Leveraging Big Data for Health Care

Derrick Schafer, Director, Enterprise Data & Analytics, Sanford Health
384
639
130
Derrick Schafer, Director, Enterprise Data & Analytics, Sanford Health

Derrick Schafer, Director, Enterprise Data & Analytics, Sanford Health

The advancement of the Electronic Medical Record (EMR) in healthcare has opened the door to a robust and diverse amount of data opportunities. Utilizing this Big Data in an effective and meaningful way has the ability to profoundly shape the effectiveness of care a healthcare organization can provide its patient population. Sanford Health is an integrated health system in the upper Midwest encompassing 45 medical centers, 289 clinics and over 28,000 employees that is leveraging Big Data to help accomplish our vision of improving the human condition through exceptional care, innovation and discovery.

  Sanford has the ambitious goal of having 80 percent of descriptive data available via self-service tools within the next year 

Sanford has recognized two critical and intertwined elements to optimize the use of Big Data in health care. The first is a common language through data governance. Although tedious at times, this work is essential in minimizing time intensive data disputes as well as diminishing organizational confusion. Sanford executive leadership has acknowledged the importance of governance and as such has supported the implementation of a Data Governance Committee that includes multi-faceted representation from the organization. This team functions to approve all terms that will be incorporated in reporting and analytics, while relying heavily on data stewards and subject matter experts throughout the organization. These definitions are then technically governed to ensure our analytics philosophy of “one source of truth.” Assessment of data integrity is another component of governance that helps identify areas of improvement in data collection as well as providing full transparency of the analytics being developed.

The second crucial element for Big Data success relates to architecture. Implementing a warehouse strategy that has the ability to bring disparate data sources into one reporting solution can allow for optimization of analytics. Sanford has determined that data virtualization will be the backbone of our data warehouse solution, which will largely eliminate the need to create resource intensive ETL’s into a physical warehouse. A significant advantage our organization has with operationalizing data virtualization for health care information is that our EMR exists on a single build, which is not always the case with large health care organizations. Sanford is developing this warehouse solution agnostic to a specific reporting tool to allow the utmost flexibility going forward. As we bring data together through technology, we lock-down governed terms, definitions, and metrics through views thus removing the need to access the data directly. Ultimately, this diminishes variations in report development and makes the information more accessible for self-service reporting.

As we turn the corner into meaningful and actionable analytics, we see an organizational culture shift taking place. Historically we have existed in a descriptive data environment but it is clear that predictive and prescriptive analytics will be where the most successful health care organizations reside. Sanford has the ambitious goal of having 80 percent of descriptive data available via self-service tools within the next year. Putting governed data at the end-users fingertips and transitioning from a push to a pull mentality for descriptive data is something we see as critical. This move, away from continuous report development, creates the capacity to evolve into analytics that in addition to identifying an issue will also illustrate what is driving that issue as well as guiding our organization on what can be done to positively affect that issue. With the addition of a talented Data Science team, all of this robust EMR information can be used to create validated models through traditional tools while exploring more elusive analytics utilizing machine learning and natural language processing. These analytics are pushing our operational teams to move away from lagging indicators and into more meaningful and actionable leading indicators.

Consider a scenario where instead of identifying a patient population that is considered high risk, you could identify which patients have the potential to become high risk and additionally what specifically could be done proactively to keep those patients from becoming high risk. The affect you could have on the health of your attributed population and in turn the health of the entire community could be remarkable. This is where the shift of descriptive to predictive to prescriptive analytics is vital. An abundance of data exists in this Big Data realm but doing something meaningful, timely, and actionable with the information remains the challenge in our current and future health care environment.

The aspiration for Big Data consumption at Sanford Health is to create an agile data environment while providing governed and actionable data to the right people at the right time. The outcome of this goes well beyond elaborate analytics and fancy dashboards being displayed in board rooms. Leveraging this incredibly diverse and valuable information can and will play a significant role as an instrument in improving the lives of the patient’s we serve.

Read Also

Maintaining Maximum Relevancy for Buyers and Sellers

Maintaining Maximum Relevancy for Buyers and Sellers

Zoher Karu, Vice President and Chief Data Officer, eBay
Building Levies to Manage Data Flood

Building Levies to Manage Data Flood

Adam Bowen, World Wild Lead of Innovation, Delphix
Resolving Disassociated Processing of Real-Time and Historical Data in IoT

Resolving Disassociated Processing of Real-Time and Historical Data in IoT

Konstantin Boudnik, Chief Technologist Bigdata Open Source Fellow, EPAM
Big Data, Small Business

Big Data, Small Business

Matt Laudato, Director of Big Data Analytics, Constant Contact