After years as an award-winning pediatric neurologist and neuroscientist, Michael M. Segal MD PhD elected to combine his passions for medicine and computer science by founding SimulConsult. Years earlier, during his residency, he had first written some decision support software for diagnosis, but the lack of computer use in medicine caused him to put the idea on hold and take a faculty position at Harvard Medical School.
As the Web emerged as a new means for delivering the diagnostic decision support, and as doctors began to use computers, Dr. Segal founded SimulConsult and began working on the product part-time while he continued on the faculty of Harvard.
He turned his attention full-time to the company in 2002 and led the effort to build the database to its present level of over 6,900 diagnoses, as well as wrote the core software, based on a detailed temporal-based statistical model, and launched case-based education using the software for residents in cooperation with the Child Neurology Society and the Professors of Child Neurology.
Lynn Feldman, Harvard MBA joined him in mid-2008 as CEO, and has helped build a business around the product. SimulConsult is a private company and got its first outside capital in October 2008. Since then it has also gotten over $5 million in NIH support to develop and test the software.
With initial support from a Massachusetts START grant, SimulConsult rewrote its software to support mobile access, releasing the new subscription version in April 2019.
Diagnostic error reduction
In multiple studies, SimulConsult has been shown to reduce errors, defined defined as not listing the diagnosis or its category in the differential diagnosis. The assessment shown here was done with case vignettes that were difficult enough that in 36% of case testing instances testers did not include the correct diagnosis, or even its category, in the differential diagnosis before using the decision support (“baseline”).
They experienced this level of error despite testers being allowed to use other current standard of care resources such as textbooks, articles or Web searches, on which they spent from 7-15 minutes prior to recording their baseline differential diagnosis and workup. Results indicate a significant overall effect of the software in reducing diagnostic errors from an average of 36% to 15% overall, a relative reduction of 58%.