There are isolated cases where clinical decision support can work well. The best example of this are the ones that analyze prescriptions for dangerous combinati

There are isolated cases where clinical decision support can work well. The best example of this are the ones that analyze prescriptions for dangerous combinations or mistaken quantities (too much for age, weight, etc...). The reason they haven't caught on in the general case is because medicine is full of unique edge cases, and like much of AI, teaching a machine about all of them becomes exceptionally difficult.