When people think of groundbreaking medical advances, they think of Boston, New York, maybe Los Angeles or Houston, but not Orlando. Florida Hospital, a major faith-based, full-service, academic and teaching healthcare system headquartered in Orlando, however, might change that.
Breakthrough medical insights are expected as Florida Hospital begins sifting through the clinical data from the system’s eight hospitals. Facilitating the analysis is Cognos 8 running on a z10 and DB2 10. The hospital’s initial plan called for the use of IBM’s Smart Analytics Optimizer (SAO) with its data warehouse on their z10, but IBM unexpectedly decided to end support of SAO on the z10. Robert Goodman, the lead DBA for System z at Florida Hospital, plans to still get there with Cognos, a product the data warehouse would have needed in any case.
Maybe the way to think of this—the z10, DB2 10, Cognos, and eventually SAO, reincarnated by IBM as IBM DB2 Analytics Accelerator (IDAA) for the zEnterprise, combined with terabytes of highly compressed clinical data—is as a country doctor’s scaled-down version of IBM Watson. It will take advantage of the data cubing capabilities of Cognos 8 and later the in-memory data analytics of IDAA, similar to Watson, and it will include the new DB2 10, with its ability to handle temporal data, or data with a time dimension.
This lands Florida Hospital at the forefront of IBM’s new vision for data warehousing, the enterprise data hub. The enterprise data hub, a more flexible systems architecture, entails consolidating the data infrastructure to reduce data mart sprawl and streamline analytics. IBM is so enamored of the vision that it has identified health information as an early application for Watson, the IBM Power cluster that beat champions of the quiz show “Jeopardy!” last year.
The Florida Hospital medical data warehouse goes beyond the efforts of Goodman and even those of the hospital itself. Rather, it’s partnering with Sanford-Burnham Medical Research Institute. Sanford-Burnham focuses on discovering fundamental molecular causes of disease and devising innovative therapies to combat them. With operations in California and Florida, Sanford-Burnham is one of the nation’s fastest-growing research institutes.
Florida Hospital, however, also is a significant medical institution. One of the nation’s largest not-for-profit hospitals, it handles more than 1 million patient visits per year as part of the Adventist Health System. The more than 2,000-bed system is comprised of eight hospitals located throughout Central Florida. Florida Hospital is home to the largest diabetes and endocrine practice in Central Florida as well as the Translational Research Institute for Metabolism and Diabetes, and it has received numerous awards for quality care.
Goodman’s high expectations for the medical data warehouse research initiative are well-grounded and not just because of the technologies involved. Sanford-Burnham ranks among the top independent research institutions nationally for National Institute of Health (NIH) grant funding and among the top organizations worldwide for its research impact. From 1999 to 2009, Sanford-Burnham ranked first worldwide among all types of organizations in the fields of biology and biochemistry for the impact of its research publications as defined by citations per publication, according to the Institute for Scientific Information. According to government statistics, Sanford-Burnham ranks second nationally among all organizations in capital efficiency of generating patents, as defined by the number of patents issued per grant dollars awarded.
The hospital describes Sanford-Burnham as using a unique, collaborative approach to medical research, and the Institute has established major research programs in cancer, neuro-degeneration, diabetes, and infectious, inflammatory, and childhood diseases. The Institute, a non-profit public benefit corporation, is especially known for its world-class capabilities in stem cell research and drug discovery technologies.
z10 Starting Point
The effort began with the upgrade from a z9 to a z10. A z196 might have simplified this effort, especially when the IDAA is added, but the timing precluded this option. The hospital wanted the z10 well before the z196 was announced.
The hospital couldn’t wait because it needed DB2 10, which doesn’t run on anything below a z10. DB2 10 offers several advantages, including support for temporal data. This proved critical to Goodman’s data warehousing plans. The hospital has been capturing and storing clinical data going back 15 years. For the past five years, the hospital also has been capturing genetic markers as part of the data. The time dimension is big.