Applying Aster’s Capabilities > Healthcare > Predicting Readmissions

Teradata Guided Analytics

While Aster itself can be used to construct Applications in almost any industry, the field of Healthcare is particularly rich with opportunities up and down the delivery chain.  In this post I will focus on one of the downstream applications that we developed to leverage predictive modeling to assess whether a patient is likely to boomerang and require readmission.

 

First, let’s understand why readmissions should be of concern to all of us and why it is especially problematic and expensive for healthcare providers and insurers. According to Kaiser Health News, nearly 18 percent of Medicare patients hospitalized last year were readmitted within a month of discharge.   That represents roughly 2 million patients costing Medicare an additional $26 billion. Under closer scrutiny, 65% of these readmissions were categorized as “Avoidable”.

 

It's no surprise that readmissions are a significant problem that escalates costs and diminishes quality of care and patient satisfaction.   For the past 3 years the Hospital Readmission Reduction Program has struggled to reduce the high readmission rates; however, 75% of hospitals being monitored are not making progress. Their patient boomerang rates remain high and they are still paying fines and being denied reimbursements.

 

The application developed on Aster ingests and trains on any number of facts including diagnoses, procedure histories, patient demographic profiles, etc. All available data are mined for their predictive value, and a classification tree is constructed to predict whether or not a patient is likely to be readmitted if discharged today.

The initial predictive model built in Aster was based on data available for diabetes readmission patients, but there is no restriction on the applicability of this approach to include other diseases or events. The limitation at present is the availability of data upon which to build (and improve) the predictive model(s).

 

If you are interested in exploring any of the potential applications of predictive modeling (or any of the other capabilities in Aster) in the area of Healthcare, please contact me to engage with the Aster Solutions Team.