Track: OPERATIONAL Sponsored by: WhereScape
Tuesday, July 15, 2014
Healthcare analytics projects are often derailed because building the integrated data sets that serve the analytics, (i.e.. the data warehouse, or other semantic layer data) often takes too long. And once built, the data-side implementation is too hard to change. MedAssets faced the challenge of blending source data from more than 10 transaction systems into a single analytics platform for customer access and clinical research after a series of acquisitions. The data represented clinical activity and supply acquisition for hundreds of acute hospitals and thousands of providers in other settings. In building the data warehouse, MedAssets faced many challenges including receiving the same clinical encounter from multiple sources and source systems that were changing to meet requirements for the new larger organization. MedAssets will detail how they used an agile development methodology and an automated approach to guide data warehouse construction while facing very aggressive timelines.
While advanced analytics holds much promise for revolutionizing how the healthcare industry survives new reimbursement models, the pressure for quick results can tempt analysts to rush through or even skip critical data due diligence tasks.
Errors can be introduced in all parts of the data lifecycle, from user mis-entry of data on the front-end, to a system data load failure on the back-end, to incorrectly applied dataset filters during the analysis itself. Insufficient time devoted to the due diligence process risks making misinformed business decisions.
This presentation will give several examples of the data complexities and analytics challenges encountered during a lab test utilization analysis conducted at Duke University Health System, and concrete examples about how overlooking certain data issues would have led to drastically erroneous conclusions and missed opportunities for improving the process of lab ordering by providers.
Through the use of data analytics, Mercy Health Partners ascertained that one of its small rural hospitals demonstrated outlier performance in the use of telemetry beds, when compared with the system average.
The potential opportunity was identified by utilizing the 3M APR-DRG grouper comparison for mortality and risk adjusted like cases, and then personnel at the outlier facility went about developing multidisciplinary teams to develop protocols to standardize telemetry care.
After 10 months, Mercy Health Partners had dramatic cost reduction in not only the percentage of cases where telemetry was used was demonstrated, but also a significant decrease in total costs for telemetry.
For too long, health care payers and providers have been doing whatever they can to support increasing reporting needs, regulatory compliance, and special requests – “just enough to alleviate and prevent an impending analytic heart attack." It's no longer enough to barely stay alive in this manner: growing information demands, consumer expectations, and heightened competition for members and patients requires health care organizations to get healthy at their core or risk the business equivalent of a coronary bypass - or worse.
But to do so requires organizations to unravel decades of siloed data, disconnected legacy software and systems, and disjointed short-term data management efforts to start fresh and do things differently. This session showcases how MVP Health Care is undertaking its own data environment transformation after years of semi-effective information initiatives. MVP built a next-generation Business Intelligence Competency Center (BICC) from the ground up to improve customer satisfaction, drive smarter decision-making, and unlock new areas of competitive advantage.
Valuable insights and success factors for health care data turnarounds will include:
- What an analytically-focused health care organization looks like and how it behaves
- Keys to delivering quick wins that build long-term executive support and funding
- New and surprising change management and data governance practices
- The single biggest factor for analytical success (hint: it has nothing to do with technology)
- Considerations for starting a BICC