However, it was not until our emergency medicine rotation several months ago that we really saw the need for Symcat. When we started the rotation, we were quietly imagining something between ER and House. We were surprised to find just how many people were there for things like sore throat, headaches, and colds. In fact, there's data that suggests that as many as 44 percent of emergency room visits are unnecessary and can be better treated at a primary care office or urgent care center. It occurred to us that patients were not getting access to good information about when and where they should seek medical care. We believe a simple interface and intelligent algorithms could help answer this question for patients.

Who should use Symcat?

We'd like for anyone with new symptoms to turn to Symcat to inform their health care decision-making. Despite the fact that four of five people with Internet access search for health information online, the medical community has for the most part shied away from offering actionable information. There are notable exceptions like the Mayo Clinic, and HealthTap is mobilizing some physicians, but when it comes to finding out what's wrong, there is still a vacuum that media companies are more than happy to fill.

Ultimately, we're hoping to build a community around helping people find appropriate medical care. We all turn to our friends and family before we go to see the doctor. We're looking for wisdom: someone who will say, "Oh yeah, I had that. I just scheduled with my doctor and it was OK." Symcat will collect that wisdom in a structured way so that it can be easily communicated back to patients. So, another answer to the question is that we'd like everyone to participate in helping each other. Symcat just makes it easier to do that.

Can you discuss some of the technical details of how Symcat works?

Symcat is what's called an expert system. At a high level, it calculates the frequencies of symptoms and diseases from clinical data and uses that past data to inform a prediction about what could be wrong. We train the algorithm on data from doctors' offices and ER visits, where patients are typically sicker than those just searching on the Internet, but it's not a bad start. Fortunately, our algorithms can very naturally learn from additional users, so they will adapt over time.

Do you have any stories of friends or family using Symcat?

Of course! I insist that all of my family use Symcat. Actually, people who have used Symcat sometimes share their stories with us about how it has helped them. One of our users had a fever and stomach pain. Symcat told him that he was likely to have appendicitis or diverticulitis and that he should go to the emergency room. He went and he was glad that he did because he was diagnosed with diverticulitis in the ER. These stories are obviously very encouraging for us.