Health IT and Electronic Health Activate your FREE membership today |  Log-in

Community Blog

Jul 26 2012   12:44PM GMT

BI and enterprise data management: Making data useful at the enterprise level

Posted by: Jenny Laurello
BI, business analytics, business analytics strategy, business intelligence, centralized warehouse, data, data integration, data warehouse, Natural language processing, NLP

Guest post by: Lisa Khorey, Vice President, Enterprise Systems and Data Management, University of Pittsburgh Medical Center

Healthcare IT has a problem, a big problem.  The realization of a digital environment, long overdue in an industry plagued by inefficiencies and high costs, is closer than ever, but the ability to produce insight from that data is a challenge and, in most cases, a chore.  IT organizations are now drowning due to a giant wave of increasing demand from the business to inform operational changes as required by payment reform, ICD10, meaningful use, value based purchasing, changing care models, and increased focus on quality.

Viewed individually, any one of these external forces creates an increase in requests for useful data and business intelligence assets. Combined, these initiatives and mandates create chaos in IT organizations that are already resource constrained under the burden of electronic health record (EHR) implementations, necessary upgrades, and a shift toward mobile platforms and cloud computing.

Further, there’s an unprecedented data explosion. Thanks to the addition of unstructured data released from new natural language processing (NLP) technologies, applications and tools that can help collect and express patient preferences, and rich genetic information now available thanks to the decreasing cost of mapping the human genome, the potential for data mining and new insights is possible more than ever before.  These capabilities have created an onslaught of Big Data webinars, sales opportunities, and industry buzz that IT leaders are quickly trying to wade through.

Organizations that desire a competitive advantage will be the first to seek and underwrite an analytics strategy, but all healthcare organizations will need one to survive. The size and scale of the strategy may vary, but unlike the electronic medical record (EMR) and business systems implementations of the past, strategic business analytics for healthcare organizations will require an IT solution that builds a virtual “cloud” of untethered data within and beyond the walls of the organization, above the vendor products, and driven by the strategic objectives of the healthcare system.

Here are three steps to consider as you sort out your business analytics strategy:

1. Inventory the strategic assets already embedded in your IT organization. A quick look at the databases and data warehouses already present in your organization may surprise you. Even in the absence of a broad strategic analytics plan, analytics absolutely exist. The trick is to inventory the systems and reporting repositories and map them against an ideal future state that is simplified and organized by categories of data versus function or vendor. Assume all data in your enterprise is electronic and begin to create a path from the source acquisition points to the center, allowing your design to be tiered. The best models balance local analytics with a centralized warehouse, plus an aligned path for movement of data, and recognition and acceptance of some redundancy.

Already have a well-functioning data integration layer? Use it to drive your data loading process. Already mapping data elements to normalized industry standard code sets? Include it. Already own an industry-standard data model? Build your data governance program over it.

The analytics strategy is an umbrella over these assets, additive, incremental, and purpose-built to add value versus rip and replace. Rationalization of duplicate data stores is step two, not one.

2. Envision the future intersection of data among consumers, caregivers, and the community. The analytics world you live in today is likely finance-driven, clinical-event-starved, and slow, slow, slow. The world you will create for tomorrow includes items of interest from outside the organization, the longitudinal patient record (versus the episodic one), and actual versus average cost of care. Plus it operates real time.

This is a major shift. To accomplish it, select technology tools to facilitate storage of discrete and unstructured data as part of your warehouse architecture and be sure that your data integration plan includes a path back from the warehouse to the point of service. Real time is not only an indicator of your ability to consume data into the warehouse but also implies that you have built the ability to monitor events, have rules to act on them, and can push information to parties of interest in their workflow or via their mobile device.  This is the value proposition of the entire architecture and the true capability you should aim to create.

3. Understand the mission of your organization and make it the “north star” for your analytic purpose. Business analytics is the evolution of a digital ecosystem from transaction-based data collection to action-based information generation. The key to driving success into your BA program is a crystal clear understanding of the actions your operations wish to take, and the only way to direct that action is to ask your organization to describe its larger goals.

Depending on the audience and their level, you will hear different expressions, but a careful listener will recognize that themes begin to emerge.  If you can engage your CEO, do so. Make every effort to understand the direction your organization is moving and design and anchor your analytics program to support those initiatives. Leave performance and monitoring to the local analytics; your enterprise business analytics data stores should serve only the strategic mission.

The only exception to the strategic north star as a driver is the compliance and regulatory needs of your healthcare organization, which should be served from this data warehouse.  Nothing puts more strain on survival than failure to keep a close eye on these measures.

Enterprise analytics projects are notoriously complex, expensive, and, in some cases, major disappointments to the organizations that fund them. But they don’t have to be, and in healthcare, we can’t afford failure.  Define your architecture, inventory your data sources, engage your executives, and get ready. Like it or not, the next generation of IT is here, and it is highly integrated, mobile, and instantly available.  The data problem represents the biggest opportunity for IT to enable better healthcare delivery, and that’s a wave we should all ride.

Lisa Khorey has spent the last 20 years in various roles in Information Technology to integrate electronic medical records and automate workflow processes at UPMC. Ms. Khorey currently serves as the Vice President of Enterprise Systems & Data Management at UPMC-the largest integrated health care delivery system in Pennsylvania and one of the leading nonprofit medical centers in the country. UPMC is a $10 billion, 55,000-employee organization that serves the health needs of more than 4 million people each year.

Comment on this Post

Leave a comment:

Forgot Password

No problem! Submit your e-mail address below. We'll send you an e-mail containing your password.

Your password has been sent to: