Analytics is not easy, and provides no miracle cure for what ails healthcare

At the start of every new year, we're bombarded with resolutions and predictions of what will happen in the next 12 months. Almost every industry in 2013 is predicted to place a major emphasis on the use of tools that provide analytics, business intelligence, decision support or some other similar-sounding initiatives.

Healthcare is certainly no exception. In fact, the latest components expected to play key roles in healthcare reform—patient-centered medical homes, accountable care organizations, penalties for readmissions within 30 days and more—scream for the need to mine our data. The focus on reducing the nation’s spending on healthcare has renewed the sense of urgency to find new pearls of wisdom from the massive amounts of data we gather every day, with the aim of using newfound wisdom to improve understanding of healthcare delivery and drive improvements.

It’s no surprise that, in this environment, analytics is becoming a larger budget item, both for IT departments and executives tasked with finding meaning from healthcare data. The typical analytics platform may consist of one or more commercial products, a self-developed data warehouse or a combination of tools. Prices for all this analytical functionality can run into the millions.

In the commercial space, many vendors are responding to—and perhaps fueling—the demand for analytics tools, bringing a steady stream of new products to market. These are coming from EHR vendors, traditional business intelligence firms and smaller startup firms that seem almost too numerous to count. Some offer the ability to find nuggets of gold in “big data,” while others claim to easily process our patient information from across the care continuum, and still others provide dashboards that suggest useful information while requiring little understanding or manipulation of the underlying data.

Emerging capabilities are amazing and show promise. For example, numerous applications use predictive modeling techniques to identify patients who are at risk of readmission, enabling providers to use early intervention to try and avoid lengthier hospital stays. Other applications use various algorithms intended to identify an inpatient whose condition may soon worsen. In addition, while results of early studies are inconsistent, many believe that genomic sequencing will enable physicians to make predictions based on DNA evidence about future health issues, helping direct well-grounded interventions before symptoms even arise.

With so many touting an “analytics made easy” approach, and executives getting pressure to achieve results, IT executives must craft an analytics strategy thoughtfully and set the stage for success within their organizations. The successful strategy will encompass many organizational factors and will look beyond merely focusing on the tools that are bought or built. Among the key organizational elements are culture, structure and empowerment for change. Another key is the team tasked to perform the analytics, which must be immersed in the data itself and analytical techniques.

The team needs to clearly know the data—where it exists in the underlying systems (EMR, revenue cycle and others), how data elements are defined and how to extract them. Experts recommend that a knowledgeable “data steward” be identified for each of the major domain areas. The analysts need to be thoroughly trained in data modeling to obtain the most meaningful information—those pearls of wisdom that will truly impact care delivery. This knowledgeable team will be equipped to evaluate and recommend the toolset best suited for their success. Whatever the technology platform, its value will be proportional only to the skill and effort of those using it.

With the tools and a strong analytics team in place, healthcare organizations can be empowered with the information that’s unearthed to solve any challenge healthcare reform may bring. However, it’s crucial that organizations examine any potential barriers to achieving needed change. Certain questions need to be addressed before analytics can have the desired effect:

• Is our culture conducive to change?

• Have we educated those who are likely to be impacted about why change is necessary?

• Are those asking the questions empowered to act on the answers?

• Have we included the medical staff?

• Recognizing the outside pressures to cut reimbursement, do we have an active process-improvement and cost-reduction program in place?

• Will we, from the CEO down, make the necessary but tough decisions?

Critics suggest that the large investments in analytics will fall short in yielding benefits. The best assurance of success will rely on how we lead our organizations through change. Our futures are counting on it.

Bill Spooner

Senior vice president and CIOSharp HealthCareSan Diego