Zeeshan Syed, PhD, is the Founder of Health at Scale Corporation and a Clinical Associate Professor and Director at Stanford Medicine. Before joining Stanford in 2016, Syed was an Associate Professor with Tenure in Computer Science and Engineering at the University of Michigan, where he was a Principal Investigator for the Artificial Intelligence Laboratory and led the Computational Biomarker Discovery and Clinical Inference Group. Syed received SB and MEng degrees in Electrical Engineering and Computer Science at MIT, and a PhD through a joint program between MIT’s School of Engineering and Harvard Medical School in Computer Science and Biomedical Engineering. He is the recipient of multiple national awards for his scholarship activities, including the prestigious CAREER award from the National Science Foundation. Syed is also actively engaged with the healthcare analytics industry, having been part of the core early-stage team for Google[X] Life Sciences (now Verily) and more recently as the Founder of Health at Scale Corporation.
What most excites you about the work you’re doing in machine learning? If you are successful, what changes in healthcare?
My work over the years has focused on the design and application of advanced healthcare-specialized machine learning and artificial intelligence technologies for population health, precision medicine and high-value care. These efforts are unified by one question: How can we leverage healthcare data that are generated from routine patient care ― including electronic health records (EHRs) from providers and claims data from payers ― along with ambulatory data, from wearables and other consumer devices (such as smartphones or smartwatches), to improve the treatment and prevention of major diseases?
There’s often talk about the “iron triangle” of healthcare and how health systems are challenged to achieve care that is simultaneously low-cost, high-quality and widely accessible. It is possible to improve one or two of these factors, but it often comes at the expense of the third. This is where machine learning and artificial intelligence can make a huge difference ― through proactive and personalized care decisions that can improve outcomes by intervening early and effectively. If the efforts driven by various stakeholders ― clinicians, hospital administrators, payers and researchers ― at the intersection of computational sciences and healthcare delivery are successful, we have a real opportunity to break through the iron-triangle paradigm and enable sustainably better care across the care continuum.
While I’ve greatly enjoyed my work in academia, I strongly believe that breaking through healthcare’s iron triangle requires innovations in the industry that can be developed and disseminated as sustainable and practical solutions for major healthcare challenges. At Health at Scale, we have substantially improved outcomes and utilization for large health organizations managing complex populations across complex care networks — reducing adverse patient outcomes and costs of care by a factor of 10 times relative to conventional analytics. These results are exciting, and we’re only just getting started in terms of tapping the true potential of machine learning and artificial intelligence.
How would you describe the current state of healthcare innovation? What big challenges need to be addressed? How could new technology make a difference?
The use of machine learning and artificial intelligence is still in its early stages, but we’re beginning to see examples of the transformational impact that these technologies can deliver. Large digital repositories of longitudinal health data are emerging that comprise a rich diversity of the kinds of data elements and the patient populations represented. We are also in the midst of an increasing awareness of the need to exploit these data sets to drive value for providers, payers and patients.
These developments are inspired by the successes of computational analytics in other data-intensive disciplines beyond medicine. Significant progress has been made in these domains, but when it comes to healthcare analytics, progress has somewhat lagged. In large part, this effect can be attributed to the methodologies for deriving insights from healthcare data sets remaining grounded in techniques that are dated and general-purpose. The focus has traditionally been on repurposing methods from other domains rather than designing innovative solutions specialized from the ground up for the nuanced needs of healthcare. This is a major challenge that needs to be addressed.
Looking back at your own career ― what particulars shaped your own development? What promoted or enabled your own productivity?
I’ve benefitted from having an exceptional set of collaborators throughout my career. I received my initial academic training through the MIT EECS and Harvard-MIT HST programs, and was extremely lucky to work with a group of multidisciplinary faculty and researchers who emphasized application impact alongside theoretical advances.
If I look back, I think that combination of being able to formally train in both machine learning and medicine; have the license to explore innovations outside traditional research areas in both disciplines with a focus on maximizing healthcare impact; and work with an extraordinary set of individuals was critically important in shaping my career.
I’m a huge proponent for continuing to develop training programs for careers in healthcare analytics that offer exposure to formal training in both EECS and medicine, and that further emphasize the ability to creatively define research directions that are aligned with practical impact.