There's no shortage of commentators who think health analytics eventually will be a game changer. But the majority of practices operating analytics systems use them primarily to review past financial performance or manage other administrative processes.
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
This is not exactly what most people have in mind when they talk about how analytics will revolutionize the delivery of care. So, how can a hospital or doctor's office take its current analytics systems to a place where they can proactively spot patients at risk for disease and deliver an appropriate intervention? Attendees at HIMSS 2013 offered some tips.
Prove analytics' value to clinicians
Terhilda Garrido, vice president of health information technology transformation and analytics at Kaiser Permanente, said clinical staff typically aren't interested in all the technical issues that go into setting up a health analytics system, but they are interested in the information these systems can provide. This is particularly true as more health systems transition to value-based purchasing models. Doctors can allocate their resources more efficiently if they can see which patients need attention. This helps the provider deliver better care and be more profitable at the same time.
People have seen [analytics] as a dark art and that scientists are the exclusive users, but in every other industry, the executive team is using the forecasts and insights that are derived from the analytics tools.
Graham Hughes, M.D.,
chief medical officer, SAS Center for Health Analytics and Insights
To get doctors interested in using them, analytics systems' technical problems need to be smoothed out. Garrido said analytics projects are often plagued by such problems as a lack of discrete data fields in a medical record or definitions that vary across databases for the same ailment or treatment. Doctors shouldn't be asked to solve these issues.
"It depends on the degree to which you can take analytics and meet physicians where they are," Garrido said.
Doctors can address some of the challenges that analytics systems face. Garrido said projects may face difficulties due to incomplete documentation and variations in documentation among staff members. But when clinical staff members aren't asked to deal with technical problems, they might be more likely to contribute to solving these softer issues.
Make the business case for health analytics
The need and opportunity for health analytics have never been greater. As the nation's providers transition to EHR systems, they are creating and storing more data than ever before. At the same time, providers increasingly are being asked to participate in value-based purchasing systems, which demand they manage patient populations efficiently. Oscar Marroquin, M.D., director of provider analytics at the University of Pittsburgh Medical Center, said these are some of the main trends that drove UPMC to implement stronger analytics systems.
Over the past year UPMC has built a system that analyzes clinical, billing, health plan and research data. Understanding all this information in context allows providers to make smarter treatment choices for their patients, resulting in more efficient use of available resources, Marroquin said.
Getting to this point wasn't easy, however. The technical challenges weren't as difficult to solve as the administrative and cultural ones, Marroquin said. Various business units throughout the health system, all with relevant data to share, were used to operating in isolation. They had to be convinced to give up some autonomy over their data.
Additionally, Marroquin said, it was important to convince upper-level management of the benefits of the analytics system. Getting the entire infrastructure up and running took a major investment of time and money, and it still hasn't shown a tangible financial return on investment (ROI). Most of the higher-ups at UPMC were nevertheless convinced that this type of system is crucial to future success, when it's likely that fee-for-service payment models will be less common.
"When we're used to living in a world of ROIs, you have to be able to educate the organization," Marroquin said. "Technical challenges are easier to deal with today than before. More important are the challenges of how to get different business units that are used to operating in isolation to see the value of sharing their data."
Look beyond the medical record
Graham Hughes, M.D., chief medical officer at the SAS Center for Health Analytics and Insights, said other industries, such as retail and banking, have figured out how use analytics to track the habits of customers and anticipate their future needs. There's no reason healthcare can't do the same thing.
Health analytics can help providers reduce hospital readmissions, identify patients who are interested in more engagement, and manage financial risks, Hughes said. To do this, systems must analyze data from a variety of sources. Demographic data is one powerful predictor of health outcomes, as is socioeconomic information. In order to understand each individual patient's unique risk factors, providers must build analytics systems that take all this information into consideration, rather than just looking at health records or medical claims data.
"People have seen [analytics] as a dark art and that scientists are the exclusive users, but in every other industry, the executive team is using the forecasts and insights that are derived from the analytics tools," Hughes said.
Understand the context of the data
Michael Zia, M.D., vice president of medical affairs and quality systems at Decatur Memorial Hospital in Illinois, said in a presentation titled "Finding the Meaning in Meaningful Use: Unlocking Excellence with Analytics" that a main pitfall of health analytics systems can be a lack of understanding of where data came from or the context in which it exists.
To highlight this problem, Zia provided an example from his own hospital. Its emergency room started tracking how long it took staff to get patients they suspected of having had a heart attack from the emergency department to the cardiac catheterization lab. The data showed many patients were making it to the lab before they left the emergency department, which obviously was not happening in reality.
It turned out that the clocks in the two departments weren't synced up. But Zia's point was that relying on data without considering its context will lead analytics projects astray. In order to derive real meaning, analysts must know where their data came from.
"It's not uncommon for people to collect data, and collect data, and collect data without anyone ever seeing where it comes from," Zia said, adding that data on its own is meaningless. Understanding its accuracy and its source is key.