Coming up with a data warehouse and business intelligence strategy can be one of the biggest challenges health system CIOs face. They are being asked to pull together data from disparate sources to help reduce hospital readmissions and to support population health initiatives, and they can’t afford to take several years to get the projects up and running.
This week I had the opportunity to speak with Becky Sykes, senior vice president of resource management and CIO at Catholic Health Partners (CHP) of Cincinnati, about her decision-making on data warehouses. CHP provides care for 5 million people across 23 hospitals and more than 250 care locations in Ohio and Kentucky. CHP has had a long-running relationship with Premier Inc., a healthcare improvement company helping member hospitals and other providers with data and analytics and supply chain solutions.
So when the health system began thinking about data warehouse needs, Sykes thought it made sense to consider helping Premier develop its cloud-based PremierConnect Enterprise platform.
“We looked at other data warehouse vendors and found the cost to be prohibitive,” Sykes said. CHP is in the process of rolling out Epic Systems’ EHR across the whole enterprise. She described Epic’s data warehouse offering as promising but two to three years away. She said although CHP is moving toward a more homogenous health IT environment, it is still quite a challenge to pool the information from across its hospitals, non-acute sites, payers and other entities into one place. CHP has been helping Premier develop the PremierConnect Enterprise platform for about a year.
“When we had to do our reporting for ACO measures, we were caught flat-footed initially, but this let us hit the ground running at a price point we could afford,” Sykes explained. “Setting up your own data warehouse could take two to three years. This allows us to quickly leverage IBM’s infrastructure and Premier’s expert team and tools.”
CHP is also one of the founding members of the Premier-led Data Alliance Collaborative, which started with four members and has grown to six, said Sean Cassidy, general manager of Premier’s ITS Emerging Business Unit. The idea is to build data analytics solutions in a collaborative format that accelerates efficiencies and cost savings while avoiding duplication of effort.
The initial members were Carolinas HealthCare System (Charlotte, N.C.), Catholic Health Partners (Cincinnati), Fairview Health Services (Minneapolis), and Texas Health Resources (Arlington, Texas). It has since added Carilion Clinic (Roanoke, Va.) and Bay State Health (Springfield, Mass.)
Cassidy said DAC is in its early development and once it is better established it will be rolled out to a broader community of health systems. Right now the DAC organizations are focused on sharing information and resources, but he added that Premier has started to look at ways that the healthcare organizations could monetize the intellectual property they contribute.
DAC members co-developed a way to quickly notify providers of groups of patients who have not filled prescriptions within 24 hours of discharge, and to immediately intervene. Now they have turned their focus to co-developing an all-cause predictive readmissions model that analyzes both EMR and administrative data to identify patients who are most likely to be readmitted before they are discharged. The DAC model will also identify risk factors leading to readmissions, tying patients to appropriate evidence-based checklists based on their condition.
Sykes said it is important to look at all possible reasons someone is readmitted, use predictive analytics to identify those people early and get that information into the hands of embedded care coordinators working in primary care practices.
It will be interesting to see if the DAC model continues to work well for its members and grows to include many more organizations in the emerging era of big data. At Healthcare Informatics, we will keep an eye on their progress and report back.
