Part 1 in a 3-part Series on Stable Discharge Rates.

So, there’s something I need to get off my chest.  Despite the fact that I have written again (and again) that I’ve given up predicting the future, I think I want to try . . . PREDICTING THE FUTURE!  

If you’ll forgive my brief digression into baseball statistics, it’s time to talk about Voros McCracken.  Back in the year 2001, he discovered something revolutionary: 

There is little if any difference among major-league pitchers in their ability to prevent hits on balls hit in the field of play”. 

The only thing that a pitcher can control are walks, strikeouts, and home runs.  This may seem like old news (and I guess it’s > 20 years old, so it is old news), but at the time it was an extraordinary notion that was roundly rejected by mainstream baseball thinkers for many years.  So, once a batter puts a ball in play against the pitcher, it’s going to become a hit about 30% of the time (.300 batting average).  Now this has changed a bit (as teams are playing defense differently the last couple of years it’s gone down), but for the most part, regardless of the pitcher, once the ball is play, it’ll be a hit 30% (or so) of the time.  However, and this is the interesting part, batters are different from each other.  Different hitters will get hits on balls in play at different rates.  And, this rate tends to be steady over time for each hitter.  Each specific hitter will tend to have a specific rate at which his balls in play become hits.  And this is where we, finally, link back up to hospitalists.  I propose that hospitalists tend to have stable discharge rates.  They might be different from each other, but over time, each hospitalist will tend to have a stable discharge rate. 

What do I mean by discharge rates?  It’s one of those measures we’ve talked about before.  To review, if you look at all patient encounters that aren’t admissions, a hospitalist can do one of two things – keep the patient in the hospital for another day or discharge them.  So, we can create a “readiness to discharge stat” that is simply: 

Readiness to discharge = Discharges / Opportunities to discharge  

Readiness to discharge = Discharges / (Subsequent visits + discharges) 

So for example, if I have 15 patients to start the day, do no new admissions, keep 10, and then discharge 5, my readiness to discharge is:  

5/(10+5) = 0.33 = 33%  

If we invert that, we get my length of stay (LOS) for the day:  

1/0.33 = 3  

So, my LOS for that day was 3, and if we look at all of my days, I believe that with enough data, we can find my personal LOS.  Regardless of the day-to-day variations in medical practice, over time, I will tend to have the same LOS.   

Where does the prediction come into it?  Find out, in. . .THE FUTURE! 

What’s next? 

Get the latest updates and news delivered to your inbox.

Subscribe to our newsletter today.