Math > Mayans: a Nate Silver Book Review

The Mayans came out with the bold prediction that the world would end two days ago.  For better or for worse, the Mayans were full of it, and we're all still here. But according to the hottest thing since sliced bread - Nate Silver - we shouldn't be surprised.  

If you've never heard of Silver - he hit it big for using math to predict the election.  Before that he developed one of the top major league baseball forecasting systems - PECOTA.  And before that he was a professional online poker hustler.  Silver's recent book The Signal and the Noise: Why Most Predictions Fail - But Some Don't is all about his specialty - predicting.  The thesis: predicting is hard, but some things are easier to predict.  For instance, Silver analyzes earthquake predictions and weather forecasts - while the computing industry has allowed weather forecasters to take leaps, earthquake hunters still have miles to go  

While Nate Silver is definitely a bonafied genius.. he's not quite an original, more like the remix. He's the same genius as Bill James and the Moneyball crew that first applied mathematical analysis to baseball.  He's the same genius as the guys who decided to infuse bioinformatic analysis to biomedical research.  Silver was simply the latest to add a splash of math to a field in desperate need of it.  

So my question is: how can we apply the same genius to medicine?

to be fair, it's not like airplanes
have advanced much either

Here's one of the main reasons I think med school education is outdated.  We need a clearer definition of a good doctor.  Since our current definition is utterly vague, med students are left with the same training program that educated doctors in the era of kitty hawk.  

In track - you have the goal of running as fast as possible.  In order to do this, you don't just run around in circles randomly.  You run according to a rigorous schedule, each day with a set purpose.  One day might be optimizing VO2 max, another increasing lactic acid tolerance, and yet others to train neuromuscular junctions to fire optimally.  It's a real science.  

We need to get med school education to that same point.  Once we have a scientific defintion of "a good doctor" it'll become a lot easier to figure out how to educate "a good doctor."

I bet you thought we were just
running around in circles

So this is where it gets hard - quantitatively defining a good doctor.  What sort of stats should we be keeping on doctors?  I don't have any good ideas, so instead here's a list of half-baked ideas Kevin Wildes-style.

1- (%age) times washed hands per opportunity

2- # of times yelled at a nurse per shift 

3- total amount of shit other doctors talk about you (respect of co-workers)

4- triathalon time (endurance is important)

5- miles walked per shift (shows commitment and effort..?)

Anyways, please send your half baked ideas my way.  I need to incubate some more ideas.  

See you on the other side,

from ken

now taking requests for future blog posts. please comment or email.