Wow, this is truly good, as long ago I did read many delays on public healthcare services are due to no-shows. I liked the fact that with the information of who were more likely to no-show, UHP then contacted these people.

UHP was able to cut no-shows for patients who were highly likely to not to show up, by more than half. That patient population went from a dismal 15.63% show rate to a 39.77%. A dramatic increase. At the same time, patients in the moderate category improved from a 42.14% show rate to 50.22%.

Of course, this article sounds like an ad for eClinicalWorks, but interesting and very good application of AI regardless.