Implementing a PrEP Prediction Model

Julia Marcus, PhD, MPH, provides advice for health care systems looking to implement a model to identify potential PrEP candidates.

Segment Description: Julia Marcus, PhD, MPH, assistant professor at Harvard Medical School and Harvard Pilgrim Health Care Institute, provides advice for health care systems looking to implement a model to identify potential PrEP candidates.

Interview transcript: (modified slightly for readability)

Contagion®: In the conclusion of the abstract you indicate that other health systems should attempt this approach. What advice would you give to clinicians who are looking to implement this method?

Dr. Marcus: Variables that floated to the top as being important in our model may apply in other health care systems, or they may not, and I think that's something we still need to test. You had asked previously what the important variables were, and they fell into different data domains, demographics, medication use, diagnosis social history, and some of those you could imagine — let’s say age or sex or race – are probably going to be important in any setting, but others may not be as important depending on what the epidemic looks like.

And so, for a health care system that wants to implement something like this, I would say I would recommend looking at existing models that have been developed like ours — and, there is a model that's been developed in Boston as well – and seeing what variables were important, and then looking at whether those are important in your system. I think what we all want to do is come up with the simplest possible model that we could then implement moving forward.

More information about Dr. Marcus’ and colleagues’ investigation on using electronic health record data to identify patients who could benefit from PrEP can be found here.