How far would you go to get access to your health data if you had a life-threatening condition? Would you enroll in a graduate medical program to get access to your own tumor as a researcher? Steven Keating did.
In May I did some reporting from the Health Datapalooza conference in Washington, D.C., and heard from many patients how difficult they find it to access and share their data for research purposes. While writing further about this topic for another publication recently I got a chance to interview someone whose dogged pursuit of his own health data has made him an eloquent advocate for patient data access and control.
In August 2014, Steven Keating, a Ph.D. candidate in the Massachusetts Institute of Technology (MIT) Media Lab and Mechanical Engineering, was diagnosed with a baseball-sized brain tumor. In trying to learn more about his condition and figure out which course of treatment was best and which clinical trials he might participate in, Keating ran into all types of interoperability and regulatory roadblocks. His story is now part of a documentary film called the "The Open Patient," part of Red Hat Films’ Open Source Stories series. The film tells the story of how open healthcare data is changing the patient/doctor relationship and the healthcare industry as a whole. It is scheduled to premier at the end of June and also explores OpenNotes, an initiative focused on giving patients access to healthcare records between them and their doctors.
Keating ended up being treated in four different hospitals in the Boston area, and while he was very pleased with the clinicians and the treatment he received, he was shocked that the hospitals were still faxing all the information back and forth.
Because patients are legally allowed to request a summary of their health data from hospitals, Keating requested his. But what we got was an 8-page report with only one mention of a neural malignancy. “The rest was all random gibberish about blood levels,” he said. “It didn’t list anything about my chemo or proton radiation.” He was being treated at different hospitals for chemo, proton radiation and surgery, and the hospital EHRs were not speaking to each other. “I would do an MRI at Mount Auburn, another hospital, and if Brigham & Women’s wanted to access it, they sent me in my car with a CD to pick it up and drive back,” he said incredulously. “Even though a lot of these hospitals are using supposedly compliant systems, they are using different installed versions of Epic, so they can’t talk to each other. All these problems I came across are very logistically silly. How come we can use online banking to track all our financial information, but for hospital data, which our lives are depending on, we are still faxing things?”
Keating said he thinks hospitals should have APIs (application programming interfaces), so that third-party developers could solve some of these problems. “If you look at the iPhone, the Apple Store has over 1.5 million apps available. Ninety percent of those apps are free, and those apps have all been vetted by Apple to not do terrible things to your phone,” he said. “So what if your hospital were an app store that could verify programs and verify third-party developers who do useful things and allow you to interact with your data?”
Keating’s story gets more involved when after surgery he sought tumor- sequencing information and was turned down. “I donated part of my brain tumor to research, and there is research being done between my hospital and my university, MIT. I donated part of my tumor because they were very interested in doing genetic sequencing of it because I had this specific mutation,” he said. “I thought this is great; I will donate part of the tissue to science, and I can get my own genome back for my tumor, which can help me make decisions in the future about clinical trials and things like that.” But after the tumor was sequenced, he heard back from the university legal team that he could not have access because the tumor was sequenced on a machine that was not CLIA-certified. Basically CLIA (Clinical Laboratory Improvement Amendments) regulation restricts any data generated on a non-CLIA machine from getting back to patients.
“What that means is that they have to run a test to make sure the machine is reliable enough and up-to-date enough that they trust it enough to give it back to the patient, so the patient can’t harm themselves,” he explained. “Of course, the sequencers at MIT were top of the line, but they weren’t CLIA-certified because that costs thousands of dollars to do. What that meant was that everyone could see my genome except me. My doctors could see it, researchers at my university — people right around me — could, but I couldn’t. And I am the one who gave them my own brain! Why can’t I accept the liability of understanding it was done on this machine? I want to share this data with the world. I want others to be able to use this data to perhaps make a discovery.”
Then Keating got the idea that perhaps he could get access to a small piece of his own tumor. “But the answer I got was no, because once you go through surgery, whatever they take out is not yours anymore,” he said. So at MIT he transferred into a medical program called health science technology and enrolled in a pathology class. “I knew this pathology class was linked to Brigham & Women’s Hospital, so I made my final project to do pathology on my own tumor. Because I was now a scientist I was able to ask for research access to my own tumor, and that is how I got it,” he said. “It was a total back-asswards way of doing it, but I had to become a medical student to get access to my own tumor.”
After that experience, Keating began writing policy comments and giving talks about putting patients at the center of research. For instance, he submitted several comments to the draft standards for Meaningful Use stage 3 regarding the requirements around APIs, which CMS eventually included in the regulations.
He has put all his health information online (stevenkeating.info) and contributed it to Open Humans and the Personal Genome Project. When a story ran about him in the New York Times, he received thousands of e-mails from doctors, patients and researchers expressing support or asking questions. “It is not just patients that have trouble accessing data, it is doctors, too,” he said. “For example, I recently moved to California and am now under the care of a Stanford neurologist. When I moved out here, only about 75 percent of my record made it,” he said. “A fair amount of the pathology records didn't get translated. Doctors are frustrated by this just as much as the patients. Why can’t we have dynamic consent so the patient can actually choose what they want to do with their data?”