Earlier this week, Perficient won the 2017 IBM Beacon Award for an Outstanding Watson Solution. These awards are selected by a panel of IBM executives, industry analysts, and media members, to recognize partners that are delivering exceptional IBM-based solutions and transforming the way clients and industries do business in the cognitive era. This year’s award recognizes our Watson-based predictive modeling solution for patient readmissions, designed to better identify patients in need of interventions and ultimately reduce readmission rates.
Need for Predicting & Reducing Readmissions
Per the Center for Healthcare Quality & Payment Reform, “one of the best ways for communities to reduce healthcare costs quickly and improve patient care in the process is to implement initiatives to reduce hospital readmissions. Research studies and quality reporting initiatives around the country show that 15-25% of people who are discharged from the hospital will be readmitted to the hospital within 30 days or less, and that many of these readmissions are preventable.” They go on to predict that billions of dollars of savings would be achieved by reducing these unnecessary hospital readmissions. The U.S Department of Health & Human Services estimates that avoidable hospital readmissions make up more than $17 billion in Medicare expenses.
Enhanced Readmissions Modeling
Perficient’s Readmissions Modeling solution is unique in that rather than focusing on a single disease or condition, this solution predicts readmissions across all diseases and conditions. Another unique aspect of this solution is that it focuses on the psycho-social needs on the patient and the family rather than a traditional clinical approach, which is the focus of most of the industry. By incorporating this data with a wealth of unstructured and structured healthcare data, healthcare providers can significantly increase the accuracy of their readmissions modeling and predictions. In fact, 80% of healthcare data is typically invisible to current systems because it’s unstructured, so the value of uncovering insights in that data is invaluable when predicting triggers for readmissions.
The ultimate goal of the solution is to reduce hospital readmissions across the entire patient population, by educating patients and their families to better manage their conditions while coordinating various services within the community. Hospitals and healthcare providers are looking to identify patients who are in the most need of interventions, identify which interventions are most appropriate and begin the intervention process as quickly as possible; whereas traditionally, interventions occurred at the end of a patient’s stay. Early intervention allows for more time for clinicians to provide instruction, increase patient and family member understanding and allow time for community services to be arranged so they are ready when the patient is discharged. This solution also helps to optimize hospital resources to provide interventions to those who stood to benefit the most.
The predictive modeling process derives a prediction of whether a person is at risk for being readmitted within 30 days for each day they are in the hospital. These predictions are assigned a probability which is used to prioritize daily patient interventions. Nursing assessment data is also segmented into different risk and intervention profiles to identify health and lifestyle behaviors that clinicians can use to develop appropriate interventions and plans of care.
Solution Components
The solution leverages IBM Watson to uncover new evidence in predicting readmission propensity, ushering in a new era of evidence-based analytics. Watson Explorer Advanced Edition and the healthcare annotators are used to analyze physician notes from the EMR to extract relevant, contextual data and transform that information into structured data points. The healthcare annotators are “reading” this unstructured information, tuned appropriately to best interpret and transform TriHealth’s data. The analysis result is exported as structured data into a data warehouse, which is then consumed by SPSS Modeler along with other existing structured data to develop the predictive readmissions model. The resulting readmission risk indicator is then incorporated into the EMR system, indicating a patient’s likelihood of readmission at the point of care.
