Medication Safety in 60 Seconds: Event Stream Data



hi welcome to met safety and 60-seconds my name is Morgan and today I want to talk about one of the challenges in working with event stream data which is the data that we see coming from smart infusion pumps so a merge health one thing we spend a little bit of time on is figuring out how to go from analyzing high level metrics of isolated events like an alert or a pump start and instead tying events together and trying to tell a more comprehensive story with the data so for example let's say that we want to look at all infusions of fentanyl that had a dose more than 400 mics per hour it would be easy enough to just see how many pump starts started at that dose or higher but if you look at those pump starts individually there really isn't a huge difference between those events they look pretty similar but instead if what we want to do is understand what the context of those events were and what the results of those events were then we need to be able to correctly interpret the sequence of events that happened before that pump started and after that pump started only then can we really tell you know was this the dose rate mismatch was the clinician bullying off of the pump or maybe it was actually an intentional and appropriate dose for that patient we really see this pattern in all sorts of events you could do the same thing for basic infusions and Furler overrides once we fully understand that even then we can start automatically analyzing the patterns around the events to tell a full story with the data

Leave a Reply

(*) Required, Your email will not be published