Today's healthcare organizations face challenges never imagined a few years ago, including pressure from public and private payers to lower the cost of the care they provide while maintaining its quality. A recent conversation with Jim Adams, executive director, research and insights, at The Advisory Board Company, underscored the value of a mature BI model in achieving these twin mandates.
Adams made it clear that the Advisory Board's BI maturity model was all about making your BI platform more sophisticated. "Many organizations are very early in their journey toward BI," says Adams. They fall into what his model refers to as the fragmented stage of development -- level 1 out of the four levels of maturity. They may have some tools in place to meet specific needs -- a system to do quality metrics reporting to meet Meaningful Use regulations, for instance. Or they may want to collect statistics on how their physicians are performing so they buy a solution that does only that.
[ For better idea generation, read 3 Ways To Foster Healthcare Innovation. ]
Organizations at this stage of development typically underappreciate the value of data and make "gut feel" decisions, according to the model. In contrast, providers who want to take a leading role in the industry will move into level 2, the enterprise stage of BI, having created a culture that "champions emerging and growing emphasis on fact-based decisions." In practical terms, that also means they've moved beyond the narrow view that simply looks at costs, reporting and performance. They have a centralized infrastructure, common policies and standards, consolidated data management tools, and an in-depth knowledge of physical and logical data modeling.
Health systems that take it to level 3 have developed the capabilities to do advanced analytics. Now we're talking about predictive and prescriptive analytics, as opposed to the descriptive analysis that less sophisticated providers are doing. Adams summed up the differences succinctly: "Descriptive answers the question: What happened? Predictive answers the question: What might happen? Prescriptive says: What should we do? to address what happened or what might happen.
Organizations that have reached this level of maturity also have a common agenda and priorities, data normalization and an in-depth knowledge of statistics and procedural programming.
Level 4 on the maturity model, big data, is pie-in-the-sky for the vast majority of U.S. healthcare organizations. Many talk about managing big data but very few really are. Adams describes it as "one of the most overused, hyped terms today ... In healthcare, we have lots of data but don't really have 'big data' as defined in many other industries." Big data occurs when the volume of a wide variety of data (not just structured data) is crunched almost in real time and fed back to end users to make better clinical, operational and financial decisions.
"If you want to identify patients who are at high risk for hospital readmission, for example, you can't do that once a month. You have to do that daily, or even near real time," says Adams. And as most hospitals know, not reducing the 30-day readmission rate for preventable complications will incur a sizable penalty from the Centers for Medicare and Medicaid Services.
The vast majority of healthcare IT shops have not entered the brave new world of big data. But it's the kind of mature setup that will generate the wisdom and insightfulness needed to make the U.S. medical system truly cost effective.