Skip to content

Center for Clinical Data Science Lays Groundwork for Success in Healthcare AI

Artificial Intelligence (AI), machine learning, and deep learning are the hottest categories in healthcare right now, but the truth is these concepts have been around for a while, says Trung Do, Vice President of Business Development for Partners Innovation. “The fundamental goal is in extracting meaningful insights from massive amounts of data and making them actionable.”

What’s new is the tremendous amount of computing power that is now available to develop, test, validate, and translate applications (e.g. algorithms and models), at scale, into clinical use, which has the potential to revolutionize the way healthcare is delivered.

Leading the Partners HealthCare AI effort is the new Center for Clinical Data Science (CCDS), a collaborative effort between MGH and BWH that is directed by Mark Michalski, MD, executive director of CCDS.

Founded in 2016, the Center is designed to support Partners’ investigators as they develop AI capabilities and to create and foster a community of like-minded people who are working to advance healthcare through AI and machine learning.

“In the past, if I wanted to find a new way to identify patients with pancreatic disease before they show symptoms, I would come up with a hypothesis and test it in new patients,” explains Keith Dreyer, DO, PhD, Chief Data Science Officer at Partners HealthCare. “But now we’re asking the machine to look into the archives for the answers.”

“I can take millions of past images that show healthy and diseased pancreases and, along with additional medical data on the patients, load them into software algorithms. With that information, the algorithms can learn to recognize those with pancreatic disease, and learn to identify markers that are precursors to their condition. It can then look at current patients’ medical records and pinpoint patients who might develop pancreatic disease and need diagnostic evaluation.”

In one recent project, for example, Dr. Michalski, used hundreds of thousands of medical images to train a computer system to detect strokes, measure tumors and look for traumatic injuries and fractures. The early results are encouraging. “The technology really does work,” he says.

AI’s potential to improve health care delivery is not just limited to imaging, Dreyer notes. “We’re also looking at genetic and genomic data, laboratory data, and other data available inside the electronic medical record and beyond.”

Partners leadership believes AI will play a crucial role in advancing research and improving the delivery of health care in the years to come, Dryer says. “We see this as a new discipline that will be part of every clinical program, every operational program within our system, to improve the way in which we deliver care and the way in which we manage operations.”

The CCDS has initiated a series of strategic relationships that are designed to support these efforts. This includes building computing infrastructure and development capabilities through collaborations with technology companies such as NVIDIA and large international healthcare companies such as General Electric.

“If we do it right, we think there are commercial opportunities for us to take these technologies, learnings and capabilities outside the walls of our system and create real value broadly for clinicians and patients,” Do says.

While the concept of AI and machine learning may not be entirely new, there are good reasons why the field has gained momentum in the past few years, Dryer explains.

One key factor is that the computational power needed to process and draw insights from large data sets is less expensive and more readily available. For many years, the only places that could afford large-scale computational platforms were big companies such as Amazon and Google.

But recently, companies such as NVIDIA have made supercomputers affordable to anyone, and a lot of that is driven by technological advancements in other industries, such as video gaming.

This has enabled Partners HealthCare to acquire a tremendous amount of supercomputing power. The CCDS announced its first agreement with NVIDA in April of 2016, and over the past year and half it has built one of the most significant supercomputing infrastructures of any academic medical center.

The next step is to combine this computing power with clinical expertise, Dryer says. “You can have really talented data scientists and engineers, but it’s difficult for them to develop health care applications and solutions without a deep understanding of the clinical use case. You need access to clinicians who understand the clinical challenges and who can integrate these applications into the workflow to optimize their effectiveness and adoption by clinicians.”

Another critical element is access to clinical data, which resides within health care systems such as Partners.

“Two years ago, we started to see a lot of companies coming to us and asking for this data,” Do says. “Thanks to the foresight of strategic leaders such as Dr. Dreyer and James Brink, MD, Radiologist in Chief at MGH, who understood the value of our system asset, we decided to make an investment in ourselves to keep the program in-house.”

And that decision is paying off. In May 2017, the CCDS announced a significant 10-year strategic collaboration with General Electric to build a learning factory—a platform with a set of capabilities and tools to support new clinical applications designed to improve patient care.

“The second thing we are doing with the GE relationship is building specific clinical applications, meaning apps that will be sitting on top of this platform or at least integrated into the platform to help clinicians manage and deliver more efficient and optimal care,” Dryer adds.

“Our view of AI is to augment what the clinician is already doing, to make them much more efficient, more accurate, and provide higher-quality discoveries, diagnostics and treatment. For us, it’s about making the clinician better, not obsolete. That’s our strategy.”

Join Us at Partners HealthCare’s 2018 World Medical Innovation Forum, which will focus on the advancements and opportunities of Artificial Intelligence. The Forum will be held on April 23-25, 2018. For more information, www.worldmedicalinnovation.org