A lot has been made of the complexity of PDPM. We’ve all heard by now there are more than 28 thousand code combinations. Let me get straight to the point: this narrative is wrong and is being used to make you feel like PDPM is too complex for you to handle.
The “more than 28 thousand combinations” is coming from people doing simple math on the PDPM categories. The table below shows the number of casemix categories for each PDPM payment category:
| Payment Category | Casemix Categories |
|---|---|
| PT/OT | 16 |
| Speech | 12 |
| Nursing | 25 |
| NTA | 6 |
If you do simple math on that, you find that 16 × 12 × 25 × 6 = 28,880. That’s where the 28 thousand number came from. The problem with that is some of those codes simply can’t go together, ever.
For example, if you find yourself with a GG score of 24 points for PT and OT, (extremely unlikely) then you will end up as one of TD, TH, TL or TP. Your nursing GG score will be 16 points. There are only 6 nursing categories that you can get into with 6 nursing GG points, not 25. (2 clinically complex, 2 behavioral and 2 reduced physical function.)
In fact, only 4 of the PT/OT categories allow you the option of the full 25 nursing casemix categories. This significantly reduces the HIPPS code universe. In fact, it reduces the universe by 39%. That means there are only 17,568 combinations.
I can read your mind: “17,568 is still a LOT.” Fair enough. There is more we can do though.
When CMS first proposed PDPM they included a “Provider Specific Impact File”. That file includes projections of category utilization within PDPM calculated from 2017 data. While using that data for serious projections is probably not a good idea, it’s good enough to make an estimate of the number of HIPPS codes we are likely to see.
So what the CMS projection tells us is “PT/OT category TA would have occurred 0.559% of the time.” What the data doesn’t tell us is how often PT/OT category TA would have appeared with SLP category SA for example. We’ll need to do some math for that.
Before we do that, we want to remove the 11,232 HIPPS codes that we’ve already determined are impossible. Now, I’m aware that there are some problems with this analysis. The biggest is that providers are already making adjustments for PDPM that are going to significantly change the nursing and NTA categories. My assumption here is that the number of HIPPS codes in use won’t change dramatically. The actual codes will shift to reflect changes in operating procedures. Even if you believe we’ll get more coding diversity, it doesn’t change this analysis much. You’ll see why in a moment.
So, on to the math. Essentially what we’re trying to do is figure out how often a particular HIPPS code should appear given that we know approximately how often each of it’s 4 components should appear. For example, for HIPPS code KAXD, we know that HIPPS codes should start with ‘K’ 36.7% of the time and should have an ‘A’ as a second letter about 40.5% of the time, etc. Now, we need to do that for 17,568 HIPPS codes, simultaneously. (Simultaneously because each code’s weighting effects every other code.)
Once that’s done we find there are a good number of HIPPS codes that will essentially never occur.
Check out this table. 98.8% of the remaining 17,568 HIPPS codes will only be seen once in every one thousand residents. Over seven thousand codes are actual “one in a million” codes.
| Odds | % of Possible | HIPPS Codes |
|---|---|---|
| ≤ 1 in one million | 41.3% | 7,252 |
| ≤ 1 in one hundred thousand | 71.6% | 12,586 |
| ≤ 1 in 10 thousand | 91.7% | 16,109 |
| ≤ 1 in a thousand | 98.8% | 17,358 |
Other Statistics
-
90% of your assessments will use less than 2,044 HIPPS codes.
-
50% of your assessments will use less than 218 HIPPS codes.
-
The top ten HIPPS codes account for nearly 10% of all use cases.
As you can see the HIPPS codes we’ll actually use are probably going to be concentrated around a few. This complexity is manageable and far less than it appears.
If you’d like help reducing the complexity of your rehab program, contact us.
p.s. – I purposely left out specific HIPPS codes due to the shifting that’s going to occur. However I’m making an exception for DLWA. The odds of seeing this HIPPS code on an actual assessment are 1 in approximately 215,877,000,000. (That’s 215 billion.) You are 739 times more likely to hit the Powerball than see this HIPPS code. Since you’ll never see it in real life, I figured I’d give it a moment in the sunshine. There you go DLWA. #15MinutesOfFame
