makes them much less expensive to manage. If you look at the cost of running patient's case on CrowdMed versus how much money would be spent on medical costs and provider visits, we can provide about 10 times the return on investment in the first year.