A population health consultant was sharing 2 projects which she was heading at a local hospital. One used EHR records to identify patients who were at potential risk for stroke if they had atrial fibrillation while the other used them to predict 12 month risk of emergency admission. It seemed like the hospital and primary care networks were able to effectively target their resources to patients who needed them even when patients did not present to healthcare establishments.
However for each of these projects the consultant had to spend a lot of time going back and forth between the data scientists/computer scientists to specify what she wanted/needed. So I thought that it would be great if we could find out from doctors how they approach these issues and then implement a common framework that would enable all these analyses once it is plugged into an EHR database.
What it does
Generates exploratory plots from various parts of EHR records based on a population obtained by filtering on various parameters like Age, gender, race, conditions etc.
How I built it
On my computer
Challenges I ran into
Understanding the way different libraries and frameworks functioned
Accomplishments that I'm proud of
What I learned
How to use flask and App engine
What's next for Data Analysis System for Senior's Health (DASSH)
Implementing filtering on composite features like BMI and clinical risk scores like CHADVASC. Implement selection of segments of the sub population from graph interactions Implement the ability to pull up individual patient details Implement additional exploratory plots for areas like socioeconomic determinants of health