Patient Re-admission Prediction using AWS

US based healthcare provider with a mission to provide high-quality, patient-centered care facilities for cancer & diabetic patients across multiple states in the US

• Diabetic patients getting re-admitted to a hospital with a short duration of discharge was creating a problem as that would create a negative impact
• High rates of re-admissions of diabetic patients not only lead to increased healthcare costs but also indicate that patients may not have received adequate care.
• Client wanted a reliable solution that can predict if a patient is fit to be discharged or needs to stay longer for further treatment

• CloudJournee’s consultants conducted continuous discussions with client medical staff and Data Analysts to understand the challenges that they are facing
• Collected and analyzed the data from multiple data sources and multiple data formats and understood the complexities of critical healthcare data
• Advised on a re-admission prediction solution leveraging AWS AI / ML services like AWS Sagemaker and developed a solution that can predict avoidable hospital re-admission of diabetic patients within 30 days of discharge
• Built multiple ML models using different algorithms and did continuous tuning and testing to get accurate results
• Deployed the ML model in client application using which doctors were able to predict patient status and take further decisions related to patient discharge

• The client was able to reduce avoidable hospital re-admission from 20% to 5%
• Reduce medical service costs by up to 12% that occurred due to readmission
• Improved trust and patient satisfaction with improved quality of care