In 2014, we started our work at Penn Medicine with one simple goal: to make our patients’ lives better, and to create a more efficient health system.

Three years later, Penn is now benefitting from predictive health care, with measurable results and lives saved. State-of-the-art applications have been deployed (some running in real-time), and clinical workflows have been updated to the benefit of all involved.

Penn’s success has depended, in large part, on a deep commitment to collaboration, transparency, and openness.

But there is still much work to be done. Integrating predictive applications into clinical workflows is complex and challenging, as is the constant adaptation required in an ever-changing technology landscape. Add to this mix the general misunderstanding of exactly what data science is and how it can help, coupled with the immense complexities of clinical data…’s hard.

My hope is by sharing our experiences we can enable other health systems to be successful using predictive systems, and help grow the community of predictive healthcare practitioners and data scientists.