Answer four key questions to transform report data into actionable data

May 22, 2017 / By Kristine Daynes

The digital age has created more noise than useful information. This is especially true in health care, where the quantity of electronic health information has exploded over the past decade. Countless observations are now ever-present and ever-accumulating in data warehouses, as analysts dedicate themselves to finding the signals and patterns that have meaning.

The sheer volume of available data can be a distraction. As organizations try to tap their IT investments for hidden value in the data, the biggest difference comes from how they use the data.

Data analysts are challenged to move from being report generators who assemble information on demand to becoming problem solvers who find answers that can influence decisions and change behavior. Actionable data starts with asking the right questions.

What problem are you trying to solve? So much effort is already spent monitoring process measures and reporting outcomes to make sure programs are on track. This shouldn’t become an end in itself. Improvements only happen when you decide what to do in response. For example, most hospitals pay close attention to their readmission rates, because they are tied to so many value-based reimbursement programs. Some hospitals only monitor readmissions as if knowing their readmission rates would solve the problem. Yet the problem most hospitals want to solve is how to reduce their readmission rates. They seek to identify what type of patients consistently readmit more than expected and explain why.

What will a user do with this information? And in what context? It’s natural for analysts to consume data as reports and organize lots of information in dashboards. But clinicians and other end users frequently consume information in limited or explicit pieces to support moment-by-moment decisions. They rely on reports and dashboards only to understand the state of things as they are. In order to act on information effectively, they also need relevant bits of analytics data within the IT systems and processes where they do their work.

What are the elements of a useful answer? With a deep understanding of how users consume the information and what they do with it, analysts can define what needs to be produced beyond a set of metrics. The most fundamental data describes what happened by patient type, diagnosis, physician and facility. But a better answer also identifies the cause-and-effect factors that explain why it happened and suggest how clinicians and care managers can intervene.  

Will other information, new guidelines or policies be needed to guide users to the desired behavior? Transformation in health care won’t be made by incremental improvements to the existing system of doing things. Ultimately payers, physicians, and patients need to change their behavior to promote health, not just treat disease. This requires policies that create incentives to change behavior and evidence-based guidelines that show how to behave more effectively. A successful program marries actionable data with the necessary policies and guidelines to support the desired behavior.

Health data analytics is in its infancy. Many analytics teams are just beginning to master reports that monitor operations. As their IT systems and capabilities mature, their analyses will undoubtedly become more complex. But the way they deliver data to be consumed by the rest of their organization should become simpler and less noisy. Analysts can begin to generate more actionable data by designing analyses to solve problems, influence decisions, and change behavior, rather than just explain history.

Kristine Daynes is senior marketing manager for payer and regulatory markets at 3M Health Information Systems.