How would you design an experiment when you cannot run a traditional randomized A/B test for a Point32Health initiative?
