A team of researchers at Carnegie Mellon University is conducting interviews to understand how data scientists select outcome variables targeted by predictive models in healthcare. Data scientists who have felt uncertain about past target variable modeling choices may be especially interested in participating in this interview.

Interviews will be conducted via Zoom, last 1 hour, and are compensated with a $35 Amazon gift card. The team will ask about times in the past in which you felt uncertain about your choice of target variable, in addition to your current model evaluation practices. You will also be asked to provide feedback on model evaluation prototypes demonstrated on the MIMIC-IV ED database.

The team will ask for consent before using any direct quotes from this interview. You will be given a chance to edit quotes or decline to have quotes included as needed. The team will not name your current or former organization or provide any identifying characteristics in publications unless explicitly requested.

Feel free to schedule an interview via this link if you are interested in participating:

https://calendly.com/lukeguerdan/picking-an-outcome-variable-healthcare

Please email Luke Guerdan at lguerdan@cs.cmu.edu with any questions.