Case Study 4L Data Intelligence
Hospice is a specialized form of healthcare that focuses on providing comfort, support, and care to individuals who are facing a terminal illness or nearing the end of their life. Our client specializes in analytics for this specific field of healthcare. Their goal is to manage hospice through predictive analytics.
The Problem
As per Medicare regulations, actual revenues in excess of the allowable expenditures will have to be refunded to Medicare. Therefore Hospice admins would like to keep the revenues within the allowed budget. The end-of-life journey for each patient is unpredictable, as patient conditions vary significantly.
It’s very challenging to accurately estimate the duration of hospice care for each case. Thus, making it hard to balance providing high-quality care while also managing costs to avoid exceeding the allowable expenditures.
Our Solution
Struggling to make the right predictions with the data they had and stay within budget, our client came to us to gain a solution. We worked with our client to develop an AI application that provides visibility into the current year as well as previous years, where there is the risk of actual revenues exceeding the allowable limit. Hospice administrators have the opportunity to suggest a measure that can bring the actual revenues back within the allowable range.
It was our responsibility to understand and document all the logic related to the Hospice release, including the business reasons behind it. Our AI solution provides subject matter expertise (SME) on hospice revenue management, improved CAP calculations, and enhanced prediction logic for patients in the current census.
Our AI system created a mapping of the new EMR systems to the 4L database, improved the existing ETL process through an enhanced and scaled cloud-native architecture, and increased the predicted patient length of stay accuracy.
In addition to documenting our understanding and contributing to application enhancements, we implemented a new ETL architecture using PySpark with Snowflake as the underlying database and redesigned the database schema to follow a star schema pattern, which is more suitable for efficient analytical workloads.
Our AI solution provides subject matter expertise (SME) on hospice revenue management, improved CAP calculations, and enhanced prediction logic for patients in the current census.
