The RocHackHealth was about hacking health
care data. One of the challenges was to try and predict hospital readmissions – whether
people who were discharged from the hospital and went home after being ill, would come
back to the hospital. Obviously, when you go home, you want to stay home. So, we had
the groups work on a data set to try and predict who might come back and the ultimate goal
is to take some of these methods and to implement them here at the University of Rochester and
translate those into ways of spotting patients that might be at high risk of coming back
to the hospital and intervening to try and change that. The second competition was looking
at prescribing patterns of controlled substances – things like sedatives and narcotics. There
is a problem with over prescribing these medications and some individuals prescribing way more
than is normal. So, we used a publicly available data set that’s put out by Medicare and the
team that won, turns out identified a number of individuals who had actually been investigated
for over prescribing and some of whom actually went to prison. The last competition was to
essentially create a social network of how medical providers are linked to each other
by the patients that they share. We ultimately hope to use that kind of mapping to look for
groups of providers that do an outstanding job of taking care of people with certain
medical conditions and then to go in and ask, “What are they doing?” All three of the groups
showed real creativity and promise and we are hopeful that within a year that we will
be able to go into much more depth and apply these methods to more complex and real world
data sets here at the University of Rochester Medical Center and our goal is to improve
patient care and improve the health of our community using these methods.