What Can Big Data Do for Doctors?

Q: What is Modernizing Medicine?

I’m a dermatologist in private practice in south Florida. My practice was basically using paper for documentation. I knew there had to be a better way. Being a Yale SOM alum, my intent was always to do something entrepreneurial, but in medicine it’s hard to figure out the right time.

When Dan Cane, the co-founder of a company called Blackboard, came in as a patient, I realized that a tremendous opportunity to fulfill that dream of starting a business had come. We created Modernizing Medicine, a healthcare software company. Our initial value proposition was to cut physician documentation time from 5-10 minutes to 1-2 minutes. Our long-term mission was to transform the healthcare industry by capturing structured data.

In six months, working out of a 400-square-foot office in Boynton Beach, we built a prototype of EMA—electronic medical assistant—for dermatology. While Dan built the platform, we realized immediately that I could not teach Dan how to become a physician. Instead, he taught me how to code software. With the advantage of understanding physician workflow, I had complete freedom to design the relationships between diseases, their treatments, and their outcomes. It’s like the product went to medical school and did a dermatology residency. A lot of the other EHR [electronic health records] platforms out there really start from scratch and put the burden of customization on the customer. But doctors don’t have time to educate computer systems on how to practice medicine.

To understand how EMA works, think of a touch screen with a picture of the body on it. You select a diagnosis and put it on the body. A 45-second interaction prints out your notes, your bill, and your prescriptions so that doctors can just focus on treating patients, not paperwork.

We launched EMA in 2010. Over the next four years, we grew to 27% of the U.S. dermatology market. Today, EMA is in eight specialties.

Q: Why choose to start with a specialty?

At the time, the American Recovery and Reinvestment Act offered physicians up to $44,000 to adopt technology. That created a mad gold rush. Hundreds and hundreds of records companies sprouted up out of nothing.

A lot of our competitors looked at the market, saw 800,000 physicians, and tried to be everything to everyone. There are a lot of word-processor-like electronic medical records that turn physicians into data-entry specialists—typing paragraphs and paragraphs even though it’s the same thing over and over again. The issue was and is that those systems don’t save physicians’ time. That leads to reluctance and pushback.

We went with the opposite approach. Each specialty has different needs. By focusing one medical specialty at a time, we can anticipate 90% of what each specialist does and provide the medical content for them as a starting point.

On top of that, no two physicians practice medicine the same way, so we built in an adaptive learning engine that understands their suture of choice and the anesthesia they use, and “sticks” their preferences as the new default, so it further customizes with use. By getting the workflow right, you save them time.

EMA works just as fast as an assembly line at Chipotle. If you have a limited menu, and you have all of the ingredients in front of you, you can deliver a high-quality product and experience in a very short period of time. Not to say that medicine is like the fast food industry, but if you make documentation fast, doctors will love your product. And that’s exactly what we did.

We’re repeating the playbook in other surgical subspecialties including ophthalmology, orthopedics, plastic surgery, otolaryngology, gastroenterology, urology, and rheumatology. Today, we have 4,800 providers using our software. People use it because the user experience is great. It’s designed for touch, using native iPad, iPhone, and Android applications. You download the medical record system from iTunes, and it’s completely HIPAA compliant. There’s no medical information on the iPad itself—all of the data is stored in the cloud.

Q: How is the data used?

Core to the mission of Modernizing Medicine, EMA captures structured data, and delivers the de-identified aggregate data back to the physician at the point of care to make better medical decisions. Investment bankers have access to information about whether their investments are going up or down. They can base investment decisions on how companies are performing using financial data and KPIs. Even though we are taught the scientific method in medical school, physicians haven’t had anything comparable because our medical data is trapped in the silos of paper charts and antiquated systems.

The right ideas to improve healthcare are out there: patient engagement tools, electronic medical record platforms, and data mining. But the industry is falling short by not collecting that information as structured data at the point of care from the physicians themselves.

Too often medicine still collects narratives that go nowhere. The patient comes in and tells me a story. I record the specific symptoms of the patient, the exam findings, the dose of the prescription, but even using an electronic health record, it stops there. The data isn’t structured, and it doesn’t go anywhere.

Modernizing Medicine’s goal is to transform healthcare by giving physicians a feedback loop. In the last couple of years, we added static global assessments of every disease in dermatology to the product. These are the same tools that pharmaceutical companies use to approve new drug therapies. Instead of clinicians expressing their patient’s conditions as better or worse, which are relative terms, doctors using EMA now rate outcomes on clinically validated scales from 0 to 5, for example. Because our platform is cloud-based, we’re able to de-identify and aggregate that data so that physicians can see what works and what doesn’t for their patients as well as learn how other physicians are treating similar diseases throughout the country.

Medically, that’s valuable with common diseases, but let’s say I see a patient with a rare disease, such as pemphigus vulgaris, which is an autoimmune blistering disease of the skin that if untreated is almost 100% fatal. Since I might only see it once in 20 years, I want to know second-line and third-line options if my first-line treatment isn’t working. I can find out the treatments that every dermatologist in the United States using our system has used for that condition. That could be over 900 patients. This could be more than what’s written up in published case studies. That’s the benefit of the cloud.

From a policy perspective, if you’re collecting structured data and using validated static global assessments of disease, then you can extrapolate and make informed decisions about which drugs are effective in the real world. This in turn can help physicians get their treatments and surgeries authorized from insurers and reduce bureaucracy.

Q: Could you explain the company’s partnership with IBM’s Watson?

Our initial products use structured data. IBM’s Watson has a cognitive computing capability that will also let us leverage unstructured data to its fullest potential. That’s a huge opportunity in medicine. Cognitive computing makes it possible to get evidence-based answers that you can apply to any patient at the point of care.

Physicians may see 40 to 50 patients a day. If we’re lucky, we spend maybe 5 to 10 minutes on each patient. At the same time, the knowledge base is changing every day. Studies come out that contradict studies from the previous month; keeping up is like drinking from a fire hose. There’s no way any human brain can do all that.

So we’re one of three healthcare companies that initially partnered with the IBM-Watson ecosystem. We built an application called schEMA, a powered-by-IBM Watson application, and we are negotiating with content partners from distinguished medical journals to feed articles into Watson so doctors can ask nuanced questions and get answers in seconds.

Cognitive computing is different than search. With search, you have to keep searching to get the answer you need. This is time consuming and less practical. With cognitive, physicians can ask very complicated questions and get very direct answers from peer-reviewed sources with confidence intervals. And you can ask follow-up questions based on those answers.

If I’m seeing a patient who has psoriasis and reactions to certain medications, I’m going say something like, “Show me the evidence for a head-to-head comparison for the only two drugs that this patient can take given the fact that they have these other diseases,” and I get an answer.

What we’re really doing is cutting down the time required for doctors to apply evidence-based medicine to all of their patients. For me to go to a website, locate, download, and read an article may take 15 minutes. And, at the end of that, I may not even get my question answered. Physicians can’t do that for 40 patients a day; they have to run a business.

In contrast, schEMA acts like a research assistant and gathers the evidence I need in a second, not minutes.

To be clear, schEMA isn’t making the decision; doctors make the decisions. But schEMA is providing access to the latest evidence from our content partners. And then physicians can put that evidence into their note to justify the decision.

Q: How far along is the partnership?

We’ve built a prototype around three diseases: psoriasis, melanoma, and atopic dermatitis. That debuted at the American Academy of Dermatology annual meeting in March. We plan to launch a dermatology product by early 2015 and expand into other subspecialties from there.

Q: Bigger picture, it seems like healthcare IT has been on the verge of transforming medicine for years. Where do you see IT and healthcare actually coming together to deliver on that promise and where has it fallen short?

The United States is the undisputed healthcare leader in terms of spend, but not in terms of how long or how well we’re living. There’s a ton spent on chronic diseases without any improvement in outcomes. An important reason for that is our system pays doctors more to do more.

Big picture, everybody wants to shift towards pay for performance. To do that well, big data needs to measure the outcomes on a per-disease basis. Once we do that, we could justify prescribing something that may be initially a little bit more expensive, but cuts down return visits to the physician or hospital admissions. Or we may be able to justify a more expensive surgery if the complication rates are lower.

To get there, we need to measure outcomes in an objective way, and we need to make sure the tools we’re using are validated. I don’t think big data is the answer for everything, but certainly it is much cheaper than doing a randomized controlled clinical trial on every disease. We can’t afford to do that. Randomized control trials are the gold standard for developing new medications, but I think you want to see some kind of hybrid of studies and analysis of real-world data. EHRs that are designed to capture structured data from the outset are well positioned to lead real-world effectiveness research and transform the healthcare industry.

Interview conducted and edited by Ted O'Callahan.