Often, rare immunodeficiency diseases are misdiagnosed to other more common or minor diseases (e.g. colds, bronchitis). Now that there are data aggregators for healthcare claims, diagnoses, etc.., over multiple years how can this data provide greater insights to diagnosing rare diseases for immunodeficiency?
Want to learn more?
What we did_
By working with aggregated claims, a data set with over 15,000 features was generated. A machine learning algorithm was trained to identify cases where the patients were positive for cases of immunodeficiency.
Improved the diagnosis accuracy by over 7% when simulating current diagnosis standards. Opened the ability to integrate data-driven diagnosis with electronic medical record systems, using historical health data.
Tags: User Research, Rapid Prototyping, Big Data, Small Data, Machine Learning
Created At: West Monroe Partners
< Previous / Next >
Improved the diagnosis accuracy by over 7% when simulating current diagnosis standard