Projects, designs, and writings on health IT


How we created the most accurate symptom checker on the web

Symcat is a consumer health information website I started with Craig Monsen in 2012. The name comes from Symptom-based Computer Assisted Triage. We used data from millions of patient encounters and predictive modeling to help predict what people would eventually be diagnosed with. We went through the Blueprint Health Incubator in New York and currently run a 5-person company. We were the topic of a Harvard study on symptom checkers and in 2015 reported as the most accurate one on the internet.

Specific Product Challenges

  • Developing a comprehensive medical ontology that includes lay terms
  • Using machine learning to train a predictive model to build differential diagnoses
  • Mashup open sources for medical content (Medline and Wikipedia)
  • Search engine optimization


Ruby on Rails, Python (offline data analysis), jQuery, backbone.js, Devise (authentication)

Validation study

Semigran et al, Evaluation of symptom checkers for self diagnosis and triage: audit study. BMJ 2015


July 2015

Washington Post -

May 2013


April 2013

February 2013

January 2013

December 2012

November 2012

InterWest - Digital Health Entrepreneurs Survey (look for us under “Which company do you wish you had founded”)

October 2012

Johns Hopkins Informatics Grand Rounds - Symcat [video]

September 2012

August 2012

July 2012

WSAV - Tech Tuesday: Symcat [video]
The JHU Gazette - Diagnoses at your fingertips

June 2012

The Health Care Blog - Lightning Strikes Datapalooza

May 2012

April 2012

March 2012

January 2012


Post a Comment