Christopher Longhurst, MD is CMIO of Stanford Children’s Health, founding program director of the clinical informatics fellowship of Stanford University School of Medicine, and clinical associate professor of pediatrics and biomedical informatics at Stanford University School of Medicine of Palo Alto, CA.
Tell me about yourself and your job.
I’m the chief medical information officer at Stanford Children’s Health. I have a faculty appointment in pediatrics and a courtesy appointment in biomedical informatics at the Stanford School of Medicine. I help to lead our clinical information technology and strategy for the health system as well as the academic fellows training program.
The new clinical informatics board certification allows physicians working in an informatics role to be grandfathered in for the first few years. Can you describe how you see that morphing into the requirement that applicants complete a clinical informatics fellowship and explain how your program is structured?
This started in 2011 when the American Board of Medical Specialties approved informatics as a board-eligible subspecialty. It’s a particularly unusual subspecialty because you can board in a subspecialty after training in any of the 24 primary specialties. Until 2017, people can grandfather in through extensive work experience and education, after which time the only way to be board eligible will be to have completed an accredited fellowship training program.
What is the audience that you anticipate will sign up for the fellowship?
When we opened the Stanford clinical informatics fellowship last year, we got dozens of applications. Some of those were from physicians with strong computer science backgrounds who wanted to write code and develop apps. While they have an important place in the ecosystem, that’s not what the fellowship program is looking for.
We’re recruiting physicians who are interested in driving improved healthcare delivery outcomes. We’re looking for people who are going to keep their eye on the ball in terms of where we’re headed and using informatics and IT as a tool to improve the delivery of the care that we provide.
We’re really excited about our first two fellows, Lance Downing and Veena Goel, who are doing some amazing work and will be future healthcare leaders. In fact, the mission of the program, we decided, was not to train physicians to become informaticists, but to train the next generation of healthcare leaders in the skill of informatics.
Once the grandfather period is over, who will offer fellowships for those people working in an applied informatics CMIO role that isn’t research based but rather feet on the ground technology adoption?
There are 140 or so medical schools in the United States and 6,000 hospitals. If the fellowship programs are only at those academic medical centers with medical schools, we’re not going to train enough people for the next generation of healthcare leaders. I anticipate, though, that we will see training programs coming up at non-academic medical centers.
In fact, I think it’s important that that happens just as with other specialties. We have internal medicine programs at over 800 hospitals. I think we’ll see opportunities for training informaticists at many, many other healthcare settings.
What’s different between this and the master’s degree programs of the past is that these fellowship programs offer experiential training. It’s the opportunity to come in, be part of the office of the CMIO or other applied clinical informatics environments, and contribute in a meaningful way to real projects. I think that this type of experiential training complements the didactics, but is a critical piece for training our next generation.
The Institute of Medicine’s recent review of medical education questioned why it’s only offered in hospital settings. Why wouldn’t a public health setting for informatics training be equally desirable given the need for population health management?
As part of our fellowship program, we offer rotations not just at the two hospitals at Stanford, but also in the clinical research informatics group at Stanford medical school. We also offer elective rotations in the industry. Our fellows have the opportunity to spend a month at a large company like HP Labs, where they do healthcare analytics research, as well as at a small startup company, Doximity. We think that there are physicians who are going to be working in all sorts of settings and having those experiences is important.
We also have our fellows rotating through the Kaiser and Sutter healthcare systems, where they have an opportunity to see a large, integrated delivery network that’s not an academic medical center.
What subjects will be covered in the two-year fellowship?
We break it down as follows. We think that it’s important that our clinical informatics fellows maintain clinical activity. They’re expected to spend 20 percent of their time seeing patients. We’ve partnered with Bill Hersh and the Oregon Health & Science University distance learning program to provide didactic support, so we anticipate they spend another 20 percent of their time with the classwork. That leaves 60 percent of their time, which is a combination of these experiential rotations and unstructured time for scholarship and longitudinal projects.
The OHSU program is rigorous and you are adding additional elements to it. It will take some work to complete the fellowship.
[Laughs] Well, we expect that the fellows will be working hard, but we also think it’s going to be a really gratifying program to complete training.
What training are medical students receiving in practicing with an EHR and then performing data analysis for research or for population health management?
At most health systems, the training for medical students is pretty limited. They may get a little introduction to the electronic health record systems, but it’s generally focused on the front-end data input and review on single patients, not on population health. Dr. Bill Hersh just co-authored a publication suggesting new competencies for undergraduate medical student training in informatics. I think that we’ll see adoption of those concepts widely moving forward.
Everybody wants to get their specific area covered in medical school education, but it’s already a busy program even though informatics is in some ways as important a stethoscope or a scalpel. Do you think the 10×10 program is meeting that need now and will that change over time?
I think the AMIA 10×10 program has played a really important role in raising the bar on informaticists. Ten years ago, any consultant with experience in clinical information systems could declare themselves an expert. But having some formal classroom understanding of what’s happened in the last 50 years in this field and where the grand challenges lie is important for coming together as a field to attack those big problems.
You’ve done work with a “patients like mine” button idea where a doctor can quickly find similar patients to the one they’re seeing. Are you doing that or is it still a concept?
We have an exciting story that was published in the New England Journal of Medicine in 2011. The story was about a 13-year-old girl with a known diagnosis of lupus who was admitted to our hospital with a flair in her lupus. One of my colleagues, Dr. Jenny Frankovich, asked a really important question, which was whether her lab findings made her at higher risk for clotting and whether we should consider prescribing anticoagulants for her.
Of course, we did what any good evidence-based clinicians would do and looked in the literature, but as in many special areas of pediatrics, there was no literature on teenagers with lupus and risk of clotting. We then asked our colleagues, and the first colleague we asked said absolutely you should anticoagulate. The next colleague we asked said absolutely not. We were left holding the bag with one of these clinical decisions that occurs every day across this nation, but has to be made in the absence of data.
My colleague Dr. Frankovich did something at this point which had not been done before. She used her IRB-approved access to a data warehouse to look at a holistic experience with teenagers with lupus at Stanford over the last five years. She found 100 similar patients, and on the day that we admitted this teenager, was able to determine that her lab findings put her at six- to seven-fold increased risk of clotting. Based on that, we made the decision to anticoagulate her. That was the experience that launched my interest in using aggregate electronic health record data for point-of-care decision making.
We just published in Health Affairs last month in the big data issue the concept of a green button. Just as a blue button is both a metaphorical and visual indication of patient’s abilities to download their own data, the idea behind the green button is that in the absence of good peer-reviewed evidence on a clinical decision, that you would be able to use the aggregate data in your electronic health record — or perhaps federated across multiple databases — to generate real-time, personalized comparative effectiveness cohorts, or “patients like mine.”
Imagine if you saw a 55-year-old woman with hypertension, asthma, and of Vietnamese heritage. Recognizing that this lady would not fit well into the American Heart Association guidelines, you could look at the experience of all 50- to 60-year-old Vietnamese women with hypertension and which medications have the most efficacious impact.
This could really change our clinical decision-making and our cost effectiveness and value of care across the United States. But there will have to be some important policy changes as well as technology developments to ensure this happens in a systematic and formalized way.
Kaiser has done interesting things given their huge database and control over all care settings for their patients. The PCORI project generates cohorts across participating health systems. Do you see the use of data going beyond the four walls to make clinical decisions as a trend?
Absolutely. I think a lot of good work on interoperability of databases is occurring. I2B2 is one example. PCORI is another and the PCORnet. Kaiser and Geisinger have been leaders in using their own data sets to make more data-driven decision about what medications they offer on formulary, for example.
But I think ultimately we need to get to a point where rather than go into a group of analysts and researchers to mine data for six to 12 months, we need to enable the clinicians with the right tools to do these queries at the front line of care. That’s really what the green button concept is about.
Do you think that’s a significant argument for virtual affiliations? The six Wisconsin hospitals jut announced plans to work together to share patient information from their Epic systems.
Unquestionably. In fact, we use the Epic electronic health record system at Stanford. One of the things that’s really exciting to us is the amount of data exchange that’s occurring in Northern California. We have such a high rate of data exchange that in the first 10 weeks on the Epic system at the children’s hospital, we connected with over 35 outside institutions for over 30,000 patients.
We know that enables continuum of care for our patients who are receiving primary care elsewhere as well as for the subspecialty care that we offer at Stanford Children’s. But the next step is using that data to provide better analytics for population health. The Wisconsin example that you describe is a great pilot and prototype for what I believe will occur increasingly as we move forward.
I should also mention that one of our clinical informatics fellows, Dr. Downing, is actively working on a project now to look at data exchange in Northern California in a 360-degree view. Most studies to date of healthcare information exchange are focused on what it means to the emergency department that they can get outside data, but in fact, the major use case that we’re seeing is that we offer tertiary care services and a lot of our patients get primary care elsewhere.
We’re really supporting the continuum of care. Being able to look at data that’s sent and received from the perspective of multiple different health systems in Northern California is one of the benefits that this fellowship offers.
Putting on your CMIO hat, what are some other interesting projects that you’re working on?
I also have the opportunity and privilege to lead our analytics and data warehouse team. I believe, as in the green button concept, that the future of leveraging these electronic health records is going to be how we use it not just for the care of an individual patient, but for the care of populations of patients.
We’ve got a number of innovation pilots in our analytics team. My colleague Dr. Jon Palma, who’s also the associate program director for the fellowship, is leading some exciting work in text analytics that’s already benefited our hospital in an operational way. We’re also looking at predictive analytics and forecasting. For example, our census report right now looks at historical trends. Shortly we will be adding the ability to forecast census trends for the next week.
Any final thoughts?
Stanford is accepting applications now for 2015-2017 fellows in clinical informatics. We welcome applications from candidates of any any clinical background.
I would close by saying that we’re at an exciting junction of the field. I believe in the future, as we see more and more physicians involved with health information technology, that this board certification will become a mark of somebody who’s achieved a certain set of core competencies and will be increasingly important across the spectrum of physicians working in these settings.