Doctors have largely moved on from keeping illegible, handwritten notes about their patients to maintaining electronic health records, thanks in part to new government rules and subsidies. Those records, of course, comprise a vast trove of information that could unlock the potential of data-driven medicine, leading to more effective and efficient health care.
That kind of advancement in medicine is sorely needed -- medical error is, after all, the third leading cause of death in the United States.
But so far, the potential of data-driven medicine has gone largely unrealized because around 80 percent of health care data is unstructured.
The six-year-old San Francisco company Apixio is now tapping into that unstructured data with a cognitive computing platform called Iris, which uses data mining, pattern recognition and natural language processing to make sense of your doctor's notes. With one Iris-based application available already, the company's business has grown quickly in the past year. On Tuesday, the company announced it's raised $19.3 million in Series D venture capital funding to help it advance new products. The funding round is led by SSM Partners, with participation from First Analysis, Bain Capital Ventures and Apixio's largest angel investor.
The global health care cognitive computing market is expected to be worth $5 billion by 2022, according to Grand View Research, and Apixio faces some formidable competitors in the field. Google's AI unit DeepMind has is working with the UK's National Health Service (NHS) to treat patients, while IBM continues to bulk up Watson Health.
Apixio has managed to gain a foothold in the cognitive computing through its existing healthcare customers. The Iris platform has already analyzed data from more than six million patients and "gets smarter with every customer we sign on," Apixio CEO Darren Schulte, MD, said.
Apixio signed on its first clients by initially offering a search tool for electronic records. About three years ago, the company started exploring the way Medicare Advantage requires doctors to document and code conditions (with a model known as Hierarchical Condition Category, or HCC coding) in order to set payment rates. That research led to its first Iris-powered product, the HCC Profiler.
"The value we're providing is that it's more accurate than a human and faster to read a medical chart to understand the conditions, and the severity of that condition," Schulte explained.
With the new round of funding, Apixio aims to use medical charts to not only understand a patient's condition but to also evaluate the quality of care a patient's been receiving. For example, accepted medical guidelines say that patients suffering from diabetes should keep their blood pressure in control or risk suffering complications. Apixio's new tool should be able to quickly scan years of records to assess whether a diabetic patient's blood pressure has been properly controlled.
This kind of data analysis is appealing to health care providers, Schulte explained, as the U.S. health care system moves away from paying for services to paying for outcomes.
"A lot of what's done with respect to medical care is based on a physician's knowledge, what they learned 20 years ago in residency, based on studies that might not have been conducted with rock solid methods," he said. "It's not as data-driven as it should be."
Ultimately, Schulte said, "We want the science of medicine, not the art of medicine."