An interview with Dr. Abel Kho, Director of the recently launched Center for Health Information Partnerships (ChiP)
Dr. Kho is a practicing physician at the Northwestern Medical Group and director of the new Center for Health Information Partnerships (CHiP) in the Institute for Public Health and Medicine at the Feinberg School of Medicine. Dr. Kho is also the principal investigator of Healthy Hearts in the Heartland, and co-executive director of the Chicago Health IT Regional Extension Center (CHITREC). His research includes creation of multi-institutional clinical data networks, health applications of geographic information systems, and data set-driven decision support.
NPHR: Tell us a bit about how you got to where you are?
Dr. Kho: I really debated whether to go to medical school and I only applied on time to one medical school. I always wanted to do something with technology. I remember while I was interviewing for my residency at the University of Michigan, the faculty member asked me, “What do you think are the two biggest innovations in health care?” And I said, “genetics and the rise of computing power”- and of course I did not get in – but I still believe that. If you look at where things are going, technology is not going to fade; it’s only going to be more and more ubiquitous and we have to do a better job and think of smarter ways to integrate technology to improve health care. Likewise we are only going to get more insight into how genetics affect our lives and health outcomes – it will happen.
How did you first get started in health informatics?
Dr. Kho: During my chief year in Internal Medicine Residency I had a great mentor in Robert Schilling , (below with Dr.Kho) the inventor of the Schilling test for B12 deficiency and a hematologist. He and his colleague Dr. Warren Slack were professors at the University of Wisconsin – Madison where I did my residency. During my chief year I had a little bit of research time and started helping Dr. Schilling with this database that he created of all the known hereditary spherocytosis patients he had ever encountered. In exchange for working on this database, he would take me fly-fishing. So each time we went fly fishing he gave me some hand tied flies and we would have a beer and just talk about cases and career.
During one of these sessions he told me if you are ever interested in really working on these kinds of database projects, I know this guy Clem McDonald, in Indianapolis, who is also interested in building electronic medical records, you should meet him. So I took time some time off and met Clem McDonald. He invited me to do a fellowship in informatics in Indianapolis but advised me that ultimately I should do what was best for me. During this same period, I had just finished transitioning the whole residency program to a fully online system for tracking schedules, so I realized I was starting to do more work on how to change systems to improve the way we deliver care. I had trained previously in theoretical and applied mechanics at the University of Illinois Urbana-Champaign College of Engineering before going to medical school. Coming from this background during my residency training I kept thinking, we can do these things better, and we can use technology to scale improvements up – not just for my patients but for a whole region. So at the end of my residency training I became interested in tools that allow us to scale up health care improvements. So as I moved to Indianapolis I started asking, “What are health care problems that can be fixed and scaled?” At this time I read the book “The Coming Plague” and one of the questions raised in the book is drug resistance. It led me to wonder: how do we track drug resistance? So we developed a database and alerting tools to track drug-resistant bacteria in our county hospital in Indianapolis. That project morphed and grew and now we track not only MRSA or VRE at one hospital but all major hospital systems in Indianapolis, and we also automatically track all the multidrug-resistant gram negative organisms too. That technology does not have to just exist there. Now that we have created it we want to make it available to any region, as our system is relatively portable. So we can pick up our system drop it on a local server – turn on a couple of data feeds to it and we will be able to track similar infections in this region.
How do you use technology to improve patient outcomes?
Our current approach is fairly straightforward – we start with a simple problem in health care and ask how can technologies make it more efficient and how scalable is it and then we try to solve that simple problem and scale up our solution. Another project we’ve been working on recently: a couple of years ago we got a grant to install electronic health records in Chicago (CHITREC) and at the time we had to fight resistance like, “what are you doing, this isn’t research!” Still, we believed this work was important and would help us develop a framework for capturing and better understanding health data. Over the course of this work we have developed wonderful relationships around Chicago in non-academic settings where we have installed electronic health records. These non-academic settings include all the federally qualified health centers in Chicago, and numerous small clinics scattered around the Chicago metro area. This preliminary work allowed us to successfully apply for a recent even larger grant to directly connect the data from all the local health centers around the Chicago area (about 250 clinics) to a central common node so we can start to first measure the baseline quality of care in a region and to assess the effect of clinical and population-based interventions on a region – all data collated from the interconnected network of local health centers. It took us 5 years to build the framework to ask research-relevant questions but now we have both the relationships and the resources to tackle research questions that will affect 3 states: Illinois, Wisconsin and Indiana.
What specific questions do you think this new infrastructure allows you answer?
Dr. Kho: Much of this work is at the interface between primary care and public health – a lot of the work we are doing depends on public health partners, because when you think broadly, informatics and public health are both broad system-wide approaches to solutions to problems. If we succeed we will be able to know whether data capture through routine primary care can effectively measure the health of certain populations by following and tracking the health of those populations through time and to determine if those interventions are working.
The question this approach raises is how representative of the general population is the sample captured by our electronic health records. So what we are trying to do in parallel is a validation study to determine how representative our medical records obtained data is of the general population. One validation is at the population level and the other is at the individual level. Great work done at the Alliance of Chicago has also been trying to validate these data at the individual level through interviews and public health surveys that are linked to electronic records data.
What are some challenges of managing a comprehensive medical records system?
Dr. Kho: As we have worked on this project a few challenges have emerged. And a major challenge for research based on electronic medical records is the question of unique patient identity – which comes up over and over again -particularly as we start doing work across multiple institutions. The reality is a person or people do not just interact with one health center, they interact with multiple health centers through time. The challenge becomes: how do you preserve the privacy of the individual, while preventing oversampling? So we had to design a novel approach that allowed us to do this in the past few years. This approach is now being employed by several institutions that are part of the Chicago Area Patient Centered Outcomes Research Network (CAPriCORN) (the PCORI-funded regional network of institutions that have agreed to form a clinical data network to conduct comparative effectiveness research). The long-term goal will be to integrate data from large institutional networks from groups like CAPRICORN to the data from smaller health centers from CHITREC to get a more comprehensive picture of the health of the region, and this can be used to track and measure health outcomes and intervention effectiveness. To help ensure we address any privacy concerns we are working closely with a health information privacy expert at Vanderbilt (Bradley Malin) for the past several years. He ensures the way we do our work minimizes the risk of identification and re-identification of participants. Fortunately he has been involved with writing the federal legislation on health privacy so he also ensures we are complying with federal regulations.
We are in a data age and everyone realizes the tremendous value that data has. But like all things of great value, folks that have it want to protect it and keep it for themselves. But greater value can be obtained from data when it is combined or compared with other data. To do this effectively you have to have appropriate stewardship, expertise and a human touch. So we believe the center is an opportunity to bring all the different components that make data powerful together, in a safe, secure, trusted and purposeful environment so that people can come up with and bring together their questions, methods, and data in a place where they can realize the greatest value from their data. This is important because a lot of the people who have the technical expertise don’t have access to the data and conversely a lot of the people generating valuable data don’t have the requisite tech savviness. We find that when it comes to data it is important to provide a safe stewardship environment that understands that people are dealing with something that is very valuable to them and need a relationship that provides clear value to all partners. This is a recurring theme in many of our projects.
What is the long-term goal of CHiP?
Dr. Kho: Our main goal through the center is to empower and enable a broader community than we currently have, to ask and answer their most pressing health questions. There is no upper bound to the types of questions community members might come up with. Personally, I would love to see work using these patient-based health records to result in true innovations in improving quality in clinical care. And I believe in time these resources will give us enough power to conduct certain types of analyses – particularly questions that relate to the influence of genes and environment on disease. A comprehensive, representative set of patient health records system over a diverse region like Chicago makes it more feasible to conduct studies like these.
What kinds of services are you able to offer at CHiP and what infrastructures do you have in place?
• Methodology experts: we have domain experts that are important for health data projects. Privacy and security foremost. Big data analytics and folks who are really skilled at that. Epidemiologists and informatics folks who understand the key methods and also library and informatics folks who understand how to categorize and index data.
• Database: We have already have built connections to extensive health records data through CAPRICORN and CHITREC and we are currently working on creating snapshots so people can get a better sense of the types of data we have as well as the geographic scope of that data especially as our data keeps expanding.
• Partners: We see an opportunity to engage industry partners; where we can’t or have not currently scaled a project, we can do so with industry partners. So what we are really hoping to engage are people, communities, methodologic experts, and data stewards as the framework of our services. Our goal is to create true partnerships that help our partners make the most of available data. We are starting off with three tracks to provide an anchor including one focused on gene-by-disease-by environment type studies.
As you look back at your career so far what lessons have stood out the most?
Dr. Kho: I think willingness to take a risk. One thing we try hard to instill in our team is the importance of taking risks and failing. A failure in one direction can be very empowering and can lead you in surprising directions. We’ve been lucky in that some risks have been really rewarding, but probability is also on your side if you fail enough times. It’s a wonderful confluence right now – it turns out that being an informatics person who is also a primary care provider is a really good thing right now and there are good funding opportunities for this type of research. I have never seen anyone succeed greatly by just following the path – it just does not happen – you have to be willing to take a risk and jump off the path.
Thanks for your time Dr. Kho.