I worked with a hospital at risk of de-acreditization. They had worse outcomes than the other hospitals around them (people were way too likely to get infections at the hospital, die, etc). Turns out they had more gut knife wounds than any other hospital. All data driven decision making shutting them down would have done was make the next nearest hospital the one with worse outcomes, and so on, and so on.

Plus I'm guessing the best data driven safety change would be a reduction of package delivery load if Amazon is really doing data driven safety changes.