Hi folks, 
I am super thrilled to say I am finally making an initial foray into the SciML ecosystem! In brief, I am in a mathematical modeling course particularly utilizing Differential Equations to model various types of phenomena. What’s also very exciting is there is a final project for this course which is basically, “Pick a problem space and attempt to model the problem with mathematical models we have been using throughout this course.”
I have the shaping of an idea for my final project that I want to run by you all; here is the (admittedly initially vague and broad) idea:
I work in the space of patient medical records, claims, and at times lab procedures. Each entry for a patient is generated when the encounter a medical system (like a clinic, hospital, etc.) and timestamped into a database. What has been found is that this information can be used to make predictions about mortality rates, risk for adverse events, and so forth. Many machine learning systems and tools have been developed to utilize this data to make such predictions or observations. My question then is how could I use differential equations to model patient phenomena and make predictions using such medical records?
I have reviewed literature, discussed with other researchers how ODEs are being used in this space, and seen that this approach in this space appears to me rather sparsely explored. Is anyone aware of anyone bringing differential equations approaches to this space? If so, could you share anything you are aware of with me? I am a novice to these approaches so I may have just been looking in the wrong place. Additionally, would it even make sense to bring differential equations to bear on this problem space versus what I’ve seen as traditional machine learning approaches?
Thanks!
~ tcp 