Print Page | Contact Us | Sign In | Join Now
News and Press: News & Views

Guest Article: PDGM and Home Health Length of Stay (With Free Data!)

Monday, July 15, 2019   (0 Comments)

By Duane Blackwell, National Home Health Analytics

We all know by now that the Patient-Driven Groupings Model (PDGM) is staring us square in the face. We all at least think we know how we might individually perform under the new payment scheme. We know that post-acute admissions are a good thing. We know that “early” episodes are much more financially attractive than ”late” episodes (although ‘better late than never’ applies here, I suppose). We also know that our contract therapy friends might be getting just a wee bit antsy about their future income streams. But have we given much serious thought to length of stay (hereafter “LOS”) and its PDGM impact? Stick with me here, I promise this is going to get interesting. Also, I’ve got some fascinating FREE agency-by-agency data I want to send your way.

OK, so home health LOS has traditionally been measured as the number of episodes delivered per Medicare unduplicated beneficiary. Nationally, the mean number of episodes per Medicare beneficiary in 2017 was 1.89. In Oklahoma, the state with the nation’s longest 2017 LOS, the measure was 2.82 episodes per beneficiary. In Hawaii, the state with shortest LOS, there were only 1.26 episodes per beneficiary. But I want you to keep in mind that the traditional measure is a kind of “flyover” look at LOS. Here at The PDGM Sweatshop (that’s ”Pretty Damn Good Metrics”), we decided to take a slightly different approach to LOS and dig just a bit deeper, state-by-state, and agency-by-agency.

The Deeper Dive

Rather than simply take the total number of episodes delivered and divide that number by the total number of unduplicated beneficiaries in a given year, we counted how many episodes each individual beneficiary received during calendar 2017 and then dropped each beneficiary into one of six “buckets” based on their individual episode counts. The chart below tells us that 14.59% of Oklahoma’s 2017 Medicare beneficiaries received six (or more) Medicare home health episodes during 2017, while 45.29% received just one. In Hawaii, on the other hand, less than 1% of the beneficiaries received six or more episodes while 81.41% received a single episode. That’s interesting enough (at least to me), but let's peel that onion a bit further.

All of the data we are considering here are from the calendar 2017 claims file. So if a patient began their home health stay in, let’s say the month of December 2017, that beneficiary would likely land in that “one episode” bucket. That patient may have been recertified in 2018 and may actually still be on census. We wouldn’t know the patient’s actual individual LOS until that patient is discharged. All we know for certain is that particular patient had a calendar 2017 LOS of one episode. Make sense? By that same token, a patient with six 2017 episodes may actually have an even longer actual total LOS if he/she was on census prior to 2017 and, perhaps, is still on census today. So it’s important to understand what the data are telling us. It occurs to me that actual LOS is really understated when measured by the standard total episodes divided by total beneficiaries methodology. Our alternate LOS measure, while perhaps not perfect, helps to turn the cube at least a bit. Ponder that for a while and let me know what you think.

We have created a spreadsheet we are calling the “Episode Frequency Report” that includes individual alternate LOS metrics for every Medicare-certified home health agency in the country. I’d be happy to send you a copy if you’re interested. Just shoot me an email at with the subject line “Episode Frequency Spreadsheet” and I’ll get it out to you by return email. All you’ll need to add to the report is a bag of popcorn and perhaps the adult beverage of your choice. Yes, it’s that interesting!


So what does all of this have to do with PDGM? It’s really about “early” versus “late” episodes, and “post-acute" versus “community” episodes. We all know that “post-acute” episodes are going to pay substantially more than “community” episodes, and that “early” episodes are going to pay substantially more than “late” episodes under PDGM. Right? (The correct response is “yes”.)

Well, looking at the episode buckets in our little chart above, you might tend to conclude that Oklahoma agencies are in for a proverbial “plow-cleaning” from PDGM while the folks in Hawaii are planning a celebratory luau. At least that’s what I would conclude. I mean, Oklahoma’s data absolutely SCREAMS ”late” and “community”, while in Hawaii’s data one can almost feel the relaxing contentment of a warm tropical breeze owing to the abundance of “early” and likely “post-acute” activity. There can be no dispute that the two states are on opposite ends of the home health LOS spectrum. It seems that PDGM success and/or disaster could be predicted based on such data observation. But as ESPN’s Lee Corso might say, “Not so fast, my friend!”

CMS PDGM Revenue Projections

Last Summer, the Centers for Medicare & Medicaid Services (CMS) released a spreadsheet projecting revenue under PDGM and comparing that projected figure with actual 2017 reimbursement under the current Prospective Payment System methodology using actual 2017 claims data. The CMS spreadsheet was released with agency-by-agency metrics with the agencies identified only by Medicare Provider Number. At National Home Health Analytics, we took that same spreadsheet and “unblinded” the provider numbers so as to identify individual agencies by name. We also totaled the actual 2017 revenue and the projected PDGM revenue state-by-state. If you don’t have that spreadsheet but would like a copy of it, send an email to with the subject line “DILLY DILLY!” and I’ll get it headed your way. SPOILER ALERT! — The CMS data projects that Oklahoma agencies will see a payment increase of 5.15% on average under PDGM while our friends in Hawaii will lose 6.17%. Huh? Something’s not right here. My dad always told me to never trust government math. He was one smart sumbitch!

Throughout the data, those states with long LOS (see “Episode Frequency Spreadsheet”) are the very ones projected to win financially under PDGM (see “DILLY DILLY!” Spreadsheet). Those states with shorter LOS? Well, not so much! Ahhh, something’s not quite right here. 

Mark Twain famously said “It ain’t what you don’t know that gets you into trouble, it’s what you know for sure that just ain’t so.” Mr. Twain was a smart sumbitch too! I’m a big fan of Mr. Twain and I’ve always particularly liked that specific example of his wit and wisdom. But there’s a little “inconvenient truth” at work here. You see, I recently learned, and much to my chagrin, that Mark Twain never said it. Bummer!

Something’s not quite right here. No indeed.

I just love me some irony!

Duane Blackwell is CEO of National Home Health Analytics. He can be reached at or (318) 229-2246 (cell).

Associate Members