The Challenge
Predicting and Managing Patient Readmissions
To achieve this, it required powerful predictive analytics capabilities that could derive actionable intelligence from patient data that would enable to effectively predict the risk of readmission using large amounts of existing clinical and claims data (historic load) that also included detailed assessments and identification of high risk hospitals and facilities that were underperforming w.r.t CKD readmissions, analysis of current readmission, tracking and clinical process for CKD treatment which should also include analysis of live clinical data along with assessment of clinical data for completeness and availability across care settings, e.g. registration, admission, treatment, discharge, etc . that are needed to effectively predict readmissions to predict high risk CKD patients
Our Solution
Developing a Predictive Risk Model
The team Identified over 45,000 CKD patients and more than 3,000 readmissions across 50+ facilities to achieve this. The predictive model classified patients into various predicted risk bands with an accuracy of 89% based on which patient specific risk indicators were identified for appropriate interventions, for a patient at admission or at discharge.
The Results
Deep Insights and Targeted Interventions
The predictive clinical BI solution provided deep insights into CKD readmissions. By accurately identifying high-risk patients and the factors contributing to readmission, the health system could implement targeted interventions, improve care management, and optimize costs.
- Improve care management processes through early and effective prediction of readmission risk
- Utilize the readmission predictions and patient specific risk indicators to identify the factors responsible for CKD readmissions and design appropriate clinical interventions
- Optimize cost and performance across all facilities, with significant improvements at facilities with high rates of readmissions