Inspiration: We need more data during clinical trials to increase statistical power and help cluster patients and to improve clinical trial testing for skin diseases by making it easier for both companies and patients while maintaining accurate data.
How we did it: Phone application that provides a way for clinicians to easily collect daily photo data of skin lesions from patients while improving patient compliance. Images can be used in addition to biopsies from more infrequent clinical visits. Statisticians can then integrate daily extracted information to make decisions on selecting patients or changing doses as the various trials progress and add support for remotely conducted global studies
How we did it: Database in MongoDB, Phone app coded from shutter/dart, machine learning to analyze lesions, python, and JSON to scan in different files and documents
challenges: connecting all the different applications together, staying awake, and solving bugs
positive experiences: Working hard together as a team, coming up with an inspiring idea and carrying through with it, distribution of work, having fun coding together!
What we learned: How to use different coding languages, different shell commands to use, how to connect to cloud and set up a database
Future for Kira: Notify patients of relevant trials -Control images -As data comes in, automatically use a model to check if the image quality and type is consistent across different days -Using data later will be much easier -Better attribute extraction -Target more disease areas -Integrate with data from google fit or apple health