Making it Click: An interview with Julia Adler-Milstein, PhD

Technology is an all-encompassing word that can mean many different things depending on who you ask. In healthcare, it could mean new radiological imaging machines or robots that deliver food to patients. One omnipresent piece of technology is the electronic health record (EHR). UCSF uses a software system called Epic as it’s EHR, as do many other hospitals. It is a complex application that is the center of many policy discussions. At UCSF we have many researchers who use EHR data but only a handful who study the EHR itself and its impact on health care delivery and outcomes. I was able to sit down with one such researcher, Julia Adler-Milstein, PhD, who is the Director of the Center for Clinical Informatics and Improvement Research, to learn more about her role in helping shape the national discussion and her own experience as a researcher who studies how clinicians interact with technology and the related policy issues. The interview is lightly edited for clarity and length.

How would you describe what you do in 30 seconds?

I would say first and foremost I'm a researcher and I try to create evidence that helps guide both federal policy makers and practitioners, which in this case would be hospitals and other health care delivery organizations, about how to make best use of their digital tools to improve health outcomes, providers satisfaction and all the other outcomes a high functioning health system would achieve.

Can you tell me more about your current research either broadly or about something you're really excited about specifically?

There are two types of research that I'm doing right now. One is a continuation of the work that I've done for a long time and one is the newer research since I've come to UCSF. The research that I've done for a long time is focused on thinking about “how do we measure adoption, use, and impact of health IT – primarily EHRs and interoperability - at a national level and be able to identify where barriers to progress lie?” I lead a lot of studies that are trying to measure how many doctors use electronic health records and where you would find them. Do you find them more in larger practices or urban settings? A lot of that work is survey-based work where we do national surveys of different types of provider organizations trying to understand how they use technology or how that technology is having an impact. As a concrete example right now the 2009 HITECH Act created these incentives for doctors and hospitals to adopt electronic health records but all other settings were not eligible for incentives. Notably, there were no incentives for long-term care providers, such as skilled nursing facilities, to implement or upgrade their technology.  This is concerning because we know that the sickest patients are the ones that make the transition between a hospital and skilled nursing facility and making sure that their information from the hospital travels with them is so critical to making a safe and smooth care transition. Some people hypothesized that there would be a spillover effect from the hospital incentives, such that a hospital might help a local SNF to improve their IT system to facilitate improved care coordination. So, we're doing a study right now in which we're surveying skilled nursing facilities and asking them about both their IT infrastructure as well as how they exchange information with hospitals. We're trying to figure out what IT infrastructure is in place, and then assess, if a skilled nursing facility has better IT infrastructure, do they have better outcomes for their patients? From a policy perspective, the hope is that you could then say, ensuring that SNFs and other providers get EHRs – not only doctors and hospitals – isn’t happening on its own and it is a really important investment.

Such evidence creates an opportunity to influence policy in a way that could lead to some funding for those organizations to adopt electronic health records. This project is a great example of my work where there's a policy angle.  We start with new data collection that will help understand the current state, and then usually I try to tie it to an outcome that policymakers care about.   So the goal is not just to learn that 50% of skilled nursing facilities have electronic health records, but to be able to say skilled nursing facilities that have electronic health records have patients that are less likely to have a medication error or and they have shorter lengths of stay.

The newer type of work that I've been doing since I came to UCSF is focused on the interaction between clinicians and electronic health records, determining how we can study those patterns of interaction to gain new insights either into how to design the software itself or in how to design the workflows around it or potentially even just to train people better on how to use the EHR. What's very cool about EHRs is that they track close to every mouse click in the system, so it's a huge amount of data to make sense of and figure out what's meaningful--what patterns are important or not. To give a very concrete example, last year in July we made a very simple switch to the way that Epic was configured here so if there was an encounter at another local health system, like Sutter, my UCSF doctor would see that Sutter encounter as part of my list of encounters, whereas before if you wanted to see that data you have had to go to a special part of the EHR just for data from outside UCSF. So we asked the very simple question of did that little change of integrating that data in a way that made it more visible impact how often our clinicians looked at my records from outside institutions? What we were able to show is there is a huge increase in the viewing of these outside records and it also seems like there's at least in association with a shorter length of stay and less likelihood of getting imaging. So again, it's a very different type of research question just using UCSF data and it's much more micro in that sense, but to be able to show that this user interface change had a big impact on clinician behaviors—how they use the EHRs—and then to be able to tie that to an outcome, that's meaningful.

How would you describe the field of informatics for someone who isn't familiar?

At its core, informatics is about the study of how humans plus technology are better than humans alone.  That work can be pursued in theory-driven way, in the same way you’d think about a traditional discipline such as psychology or physics.  However, with the huge push to adopt EHRs in healthcare, there is a large contingent of people working on highly applied questions related to adoption and use of health IT.  It feels a bit like the field is having an identity crisis because on the one hand there are practical skills related to understanding how to implement and manage an EHR, but does that make you an informaticist and someone who has mastered the discipline?  So you’ve actually asked quite a complex question and one for which there isn’t a clear answer today.

What do you need to study to become an expert?

The most traditional route is via biomedical informatics masters and doctoral programs. From a personal perspective it's always been a little bit odd because I didn't go through one of those programs. My training is in health policy and policy is my first passion. It just happened that the policy issues I was most drawn to were policy issues related to technology. But if you ask me if I'm an informaticist, I would say not. I have not been formally trained and you’d never want me to get close to designing or implementing an EHR!  So would I even say I'm a woman in technology? I guess technically yes, but it always feels a little weird to say that because my background and my formal training has never been in technology. I don't know how code, or do these things that are typical hallmarks of someone who is in informatics or in technology.

I think that's something that we are grappling with the group too, because technology is so pervasive and so many people at UCSF work with technology in some form or another. Some people are experts in certain systems, but they don't consider themselves technologists. It's just a very interesting to hear someone who has conducted such in-depth research on informatics and technology say that you don't identify like that. It’s a gray area.

It’s taken me a long time to feel at home in the informatics and health IT circles. For example, I’ve submitted my work and attended the American Medical Informatics Association’s annual meeting for the past decade, but I always felt that my work was fringe. In fact, one of the first times I presented, someone stood up at the end and said “this isn’t informatics!”  I can still picture it! But I also got a lot of positive feedback and support, particularly from senior members of the community. Now when I attend, I know that the community  sees me as one of them and they recognize the value of my work.

Thanks to Julia for her time and for sharing her thoughts with our members.