A week or two ago, I got to correspond with Danielle Brooks of Disruptive Women in Health Care about the work I do here at O’Reilly. The following interview is reprinted here with their kind permission.
Tell us about your work. What drew you to the area?
I have mostly worked as a book editor, until just a year or two ago. I was working on books about databases, machine learning, visualization, and other relevant topics when O’Reilly launched its Strata conference on data science, and so I became involved in that conference. But as Strata took off, it became apparent to us that certain communities — and certain types of data — were special. Health care is one of those areas: the insights that data analysis can give us about ourselves and the things that ail us are enormous, but the risks of over-sharing and the resulting constraints such as HIPAA also present very real challenges.
In 2012, O’Reilly decided to launch a new edition of its data science conference to focus on health care, and that’s how Strata Rx was born. I was asked to become its Program Chair, along with Colin Hill, CEO of GNS Health care, and so I have spent that last 18 months learning everything I can about the (very complicated!) health care industry. Colin and I are great partners because of the complimentary backgrounds we bring together — Colin from the health care industry side and myself from the technology side. Ultimately, that’s what Strata Rx aims to do, too: we hope that by bringing together professionals from all parts of the industry (payers, providers, researchers, analysts, advocates, developers, investors, and caregivers, just to name a few) we can begin to solve some of the large and complex problems facing us in this area.
You work primarily in Data and Analytics. Tell us how you apply your work to the space of health care.
As an Editor and Program Chair, I work primarily in the business of sharing knowledge and ideas, as well as the context for those ideas. Health care faces a number of significant challenges, from the staggering costs (about $2.6 trillion every year in the United States) to the widespread occurrence of chronic conditions such as heart disease and diabetes to the highly variable responses of different individuals to a given treatment. I’m interested in helping to connect people with a deep knowledge of things like metrics, statistics, and interaction design to others with a deep knowledge of genomics, epidemiology, drug development, and patient advocacy. Data science and analytics are already making a huge difference in other fields (such as marketing, finance, and retail, just to name a few), and health care is similarly ripe for innovation and advancement.
What is health care doing right in data and analytics?
It’s not really possible to speak of the industry in monolithic terms. Just as in any discipline, there are some people doing cutting-edge work, and many others lagging behind. But there are some great examples of where progress is made. Some researchers and companies are using data and analytics to create targeted therapies for specific gene mutations. Some patient communities are sharing their own information to help each other out and identify patterns. The Quantified Self movement uses wearable devices to monitor and change their own behaviors. Doctors and hospitals are using electronic medical records to centralize information and reduce errors, and programs like the VA’s Blue Button initiative and online patient portals are helping give access to those electronic records back to patients themselves.
The real advantages will come as these innovations start to cross boundaries between groups of professionals. When you can share the information from your wearable device with your doctor, who can upload that into an electronic record that works with the systems your specialists are using, and they can compare that data against the things your genome suggests you might be at elevated risk for and consider the interventions that are most likely to work for you as an individual — then we’ll really be onto something.
What is the industry doing wrong?
Despite the privacy and sharing constraints of legislation like HIPAA, it seems to me that some of the most serious challenges preventing health care professionals from making more use of data and analytics are cultural.
On the patient side, there is a generational divide between people who are used to sharing lots of personal information and people who have been trained to keep everything to themselves. On the provider side, there is an ingrained way of thinking about how to make good decisions (with an over-reliance on “gut instinct” and subjective experience). On the research side, the practice of publishing only successful studies — some with dubious definitions of success — means that failed research is never shared, and we lose a lot of available context for the studies that are published, misleading us all about the significance of various findings. In the entire system, incentives are misaligned so that the care and health of the patient isn’t actually the primary concern.
How can the health care industry begin to think about data and analytics in a different way to make it more useful in the field?
The biggest difference that data and analytics can make in health care is increasing the level of granularity at which we can understand ourselves and make decisions. For example, right now most people have their blood pressure and heart rate measured once a year — and that’s only if they actually show up to an annual physical. Wearable devices can now measure and report those statistics multiple times per day. That’s a huge difference in how much information we can use to paint a detailed picture of our health. Another example would be genome sequencing, which is becoming faster and cheaper all the time. It can potentially tell us as individuals which conditions we may be at risk for, and which treatments we’re likely to respond to, and allow providers to target interventions more precisely (known as precision/personalized medicine).
Another significant opportunity I see is to help us measure the interventions and processes that work, so we can standardize best practices. Right now, health care providers mostly rely on a combination of gut feeling and subjective experience. But by carefully tracking and assessing a much broader experience base, we can develop checklists (like the ones that already exist for airline pilots and other professionals who hold lives in their hands). These checklists and standards are already being developed around goals such as preventing the spread of sepsis in hospitals, but aren’t widely adhered to yet, and could be useful for so many other health care goals.
Finally, better decision-making and personalized medicine will help us direct our financial resources to where they can do the most good. There is a lot of monetary waste right now, both at the individual level (excessive testing and treatments that don’t work) and at the systemic level: a small number of individuals who fail to manage chronic conditions end up back in the hospital over and over again, racking up costs. Reducing readmission rates by identifying and focusing on at-risk individuals can reduce costs to the system. And since prevention and disease management are as much in the interest of insurance companies as patients, identifying which treatments work for which individuals can help payers make customized reimbursement decisions to pay for the things that do the most good (known as value-based reimbursement).
Where do you see the use of data and analytics in the future of health care in the next five years?
Five years is typically a long horizon in the technology space, but in health care I fear it is barely a wink. Many of the industry professionals I’ve spoken with agree that we’d be better off thinking in terms of generations. Some of the cultural challenges to innovation I mentioned are so deeply ingrained that many people in the industry believe that our best hope is to sit back and wait for a changing of the guard (I remain hopeful that we won’t have to, but I haven’t heard any ideas about how to hasten the necessary cultural shifts).
In the meantime, I do think we’ll see some significant leaps forward in the next few years. Manufacturing techniques such as 3D printing will continue to improve: we’re already using this method to create custom implants and prostheses out of titanium and other metals, but we are also making rapid strides with the ability to print tissue and organic matter; the holy grail is a 3D printed kidney or similar organ for transplant. We’ll also see huge gains in the ability to make use of what’s called unstructured data — such as information that exists in text instead of in a spreadsheet — think of the huge body of insurance claims or hand-written notes on doctors forms. Our ability to make sense of this information in aggregate and to place it in context will mean lots of learning, and there is even bigger potential when we combine it with other available data.
There is so much we still don’t understand about ourselves, our bodies, our diseases, and our environments. But the promise of data and analytics is that it can help us detect patterns and increase the precision of our insights and our actions. It’s a very exciting time to be in the industry, and I’m especially looking forward to the discussions at this month’s Strata Rx conference in Boston. I’m glad to have the privilege of being involved.
Editor’s note: This post contains material previously posted on the Disruptive Women in Health Care blog.