One of the things that annoyed me most when I worked in business was the insistence of upper management that we achieve an error rate of zero.
First of all, such a rate is *actually* unachievable (something about calculus and limits and asymptotes and all that) though possibly achievable theoretically.
Second of all, if you were to try to achieve an error rate of zero, you'd expend an inordinate amount of resources doing so that would be most likely better spent on something else.
And finally, in trying to achieve an error rate of zero, you'll create perverse incentives, which may cause people to partake in other unattractive behaviors in order to game the system.
(Sorry this is so unscientifically phrased.... it's been a while since I tried to speak econ.)
The problem was, if you raised any of these points in a business environment, and suggested deciding on what the optimum error rate should be, some doof manager would decide you were just lazy and "not a team player" and that would be it for you.
And now you're wondering how this is related to my research.
I went to a talk on hospital readmissions last week, and there it was again. A guy was talking about ways to reduce the 30 day hospital readmission rate to zero. They were identifying patients who had a "high" probability of being readmitted using available hospital admission data, and "targeting" them in order to reduce readmissions. He'd come up with a model that had a sensitivity of (drum roll please) 25%, which basically means that you cannot identify these patients using his data. And he'd "designed" a fancy Excel spreadsheet tool (oooooohhhhhhhh!!!!) that allowed him to fiddle with the numbers in an expedient way in order to determine whether or not an intervention program would save money.
(As a side note, lesson #1 in my career was that many managers are easily impressed by things like one's ability to manipulate an Excel spreadsheet. No, not by your novel thoughts, but by your ability to make Excel dance. Yes, that is disdain you are picking up upon.)
Ok, let's put aside the fact that he wasn't even looking at cost-effectiveness, and that there were no confidence intervals around the numbers he was producing, or the fact that his model could not predict shit. What disturbed me the most was when the only economist in the room (who is a brilliant brilliant man) said, "The optimum readmission rate isn't zero," and this guy replied, "The optimum PREVENTABLE readmission rate is at zero."
I swear, it was like talking to a wall.
Which made me realize that the guy giving the presentation truly had no idea what he was doing. Or if he did, he had been so steamrolled by some moron manager he worked for that he was actually unable to understand the point the economist was trying to make, or at least felt he couldn't acknowledge such a point in public for fear of angering some higher-up.
It also gave me a chance to reflect on the fact that several people have recommended this guy's career trajectory to ME, and made me want to cry just slightly. I'd much rather be thought of like the economist....
Anyway, none of this is to say that quality improvement is a bad thing, or that we shouldn't try to do better. It IS to say that there is a reason (actually reasonS) that I ran screaming from the business world.
And also, I'm wondering whether a six sigma approach might work as alternative.
5 Pearls of Wisdom:
it's amazing how little statistics most people making the decisions actually understand.
I also don't know in what universe these guys exist where they'll have zero readmissions. All it takes is one really sick guy or one crappy nursing home to ruin that.
Yeah, realistic goals & expectations are key; and trying to decrease ANYTHING to zero creates SO MUCH unnecessary cost! Does this guy REALLY think spending $100000s to keep, for example, ONE person out of the hospital is reasonable? No treatment has a 100% cure rate---relapse of disease is GOING to happen in some cases. It's making my head hurt to think that people actually believe otherwise!
It is interesting that this happens even MORE in the business world---i always thought it was the lack of economics education in the medical field that led to this kind of b.s.
Ana -- Add to the fact that there's only so much one can do on the inpatient side of things to prevent a 30 day readmission.
I was thinking, why not start by looking at readmissions that occur within 48 hours, which are more obviously the hospital's responsibility than a readmission that occurs after 30 days? Or readmissions that occur from SNFs, LTACs, or rehabs, or even nursing homes. Or readmissions due to hospital acquired infections?
I think the "zero" language is mostly rhetoric. The public wants to hear that hospitals are striving towards "perfect" care, so that is what the business people who run them tell them they're doing. The truth is, most people in business don't have extensive econ training either. If they do, the common sense gets turned off the minute they need to play politics with a manager who doesn't have it.
I love QI, and my hope is to be able to do it with some level of common sense and an eye for creating productive incentives for patients and doctors. It's just so difficult to get business people (or doctors, or the public....) to talk about behavior and incentives in a rational way sometimes.
That's all stuff I never thought about really, until I took my regressions course this semester. I guess I've just forgotten about everything bigger than a protein. :) It's interesting to keep in mind!
Also, question-when you were writing your manuscripts, did you feel as though the to-do list on the manuscripts kept getting bigger and bigger? The minutiae I keep remembering to do (add that figure, take out that figure, change that legend, forgot that method...change those colors) is just piling up.
DV -- It's the difference between social science research and basic science research.
As to your question, yes the to do list just keeps getting longer and longer. It's easy to get overwhelmed and sidetracked when writing. One thing that helped me was to keep a list of all the minutiae that I still had to take care of. That way, when I was working on one thing, I could set aside the small things I thought of in the process for later and not worry about forgetting to do them. Then when you've gotten the big stuff completed, you can refer back to your to do list and take care of all the figure annotations and formatting, references, etc. then. If I didn't do this I often found that I felt I'd get nothing done. Eventually you'll address everything and your to do list will start to get smaller.
:-)
Post a Comment