According to the Congressional Budget Office, health care costs are projected to be upwards of 26 percent of the gross domestic product by 2035. Even now, the United States has the world’s highest per capita health care costs at $8,508per person. These accelerating costs are not sustainable and we have to operate differently to break the trend. This is one of the most exciting times in the health care industry to be in the health care IT. Now, IT is playing a vital role in leveraging data in order to slow the cost trends. As we shift our care model from fee for service to population health, IT must invest in the analytic capabilities to focus on what has happened, what is happening, and what is likely to happen. Successfully addressing this across the continuum of care will ultimately improve the way our caregivers provide care.
Opportunities to Generate Value
The age old adage, ‘you can’t manage what you don’t measure’ rings so true today in health care. We have so much data relative to knowledge–the key is how you leverage that data for hindsight, insight, and foresight. To do this, we need to leverage the data collected through electronic health records (EHR) and a host of other operational systems. Leveraging data scientists to mine structured and unstructured data for insights to predict readmission risk or the onset of sepsis are just a couple of hundreds of use cases that will help us improve our quality, safety, and operations. At Dignity Health, we are creating an intelligent health care ecosystem with Big Data to maximize delivery of our use cases.
Creating an Intelligent Health Care Ecosystem with Big Data
We are building a highly scalable, fault tolerant, open architecturebased Big Data ecosystem using commodity hardware, cloud infrastructure, and open source software to deliver our use cases. Our platform is designed to seamlessly integrate with both internal and external systems, regardless of location. The platform will be equipped with the ability to capture an ever-growing variety of healthcare data (sensor signals, images, text, HL7), process it in near real-time, and store it forever in a cost-effective manner. Our Big Data ecosystem, referred to as Dignity Health Insights, is designed to handle both analytical and computing workloads. This capability not only empowers data scientists to discover insights quickly, but also enables them to build predictive models using the full volume of data instead of traditional sample-based modeling. The bottom line is more accurate models, stronger predictions, and better patient outcomes.
Key Adoption Challenges & Best Practices
While Big Data ecosystems are becoming more common place, many firms struggle with start-up activities and with slow adoption. These challenges include a shortage of big data infrastructure and application development technology skillset, complexities in moving data from traditional systems, and integration with third party products. Firms should consider building a comprehensive big data rollout roadmap that sets guidelines for changes that will be required for human resources (Ex: new job descriptions), application development, infrastructure and support functions.
“Future applications of a Big Data ecosystem will be focused on implementing machine learning algorithms off real-time data feeds where algorithms learn from streaming data and initiate actions”
Dignity Health is planning to train and build Big Data and data science skills among internal leaders who have extensive health care knowledge. Recent advances in the Big Data tools like Apache Hive and Apache PIG have eliminated dependency of high skill Java/ MapReduce talent. Now, organizations can transform resources with basic structured query language skills into Big Data programmers.
Getting data in the right place, at the right time, in the right form, to the right people sounds simple. Establishing a data governance and stewardship function, developing key internal and external alliances, and adopting a use case centric approach to delivery are critical success factors. Building the right group of executives to help prioritize initiatives and lead change is imperative. An EDW needs broad-based business alliances to succeed and don’t run as effectively as an IT-only initiative. For example, our CMIO is both the executive sponsor for our Enterprise Analytics program and the lead of our Data Governance and Stewardship Council to ensure we are always putting the needs of our patients and care givers first.
Our intention is to seamlessly integrate services that clinicians really want into their regular workflow. By delivering analytics at the point of care, physicians are able to make faster, more accurate decisions that improve patient care, accuracy, and quality.
Preparing for the Future
Future applications of a Big Data ecosystem will be focused on implementing machine learning algorithms off real-time data feeds where algorithms learn from streaming data and initiate actions. Driving insights out of sensor data is a trend that has the potential to create leapfrog applications in the healthcare domain. The concept of the “Internet of Things” which will connect devices, systems, and services is another exciting aggregation and analysis opportunity for big data.
Just imagine a world where all our care givers have access to timely and accurate data across the continuum of care for their patients. Basis point improvements in predictive analytics will have a huge impact in improving quality of care and enabling innovation. Big Data is a means to enable our mission to improve health in the communities we serve.