I understand that “slow news days” call for Evergreen stories, but I implore the SNF-stream media to end its “Faux News” practice of covering a certain “Skilled Nursing Monthly Report” as legitimate dispatch. Enough with the three-letter words already. Let me be more specific – benchmarks contained in these reports hold no statistically relevant applicability to industrywide SNF performance metrics. For example, the following excerpt is from this week’s offering:
No, it didn’t. That statement is unequivocally false. Worse, propagating such obtuse “data” is detrimental to provider interests. Platforming skunky data gave it gravitas. As I explain, it’s then used against the industry to temper reimbursement. Nevertheless, this group perpetually delivers a slew of spurious and curious statements that misrepresent Skilled Nursing’s reality. The Report’s source data is self-reported by a small, homogeneous sample of facilities. Among many statistical shortcomings, the absence of geographic neutralization is perhaps the most egregious. For context, adding a group of SNFs from high-occupancy markets with outsized wage indexing would reshape its entire profile; the next release would elevate Skilled Nursing to “blue-chip” status. Is it industry trends or sample bias? Don’t know, don’t care. It’s absurd. Inexplicably, each update the-three-letter crew compiles is a months-long effort; this week spotlights February utilization… it’s May 2022. More current and relatively reliable industry metrics are readily available – April occupancy data forthe entire nation is just a few clicks away. I don’t even know what to call their deliverable. There seems to be no adjective that conveys a quality that concurrently expresses danger, misinformation, inaccuracy, humor and irrelevance. The closest word I found is “malarkey,” but that omits the “danger” element. How about a portmanteau of “danger” and “malarkey”? “Danlarkey!” That’s it. “This monthly report from the three-letter-crew is danlarkey.” Let’s take a closer look at their February analysis anyway
Nope. The National Health Safety Network data detailed below makes clear national occupancy in late February straddled 72%, not 76.7%, while pandemic-era utilization bottomed out at 67.5%. I guess things didn’t get as bad on the surface of Mars.
Note: Applying national “anything” to SNF performance is problematic, given the extreme variation across markets and shoddy industry data, but there’s more to the story. As we explained back in March, NHSN is a great resource, but the data still requires adjustment. The National Data Mavens must have missed that memo.
To properly calculate occupancy, we need two reliable datapoints: Number of patients and number of beds; that second one is a problem. NHSN’s data is self-reported by providers, who must enter a SNF’s bed count. That can be a tricky question these days. We estimate as many as 150,000 certified SNF beds are currently “offline” for various pandemic-related reasons. Data entry errors helped develop our estimate, as many operators mistakenly omit their phantom beds from the count. As a result, every stat you’ve heard about recent occupancy is overstated. When all “dark beds” are removed from the equation, February national occupancy comes in around 69.8%, yet we’re still not done qualifying the statistic.
A confluence of events has prompted many states to update their respective Medicaid rate setting methodology – the “Great Transition” – an unprecedented opportunity to rationalize reimbursement on a national scale. Unfortunately, states are recycling the same ill-conceived constructs that will do nothing to address systemic mispricing across the country. I have studied and modeled SNF reimbursement for 30 years; states must approach rate setting differently, or no amount of funding will correct the problem.
“Occupancy” is not Synonymous with “Census”
Words are powerful. In this case, we are inapplicably conditioned to process the term “Occupancy” as a measure of demand. Case in point from today’s inspiration:
These people make no sense. The occupancy equation requires context before demand can be measured; all points of reference must be time-stamped, or trends get distorted as bed count fluctuates. In other words, a stable census in a market shedding beds will report rising “occupancy” while “demand” (i.e., census & revenue) remains flat. Every market is shaped by unique factors – few can be compared without adjusting for countless variables. That’s why national performance metrics should not be used to shape local policy. Worst case scenario, regulations and payment systems destabilize the industry to the point of crises… exactly where SNFs find themselves today.
The Great BEDsignation!
Upwards of 300 SNFs have closed since the pandemic’s onset; another 550 or so discharged their final patient in the five years leading up to covid-chaos. Official numbers are frustratingly unavailable, but z-INTEL approximates net loss in the 600 facility / 50,000 bed range.
Cutting to the chase, we need a reference point to gauge demand; z-INTEL uses 2015. For example, “SNF occupancy dropped from 82.1% in 2015 to 72.3% in February;” but that statement is deceiving. z-INTEL recommends scaling the expression as follows:
“Relative Occupancy” is key to understanding Skilled Nursing’s plight. To ensure apples-to-apples, we divide the total number of patients in beds (today) by the total number of certified beds in 2015. Using 2015 as the common reference point, we have insight into underlying market distress.
SNFs struggle at a relative 68%, yet just the other day we heard the industry is nearing full recovery as 77% of beds are full. Comparatively, conventional wisdom maintains that pre-pandemic occupancy was 81%. That’s why overstating this perceived measure of demand is absolute danlarkey!
Am I overreacting? Admittedly, I get annoyed when people who should know better quote the Faux News, but trust me, it’s incredibly important. Case in point…
The (once) estimable, independent nonpartisan agency responsible for Chapter 7 of the eponymous 2022 Report to Congress that I plan to expose for all its hypocrisy later this month, tells Congress every year that SNF Medicare payments should be reduced. MedPAC claims its inviolable position is supported by data from the same Skilled Nursing Monthly Report that sparked my polemic today. This does not exactly restore my faith in government.
Wouldn’t you think MedPAC had some secret stash of Grade-A data pumps? Apparently not; its 2019 Report to Congress digs right into two scoops of Faux News from the producers of “All’s Well on SNF Street.” Looks like the “data” MedPAC trusts to claw back your Medicare revenue is as ridiculous* as that February-Fantasyland. Reminds me of an accountant I knew who did taxes out of a gas station. Gross.
For context, the closed SNF’s Medicare Part A utilization is graphed below, next to Mercer County, NJ’s Medicare Advantage penetration over the same period. Traditional Medicare days decline as Medicare Advantage grows in each market.
*Note that I did not use the term “danlarkey” here because once the damage is done, the danger is technically over.
Epilogue
Occupancy through April did indeed trend up, well at least it trickled up. NHSN will show the broader industry reached 73%, but I can point to another 700+ beds already removed this year from the calculation’s denominator, further distorting putative SNF demand. After requisite adjustments, SNF census increased just over a half-percent since February (based on accurate occupancy reporting, not the Faux News).
Meanwhile, the pandemic lingers. I’ve never seen so many Skilled Nursing Facilities in distress. Outliers aside, the Public Health Emergency has taken its toll. z-INTEL has identified more than 100,000 beds at risk of closing within 18 months without relief from intractable staff shortages, untamed inflation, and the treachery of a Medicare/Medicaid Rate Construction Crisis. This is how the stage is set for Comprehensive Reform, a plan so grand it needs no luck or logic. z-INTEL’s mission is to support development of rational payment policy with meaningful data, because from our perspective…
Faux News
Marc Zimmet
Data musings on a slow news day…
I understand that “slow news days” call for Evergreen stories, but I implore the SNF-stream media to end its “Faux News” practice of covering a certain “Skilled Nursing Monthly Report” as legitimate dispatch. Enough with the three-letter words already. Let me be more specific – benchmarks contained in these reports hold no statistically relevant applicability to industrywide SNF performance metrics. For example, the following excerpt is from this week’s offering:
No, it didn’t. That statement is unequivocally false. Worse, propagating such obtuse “data” is detrimental to provider interests. Platforming skunky data gave it gravitas. As I explain, it’s then used against the industry to temper reimbursement. Nevertheless, this group perpetually delivers a slew of spurious and curious statements that misrepresent Skilled Nursing’s reality. The Report’s source data is self-reported by a small, homogeneous sample of facilities. Among many statistical shortcomings, the absence of geographic neutralization is perhaps the most egregious. For context, adding a group of SNFs from high-occupancy markets with outsized wage indexing would reshape its entire profile; the next release would elevate Skilled Nursing to “blue-chip” status. Is it industry trends or sample bias? Don’t know, don’t care. It’s absurd. Inexplicably, each update the-three-letter crew compiles is a months-long effort; this week spotlights February utilization… it’s May 2022. More current and relatively reliable industry metrics are readily available – April occupancy data forthe entire nation is just a few clicks away. I don’t even know what to call their deliverable. There seems to be no adjective that conveys a quality that concurrently expresses danger, misinformation, inaccuracy, humor and irrelevance. The closest word I found is “malarkey,” but that omits the “danger” element. How about a portmanteau of “danger” and “malarkey”? “Danlarkey!” That’s it. “This monthly report from the three-letter-crew is danlarkey.” Let’s take a closer look at their February analysis anyway
Nope. The National Health Safety Network data detailed below makes clear national occupancy in late February straddled 72%, not 76.7%, while pandemic-era utilization bottomed out at 67.5%. I guess things didn’t get as bad on the surface of Mars.
Note: Applying national “anything” to SNF performance is problematic, given the extreme variation across markets and shoddy industry data, but there’s more to the story. As we explained back in March, NHSN is a great resource, but the data still requires adjustment. The National Data Mavens must have missed that memo.
To properly calculate occupancy, we need two reliable datapoints: Number of patients and number of beds; that second one is a problem. NHSN’s data is self-reported by providers, who must enter a SNF’s bed count. That can be a tricky question these days. We estimate as many as 150,000 certified SNF beds are currently “offline” for various pandemic-related reasons. Data entry errors helped develop our estimate, as many operators mistakenly omit their phantom beds from the count. As a result, every stat you’ve heard about recent occupancy is overstated. When all “dark beds” are removed from the equation, February national occupancy comes in around 69.8%, yet we’re still not done qualifying the statistic.
A confluence of events has prompted many states to update their respective Medicaid rate setting methodology – the “Great Transition” – an unprecedented opportunity to rationalize reimbursement on a national scale. Unfortunately, states are recycling the same ill-conceived constructs that will do nothing to address systemic mispricing across the country. I have studied and modeled SNF reimbursement for 30 years; states must approach rate setting differently, or no amount of funding will correct the problem.
“Occupancy” is not Synonymous with “Census”
Words are powerful. In this case, we are inapplicably conditioned to process the term “Occupancy” as a measure of demand. Case in point from today’s inspiration:
These people make no sense. The occupancy equation requires context before demand can be measured; all points of reference must be time-stamped, or trends get distorted as bed count fluctuates. In other words, a stable census in a market shedding beds will report rising “occupancy” while “demand” (i.e., census & revenue) remains flat. Every market is shaped by unique factors – few can be compared without adjusting for countless variables. That’s why national performance metrics should not be used to shape local policy. Worst case scenario, regulations and payment systems destabilize the industry to the point of crises… exactly where SNFs find themselves today.
The Great BEDsignation!
Upwards of 300 SNFs have closed since the pandemic’s onset; another 550 or so discharged their final patient in the five years leading up to covid-chaos. Official numbers are frustratingly unavailable, but z-INTEL approximates net loss in the 600 facility / 50,000 bed range.
Cutting to the chase, we need a reference point to gauge demand; z-INTEL uses 2015. For example, “SNF occupancy dropped from 82.1% in 2015 to 72.3% in February;” but that statement is deceiving. z-INTEL recommends scaling the expression as follows:
“Relative Occupancy” is key to understanding Skilled Nursing’s plight. To ensure apples-to-apples, we divide the total number of patients in beds (today) by the total number of certified beds in 2015. Using 2015 as the common reference point, we have insight into underlying market distress.
SNFs struggle at a relative 68%, yet just the other day we heard the industry is nearing full recovery as 77% of beds are full. Comparatively, conventional wisdom maintains that pre-pandemic occupancy was 81%. That’s why overstating this perceived measure of demand is absolute danlarkey!
Am I overreacting? Admittedly, I get annoyed when people who should know better quote the Faux News, but trust me, it’s incredibly important. Case in point…
The (once) estimable, independent nonpartisan agency responsible for Chapter 7 of the eponymous 2022 Report to Congress that I plan to expose for all its hypocrisy later this month, tells Congress every year that SNF Medicare payments should be reduced. MedPAC claims its inviolable position is supported by data from the same Skilled Nursing Monthly Report that sparked my polemic today. This does not exactly restore my faith in government.
Wouldn’t you think MedPAC had some secret stash of Grade-A data pumps? Apparently not; its 2019 Report to Congress digs right into two scoops of Faux News from the producers of “All’s Well on SNF Street.” Looks like the “data” MedPAC trusts to claw back your Medicare revenue is as ridiculous* as that February-Fantasyland. Reminds me of an accountant I knew who did taxes out of a gas station. Gross.
For context, the closed SNF’s Medicare Part A utilization is graphed below, next to Mercer County, NJ’s Medicare Advantage penetration over the same period. Traditional Medicare days decline as Medicare Advantage grows in each market.
*Note that I did not use the term “danlarkey” here because once the damage is done, the danger is technically over.
Epilogue
Occupancy through April did indeed trend up, well at least it trickled up. NHSN will show the broader industry reached 73%, but I can point to another 700+ beds already removed this year from the calculation’s denominator, further distorting putative SNF demand. After requisite adjustments, SNF census increased just over a half-percent since February (based on accurate occupancy reporting, not the Faux News).
Meanwhile, the pandemic lingers. I’ve never seen so many Skilled Nursing Facilities in distress. Outliers aside, the Public Health Emergency has taken its toll. z-INTEL has identified more than 100,000 beds at risk of closing within 18 months without relief from intractable staff shortages, untamed inflation, and the treachery of a Medicare/Medicaid Rate Construction Crisis. This is how the stage is set for Comprehensive Reform, a plan so grand it needs no luck or logic. z-INTEL’s mission is to support development of rational payment policy with meaningful data, because from our perspective…