Congratulations are due to John Schrom, at 26 years old both an
epidemiologist and now a bona-fide hacker can boast of first prize in
Practice Fusion’s latest challenge
to find a use for their massive health data sets. Schrom, an
employee of Hennepin County Medical Center serving Minneapolis, also
develops code for a company named Epicenter, and found a free
weekend to code up a program that displays data in different
groups–by state, by diagnosis, by age range, and so on–and lets the
user quickly find where at-risk populations are. He explains the
purpose and use of the program in a brief
video, embedded below.
This kind of epidemiology is not the disaster kind, like trying to
quickly identify flu outbreaks (which Google found it could do by
totaling up significant search strings). Rather, it answers such
questions as “which age range or racial group should we concentrate on
when promoting preventative measures?” or “how well did our hospital
perform compared to other hospitals facing similar diseases?”
Thus, the Epicenter app reflects a recent concern with delivering more
carefully delineated treatments to populations. Genetic testing, for
instance, is starting to identify more precisely who responds to a
particular drug. Whereas up to now doctors could only say, “This seems
to cure 80% of the people we give it to,” they are starting to predict
better who falls into that 80% and who is among the 20% for whom the
drug would be a waste. Many institutions–notably the Veterans
Administration, but even a single large hospital–are also learning to
determine which types of patients they are treating well, and where
they fall down, by collecting data on each intervention and running
the stats. Schrom’s app fits into this trend.
Emily Peters, Director of Communications at Practice Fusion, who was
also a judge in the contest, says, “John comes from the frontlines of
epidemiology and applied his knowledge and experience in building
Epicenter. We love that the concept is so cool and that it offers a
very tangible benefit and practical use for the medical community at
large.”
The Practice Fusion data is quite detailed while being sufficiently
de-identified (at least according to current standards in the medical
field) to meet privacy requirements. So with Schrom’s app you can
drill down through demographic data to the level of a single patient.
Schrom told me the app is a preliminary prototype and will be
developed in several ways. Currently you can look at data by state or
by groupings of states that you choose to combine; he would like more
targeted geographic comparisons. He’d like to add lab data and make
the types of comparisons more flexible. Finally, he’s aware that the
graphical interface could use a facelift. So check it out, and come
back for more.