|Email:||pcalv (at )eden (dot) rutgers (dot) edu||pcalv (at) reu (dot) dimacs (dot) rutgers (dot) edu|
|Home Institution:||Rutgers University|
|Project:||Statistical analysis of medical devices and their adverse effects|
Medical devices are subject to unforeseen failures. Beginning with data culled from user-submitted reports on their experiences with a given medical device, we will use statistical tools to detect patterns identifying links between a device and possible adverse effects. In this way, we hope to increase safety and expedite the process of assessing medical device defects.
I read Drs. Shetty and Dalal's paper on combing through PubMed publications to discern drug-side effect links; we extrapolate the concept to medical devices. It is immediately clear that carrying over the techniques is not straight-forward, owing to differences in the data set, among other things. I have begun to look at R, and am considering possible means of sorting our data.
I looked into making a clustering algorithm with either Python or C++.
Switching gears, I began work on compiling and building a program that would implement our data clustering. Having done that, I am experimenting with what the program will accept as its input.