Finding Algebraic Structure of Care in Time: A Deep Learning Approach
This addresses the challenge of capturing complex healthcare processes for improved patient risk prediction, though it appears incremental as it builds on existing embedding and recurrent methods.
The paper tackled the problem of predicting future risk from Electronic Medical Records by modeling the dynamic interactions between diseases, treatments, and recording practices using an algebraic deep learning approach, reporting preliminary results on predicting unplanned readmission for chronic diseases like diabetes and mental health.
Understanding the latent processes from Electronic Medical Records could be a game changer in modern healthcare. However, the processes are complex due to the interaction between at least three dynamic components: the illness, the care and the recording practice. Existing methods are inadequate in capturing the dynamic structure of care. We propose an end-to-end model that reads medical record and predicts future risk. The model adopts the algebraic view in that discrete medical objects are embedded into continuous vectors lying in the same space. The bag of disease and comorbidities recorded at each hospital visit are modeled as function of sets. The same holds for the bag of treatments. The interaction between diseases and treatments at a visit is modeled as the residual of the diseases minus the treatments. Finally, the health trajectory, which is a sequence of visits, is modeled using a recurrent neural network. We report preliminary results on chronic diseases - diabetes and mental health - for predicting unplanned readmission.