Jawad Khan, Director, Data Services & Knowledge Management Information Services Division, Rush University Medical Center
Healthcare is in the middle of a major digital transformation. It is allowing patients to generate, access, and manage their healthcare information from anywhere. Technology is playing a key role in enabling and even accelerating the change. For newer generations of patients and providers, clinical care and technology are inseparable. Data is the center of gravity of this transformation.
Previously, Electronic Medical Records (EMR) were the only source of healthcare data. Now in the era of modern healthcare, solutions must provide patient-centric care that addresses all aspects of a patient’s journey. Institutions must collect data from multiple sources like wearable devices, social media, bedside devices (within or outside of healthcare facilities), and genetic labs. Data from these sources is voluminous, it has veracity, and it has velocity; it is classic big data. Big data problems may be difficult, but they are extremely important and must be solved. Institutions that harness the power of data from such disparate sources will be able to embrace the digital transformation. These institutions will be successful in providing innovative, top quality, and cost-efficient healthcare. Institutions who continue to use past methods will fall behind, and may even become extinct.
What needs to be done to use data effectively?
Building a robust platform that can grow with data is the first step to the healthcare transformation. Institutions must look for platforms that are beyond the walls of traditional data centers. Virtualization and even service-oriented approaches to infrastructure (e.g Infrastructure as a service) are no longer sufficient. Entire platforms including storage, compute, ETL tools, databases and data warehouses need to be provisioned as services.
Big data problems may be difficult, but they are extremely important and must be solved
The salient features in this approach are its ability to expand, interface, and easy-fast deployment. These features make the platform as a service ideal for next-generation solutions. Essentially, platform as service solutions are necessary and fundamental to provide data as a service.
Moving data is messy, especially big data. Applications, analytics, and visualization tools must be decoupled from data and only consume data using “just in time” approach. Whether institutions develop data solutions in-house or leverage third party products, they must avoid moving data sets. It is typical for institutions to develop extracts and ship data using file-based transports. Such solutions will not scale to consume and process big data. All design and procurement considerations must now address a fundamental aspect – consume data as a service.
What are the challenges?
Service-based architecture can become expensive. This is true for institutions that have to manage data centers that host monolithic large applications and, at the same time, subscribe to service-oriented platforms. They will have to continue to support existing data centers and pay for service based platforms. This is perceived as an added cost when these applications can be deployed on existing infrastructure. It is easy for institutions to lose sight with this quandary and hurt transformation and innovation.
Institutions will have to overcome cultural biases. Teams must be transformed to realize the power of data, cliché, but “data is the new oil” (Clive Humby). In order to improve healthcare, Data must become the focal point of technology build to remove friction for people and systems to access the data. Existing solutions have erected rigid boundaries around data and this is a prevailing problem in many healthcare institutions.
Most industries have moved forward with the advent of Semantic Web (Web 3.0) by embracing the power of data. Healthcare, in general, is lagging behind with the exception of a few institutions like Rush. We are bending the curve in the adoption of the digital transformation. We are now able to predict cost and understand care gaps in geographies of interest and intervene using data insights. As next steps, we want to further improve on these models and automate clinical interventions to reach out to patients proactively. This automation can be achieved by encapsulating patient-provider interactions in apps and push notifications. Such interventions, we believe, will help in closing care gaps and save cost for the patients and healthcare institutions and data will drive this change.