Health IT

Startup strives to build a better symptom search engine using patients own words, open source data

Photo Apr 18, 3 24 06 PMA patient facing search engine app MedWhat wants to achieve something its co-founder believes is lacking from similar tools — fast, comprehensive responses to patient questions no matter how simple or complex.

In an interview with MedCity News, entrepreneur and co-founder Arturo Devesa said he believes two ingredients are essential to achieving that: open source data from respected medical institutions and natural language processing — allowing people to ask questions in their own words. He envisions a platform that can transform mobile phones into virtual primary care physicians.

The company was recently accepted into the nonprofit student startup accelerator program at Stanford University– StartX. The MedWhat app is currently available on Android and iOS systems.

Here is how it works. Users are encouraged to ask full questions — the longer the question and the more relevant words they enter the better the answer. It distinguishes what its search engine does from the basic keyword search system, which determines the importance of a website based on things like the words it contains and  how often other websites link to it.


MedWhat uses a technology referred to as “semantic search,” a process of understanding the actual meaning of words, concepts, attributes and providing an answer.

Devesa admits it is a work in progress but the goal is to take the company’s current 60 percent accuracy rate and increase it to 90 percent by securing investment to expand its database of articles from groups such as Medline Plus and the National Institutes of Health. “We are only using 500,000 articles, but once we get more resources we will be capable of using 12 million articles.”

As Devesa sees it, WebMD and iTriage have pioneered consumer facing medical queries but MedWhat is offering something better. HealthTap can generate vague responses from its team of doctors or directs users to questions asked from other users that have already been answered. Although symptom checker iTriage has been very successful, it often sends users articles to read in response to their medical queries. Devesa envisions a service that’s more specific and contextualized.

“iTriage is missing the quality data because it doesn’t let people ask full questions,” Devesa observed

Asked about the MedWhat’s business model, Devesa points out that it is reliant on providers and payers and has three components. A health insurer, for example would integrate it into their website or through their platform. Advertising offers some interesting opportunities. Say you had a stomach ache after eating a salad. You could query MedWhat and find that spinach contains folic acid which may have led to the belly ache. Cue an advertisement from a drugstore offering a discount on an antacid.

Data mining also provides some interesting opportunities that could provide useful information for companies based on de-identified data from user queries. Patients Like Me, Devesa observed from his research, generates a great deal of business through data mining alone. It could provide companies with de-identified data that could be used for drug side effects, health trends, correlation of symptoms and diseases and relationships that have not been previously considered. But data mining won’t really be practical until the company has grown more.

The company sees a lot of scope for growth, particularly if it can attract a chunk of the millions of health queries Google processes every week. By creating a smarter search engine for consumers, their medical queries can be more useful and help consumers manage their conditions more effectively. It will be interesting to see as the company evolves whether it will achieve its vision of creating virtual primary care physicians within an app.

We’re bringing active investors and corporate development execs together with promising startups at MedCity News INVEST. Join us April 12-13, 2016, in Chicago for the premier national investing conference based in the Midwest.

1 comment