There are times in every healthcare leader's life when they feel overwhelmed by information and they struggle to make sense of it all. The chatter of big data and analytics tools indicate there are solutions, but some don't know what these words mean. Many healthcare professionals have already embraced data-driven decision-making, so what exactly do data analytics have to offer?
In short, data analytics involve using technology to leverage raw data from one or more systems for a more complete picture.
Anyone who has faced a CFO with a report showing costs rising in a certain department or line item will understand that looking at that one data point alone doesn’t provide the underlying story or explain variances to the budget.
To get to the why, a healthcare manager would have to dig through historical data and talk to their team to try and figure out the why, much less the "how should we respond." Data analytic technologies facilitate both of these processes by leveraging all of the information in systems to understanding past behaviors, gain insight into probabilities and present suggested action to be taken.
Halo, a business intelligence platform, breaks data analytics into three main categories.
- Descriptive analytics, which use data aggregation and data mining to provide insight into the past and answer: "What has happened?"
- Predictive analytics, which use statistical models and forecasts techniques to understand the future and answer: "What could happen?"
- Prescriptive analytics, which use optimization and simulation algorithms to advice on possible outcomes and answer: "What should we do?"
The best way to understand how data analytics can help improve healthcare is to look at examples where they have improved patient outcomes, led to more effective use of resources and reduced costs.
The 2015 Most Wired List included a story about Lakeland Regional Health System and their use of data analytics to fight sepsis. The technology mines data in the electronic health record (EHR) system and uses algorithms to predict which patients are becoming septic.
Care teams also use the technology to identify patients at risk of deep vein thrombosis, catheter-related infections and congestive heart failure patients who may be at risk for readmission within 30 days of discharge. This helps clinicians more effectively allocate and deliver resources where and when they are needed most.
The Defense Health Agency uses its Military Health Population Health Portal to help care teams at its military treatment facilities across the world to identify those patients who are at-risk for over-utilization of services, chronic disease or poor outcomes. The risk-adjusted system leverages multiple sources of information, such as the EHR, registries and claims data, to help disease managers and care coordinators identify and stratify patients.
They can track diabetic patients who haven’t had a recent A1C level or those or haven’t had a mental health follow-up appointment after an inpatient psychiatric hospitalization. Metrics on important dimensions of care, such as HEDIS measures, are improving because patients who failed to meet the metric are identified and care teams are able to determine needed interventions. The improvements in care for active-duty, retirees and their families also contributes to cost savings.
Imagine what could be next? Could data analytics help providers identify their patients who are at risk for controlled substance abuse and death? There are 20,000 deaths due to an overdose of opioids in the U.S. annually.
Imagine a day when data analytics not only identifies those at greatest risk, but also those at increasing risk. Providers would be alerted to patient behaviors along with suggested interventions. The technology could leverage data already being collected by pharmacy benefit managers, state registries, EHRs and other systems.
Right now, even though information is available electronically, the analysis process is more time consuming for physicians and their staffs. The greatest risk is that they are only evaluating those patients with potentially obvious signs. What about those who are skilled at hiding the signs of abuse?
Data analytics is about letting the computer analyze historical data and forecast the future and removing some of the subjectivity from a completely human process. Data is being broken down so the information can be better understood, modeling and machine learning are then helping to forecast the future and suggested actions are being offered.
This all leads to a more robust analytic environment that will help leaders better understand the data and make more informed decisions.
