Case Study: Leading Healthcare Company predicts patient relapse with Turing

Predicting Patient Relapse With 80% Accuracy AI-Powered Behavioral Health Insights

Leading Healthcare Company, a U.S.-based enterprise behavioral health provider, needed a more accurate way to predict patient relapse so it could tailor treatment plans and allocate resources more effectively. The company worked with Turing, using Turing Intelligence, to build an ML-based prediction framework for relapse risk.

Turing helped define relapse categories, engineer features from five years of patient data, and train and optimize models, ultimately deploying a GBTClassifier solution on Azure Cloud for real-time use. The result was an 80% ROC/AUC model, 4% quarter-on-quarter revenue growth from more customized treatment plans, improved treatment customization, and better admission-stage predictions that supported stronger patient outcomes and resource allocation.