search

data2health / DREAM-Challenge

EHR DREAM Challenge
7 stars 2 forks source link

Identify how best use the WashU synthetic dataset #31

Closed tschaffter closed 4 years ago

tschaffter commented 5 years ago

These are "meaningful" synthetic data that we could use for a sub-challenge, for example.

We have a call with Randi on June 18 between 8 and 9am PDT.

tschaffter commented 5 years ago

@jguinney @sdmooney @trberg Randi presented how the WashU data are generated. Generating a dataset for the challenge will however be difficult because the dataset generated would be missing many of the EHR tables/properties that are expected to be found in a standard EHR.

While WashU data may not be used for the challenge itself, it could be used to address additional questions after the end of the challenge. An idea was also to generate a tailored WashU dataset for each of the best-performing model so that a give dataset includes at least the EHR information corresponding to the main features used by a given model.

tschaffter commented 5 years ago

@jguinney @sdmooney @trberg Randi will be at the CD2H All Hands F2F Meeting and is looking forward to meet and discuss on how to best use the WashU data in the context of this EHR DREAM Challenge.

trberg commented 5 years ago

Currently, we are in the process of working with WashU to not just used their synthetic data, but their non-synthethic OMOP repository for this challenge.