The Patient Driven Payment Model is a rare bright spot in Skilled Nursing’s dreary reimbursement landscape, but storm clouds are on the horizon. CMS has proposed recalibrating SNF rates to correct for unintentional spending relative to the RUG system it replaced. The agency refers to this reduction as a “Parity Adjustment,” but in the context of a post-pandemic operating environment, the 2023 proposal reads more like a “Reimbursement Parody” to me.
“Budget Neutral” and “Parity” are not mathematically synonymous; the nuances of recalibration must be respected. CMS purportedly removed COVID & 1135 Waiver distortion from its calculation in last year’s Rule, but now recognize that effort failed to isolate all such claims. CMS expanded exclusion criteria for 2023, but we believe their process still does not fully neutralize for Public Health Emergency variation. Either way, SNFs are facing a 4.6% rate reduction effective October 1, 2022, or about $30 per patient day on average. While the 2023 Market Basket update offsets much of the net impact (down to about $5 or $6 PPD), it does nothing to lessen the philosophical sting of recalibration. All that said, there is much more to the PDPM story.
Recalibration is not the only change to rate construction; keep in mind that other modifications impact providers differently. Area Wage Index updates create the typical array of winners & losers (downside is now capped), while fine-tuning the labor ratio means that lower cost counties are hit a bit harder than regions with relatively high wage adjustment. For example, Monmouth County, NJ is gifted a $10 increase, but Salem County, NJ drops roughly $30 PPD.
PDPM is the fourth iteration of Skilled Nursing’s Prospective Payment System. There were also six intermittent tweaks to the respective RUG-III / RUG-IV models and two systemic “false-starts.” Alas, the RUG system was beyond repair; even in its final form, RUGs lacked the statistical sensitivity to appropriately calibrate payment to acuity.
PDPM delivered on CMS’ promise of a more rational reimbursement system, most notably by eliminating artificial therapy targets and the perpetual cycle of “Ultra High” audits & recoupment that followed. It also arrived “just in time,” as RUG-IV would have been an unmitigated disaster for SNFs through COVID, with rates $200/day lower than the prior year. There is simply no way to reconcile pre-pandemic payment drivers to clinical and financial realities of the Public Health Emergency. This irregularity alone should abrogate using RUG payments as a comparative parity measure. Nevertheless, CMS’ recalibration is based on hypothetical spending had RUG-IV classification continued unabated.
Before we drill into specifics, consider the overriding logic of this reimbursement system transition. The RUG model was universally derided for its limited predictive power, perverse treatment incentives and payment-driver incongruities. PDPM’s intent was to shift rate-sensitivity away from therapy volume to distinct medical conditions and services; an absolute necessity as SNF acuity rises.
COVID changed everything. Recalibration is now effectively CMS intervening to prohibit the system from working as intended. Capping aggregate PDPM payment at the amount paid under ill-conceived RUGs is apocryphal by any measure. If CMS was committed to paying rates commensurate with patient-specific acuity, how can PDPM aggregate reimbursement equal the amount paid under a system with little correlation to acuity? It makes no sense; why bother making the change?
As we will explore in the coming months, revenue distribution is the greatest threat to SNF stability; tinkering with the one rational methodology is at best counterproductive, at worst dangerous.
Recalibration
CMS’ recalibration methodology is surprisingly simple, but widely misunderstood. Total Medicare spending is not the benchmark; in fact, since the RUG-IV recalibration in 2012, aggregate Medicare SNF spending has remained flat. Average rates increase each year, but lower fee-for-service utilization and growing Medicare Advantage penetration (a ratio we call “ FFS Attrition Rate”) offset inflation updates.
Purported “overpayments” are based on the average gross PDPM rate per day, benchmarked against a modeled RUG-IV per diem. “Gross rate” means offsets including copayments, sequestration and value-based adjustment are not applied. In essence, it comes down to the average rate paid per covered day – as if nothing else matters.
CORE Analytics has extensively modeled hundreds of thousands of PDPM claims in the context of CMS’ recalibration methodology. Our database includes more than 3,500 SNFs in 49 states; half our users have already uploaded March 2022 claims.
We conclude that CMS’ approach to budget neutrality overstates the impact of provider behavior on rate escalation. Consistent with the system’s design, rate variation is fully explained by the following pandemic-induced factors that must be applied to CMS’ recalibration equation to ensure equitable payment for SNF services.
See the CORE Reporter for info about our analysis and PDPM capture trends through 2021.
Share of nursing home residents with depression Anxiety and depression
may be exacerbated by fear, worry, or social isolation due to COVID19… In Washington, where media attention has been centered on the outbreak of coronavirus in nursing facilities, almost half of residents have experienced depression or depressive symptoms. Research on family involvement in long-term care has shown that family visitation can have potentially positive effects on cognitive and behavioral health diagnoses.1 Thus, visitor restrictions… may also have negative impacts on residents’ mental health and increase the incidence of depressive symptoms.2
Certain conditions are difficult to compare across payment systems. As detailed in the CORE Reporter, we believe CMS inappositely calculated the RUG benchmark by assuming Depression is equally distributed among all acuity-profiles.
CORE’s analysis takes no liberties with the data; our approach is the most accurate method of measurement. We compare “apples-to-apples” by including only Depression-sensitive days under both PDPM and RUG-IV because the discussion is specific to reimbursement, not overall accuracy of assessment (if it were the latter, there could be no justification for using RUG scores at all!). As such, apples-to-apples, PDPM Depression increased by only ~ two percentage points during the first three months after implementation. Then the world changed.
CMS ignores the reality that every SNF resident was impacted by the pandemic; removing COVID & Waiver claims made no one happier to be shut off from their families. Excluding COVID & Waiver claims actually inflates the recalibration. ES and Behavior/Physical rates are not impacted by Depression and historically have lower Depression capture rates, yet these are scores most likely to be COVID & Waiver related. As a result, CMS’ PDPM benchmark was further inflated.
When limited to Depression sensitive scores only, capture was 12% during October 1 – December 31, 2019, yet finished FY 2020 at 35.5% before dropping to 34.8% in 2021. CORE’s data through March 2022 supports this downward trend as SNFs reopened to visitation. Depression did not “creep,” it came along for the ride
2. Extensive Services & NTA Isolation
Only 1.88% of days were classified Extensive Services in the 2019 RUG baseline (including Rehab-X/L), yet after excluding CMS’ “high COVID months,” COVID Dx & 1135 Waiver claims, ES increased to 3.15% in 2020 and 3.35% in 2021. Targeted reimbursement strategy does not explain the PDPM growth. ES is driven by high-acuity clinical qualifiers impossible to overlook and not obfuscated by Rehab under RUG-IV; in fact, adding an ES qualifier to Rehab RUGs added up to $200/day. In other words, “you don’t creep your way to Extensive Services.” Additionally, Isolation adds one point to the NTA component, overstating PDPM rates for many claims that should be removed from the recalibration benchmark.
CMS’ subset of Extensive Services benchmark claims were 80% above baseline for two reasons: COVID-related days were not removed from their calculation, and/or nonCOVID acuity increased overall; either way, SNFs are unfairly penalized. Moreover, CMS seeks input regarding expanded Isolation capture criteria. This suggests tacit acknowledgment that Medicare underpaid SNFs for patients who were Isolated per CDC guidelines but did not satisfy RAI rules for MDS capture
3. Higher Patient Acuity
As homecare adapted to new levels of medical complexity, SNF acuity increased as well, irrespective of COVID status. This elevation is difficult to quantify, so CORE created a proprietary Risk Adjustment Factor using data contained on the Medicare UB-04, which allows for more consistent measurement than the MDS. Initial analysis of “Z-RAF” scores indicates a sharp acuity increase beginning April, 2020. CMS should apply its own claims-based CMS-HCC scale (as opposed to the MDS), to quantify how patient acuity impacted PDPM reimbursement (as intended).
4. Entrapment!
CMMI gave us ACOs & BPCI. More accurately, CMS created shared-savings opportunities for health-systems at the expense of SNFs. SNF were told CMMI would reward quality providers as “valued downstream partners.” In reality, “valued downstream partner” was code for “primary target for cost reduction” and the result was shorter SNF Medicare stays. Overall ALOS may not have decreased during the pandemic, but CMMI unequivocally condenses SNF episodes. Because PDPM payment is “frontloaded” (NTA rates are tripled for days 1 – 3), average rates increase for each day removed from a stay. CMMI episodes are non-comparable against historical RUGs. Ignoring this effect inflates the PDPM benchmark and increases recalibration and will continue to distort the benchmark every year. I may have watched My Cousin Vinny a few too many times, but this feels like a set up to me.
5. The “Soft Landing”
Patients covered by Medicare on October 1, 2019 (PDPM’s start date) were treated as new admissions for purposes of day weight adjustment, regardless of how many days remained in their benefit period. This one-time allowance was a pragmatic way for CMS to offset the provider burden of system transition. As a matter of principle, transition assistance should be exempt from any comparative spending analysis.
6. Hospital TCUs & TCU Transfers
Hospital-Based Transitional Care Units are designed for relatively short stays, making TCUs disproportionate winners under PDPM. These dollars are effectively redistributed from Freestanding SNF to hospital systems. The recalibration impact is exacerbated when patients are discharged from TCUs and sent to Freestanding SNFs for continued care. “Double-dipping” by hospitals not only turbocharges NTA payments but shortens overall ALOS (increasing PDPM rates against the RUG benchmark). TCUs should therefore be removed from the recalibration calculation.
7. Savings to Other Payers
Financially, Medicare Advantage added insult to injury. Traditional Plans offered no relief as they subtly added to the Medicaid burden yet receive an 8.5% premium increase concurrent to the SNF payment reduction. Meanwhile, operator sponsored ISNPs, ideally positioned for clinical support, were excluded from CARES funding, then saw enrollment hobbled by unintended 1135 Waiver incentives.
Irrespective of how many 1135 Waiver claims CMS removed from its analysis, the allinclusive nature of Medicare Part A reimbursement mitigated some measure of Medicare Part B & D spending, but Medicaid program savings were significant. Per the 2022 MedPAC Report to Congress, “(Medicaid SNF) Spending was $39.8 billion in 2020, 3.8 percent less than in 2019.” This $1.57B decrease was caused by the pandemic, not PDPM, and the aggregate value of Medicaid dollars not spent on long-term care is eerily close to the $1.7B SNF Medicare “Parody Adjustment.”
Based on publicly available information, 27 states issued emergency Medicaid SNF funding as of October 2021, without which aggregate savings would have been even greater. So, where did this leave the 23 states that received no such relief? Despite CMS’ authority over both programs and the extreme variability across markets, all SNFs are subject to the same reimbursement policy. But that’s a story for another time.
ZHSG, CORE Analytics & z-INTEL are exploring solutions to the fractured SNF revenue cycle. As presently constructed, forthcoming reform initiatives cannot be sustained by the current SNF financial model. Even with additional funding, irrational distribution across the provider community will only serve to further destabilize the system. It’s time to have an honest discussion about SNF economics.
Skilled Nursing Facilities are relying on CMS to base policy on sound data, but we have ample reason to question the agency’s approach. z-INTEL is dedicated to balancing the SNF data equation.
Marc Zimmet is CEO of Zimmet Healthcare Services Group, Z-CORE Analytics, & Cofounder, z-INTEL.
2023 SNF PPS Recalibration Adjustment
Marc Zimmet
The Patient Driven Payment Model is a rare bright spot in Skilled Nursing’s dreary reimbursement landscape, but storm clouds are on the horizon. CMS has proposed recalibrating SNF rates to correct for unintentional spending relative to the RUG system it replaced. The agency refers to this reduction as a “Parity Adjustment,” but in the context of a post-pandemic operating environment, the 2023 proposal reads more like a “Reimbursement Parody” to me.
“Budget Neutral” and “Parity” are not mathematically synonymous; the nuances of recalibration must be respected. CMS purportedly removed COVID & 1135 Waiver distortion from its calculation in last year’s Rule, but now recognize that effort failed to isolate all such claims. CMS expanded exclusion criteria for 2023, but we believe their process still does not fully neutralize for Public Health Emergency variation. Either way, SNFs are facing a 4.6% rate reduction effective October 1, 2022, or about $30 per patient day on average. While the 2023 Market Basket update offsets much of the net impact (down to about $5 or $6 PPD), it does nothing to lessen the philosophical sting of recalibration. All that said, there is much more to the PDPM story.
Recalibration is not the only change to rate construction; keep in mind that other modifications impact providers differently. Area Wage Index updates create the typical array of winners & losers (downside is now capped), while fine-tuning the labor ratio means that lower cost counties are hit a bit harder than regions with relatively high wage adjustment. For example, Monmouth County, NJ is gifted a $10 increase, but Salem County, NJ drops roughly $30 PPD.
Model your 2023 rates using z-INTEL’s PDPM Rate Simulator here.
No More “Life on the RUG”
PDPM is the fourth iteration of Skilled Nursing’s Prospective Payment System. There were also six intermittent tweaks to the respective RUG-III / RUG-IV models and two systemic “false-starts.” Alas, the RUG system was beyond repair; even in its final form, RUGs lacked the statistical sensitivity to appropriately calibrate payment to acuity.
PDPM delivered on CMS’ promise of a more rational reimbursement system, most notably by eliminating artificial therapy targets and the perpetual cycle of “Ultra High” audits & recoupment that followed. It also arrived “just in time,” as RUG-IV would have been an unmitigated disaster for SNFs through COVID, with rates $200/day lower than the prior year. There is simply no way to reconcile pre-pandemic payment drivers to clinical and financial realities of the Public Health Emergency. This irregularity alone should abrogate using RUG payments as a comparative parity measure. Nevertheless, CMS’ recalibration is based on hypothetical spending had RUG-IV classification continued unabated.
Before we drill into specifics, consider the overriding logic of this reimbursement system transition. The RUG model was universally derided for its limited predictive power, perverse treatment incentives and payment-driver incongruities. PDPM’s intent was to shift rate-sensitivity away from therapy volume to distinct medical conditions and services; an absolute necessity as SNF acuity rises.
COVID changed everything. Recalibration is now effectively CMS intervening to prohibit the system from working as intended. Capping aggregate PDPM payment at the amount paid under ill-conceived RUGs is apocryphal by any measure. If CMS was committed to paying rates commensurate with patient-specific acuity, how can PDPM aggregate reimbursement equal the amount paid under a system with little correlation to acuity? It makes no sense; why bother making the change?
As we will explore in the coming months, revenue distribution is the greatest threat to SNF stability; tinkering with the one rational methodology is at best counterproductive, at worst dangerous.
Recalibration
CMS’ recalibration methodology is surprisingly simple, but widely misunderstood. Total Medicare spending is not the benchmark; in fact, since the RUG-IV recalibration in 2012, aggregate Medicare SNF spending has remained flat. Average rates increase each year, but lower fee-for-service utilization and growing Medicare Advantage penetration (a ratio we call “ FFS Attrition Rate”) offset inflation updates.
Purported “overpayments” are based on the average gross PDPM rate per day, benchmarked against a modeled RUG-IV per diem. “Gross rate” means offsets including copayments, sequestration and value-based adjustment are not applied. In essence, it comes down to the average rate paid per covered day – as if nothing else matters.
CORE Analytics has extensively modeled hundreds of thousands of PDPM claims in the context of CMS’ recalibration methodology. Our database includes more than 3,500 SNFs in 49 states; half our users have already uploaded March 2022 claims.
We conclude that CMS’ approach to budget neutrality overstates the impact of provider behavior on rate escalation. Consistent with the system’s design, rate variation is fully explained by the following pandemic-induced factors that must be applied to CMS’ recalibration equation to ensure equitable payment for SNF services.
See the CORE Reporter for info about our analysis and PDPM capture trends through 2021.
1. Depression
CMS attributes elevated PHQ scoring to provider strategy, yet the science says otherwise, as explained by Kaiser Family Foundation:
Share of nursing home residents with depression Anxiety and depression
may be exacerbated by fear, worry, or social isolation due to COVID19… In Washington, where media attention has been centered on the outbreak of coronavirus in nursing facilities, almost half of residents have experienced depression or depressive symptoms. Research on family involvement in long-term care has shown that family visitation can have potentially positive effects on cognitive and behavioral health diagnoses.1 Thus, visitor restrictions… may also have negative impacts on residents’ mental health and increase the incidence of depressive symptoms.2
Certain conditions are difficult to compare across payment systems. As detailed in the CORE Reporter, we believe CMS inappositely calculated the RUG benchmark by assuming Depression is equally distributed among all acuity-profiles.
CORE’s analysis takes no liberties with the data; our approach is the most accurate method of measurement. We compare “apples-to-apples” by including only Depression-sensitive days under both PDPM and RUG-IV because the discussion is specific to reimbursement, not overall accuracy of assessment (if it were the latter, there could be no justification for using RUG scores at all!). As such, apples-to-apples, PDPM Depression increased by only ~ two percentage points during the first three months after implementation. Then the world changed.
CMS ignores the reality that every SNF resident was impacted by the pandemic; removing COVID & Waiver claims made no one happier to be shut off from their families. Excluding COVID & Waiver claims actually inflates the recalibration. ES and Behavior/Physical rates are not impacted by Depression and historically have lower Depression capture rates, yet these are scores most likely to be COVID & Waiver related. As a result, CMS’ PDPM benchmark was further inflated.
When limited to Depression sensitive scores only, capture was 12% during October 1 – December 31, 2019, yet finished FY 2020 at 35.5% before dropping to 34.8% in 2021. CORE’s data through March 2022 supports this downward trend as SNFs reopened to visitation. Depression did not “creep,” it came along for the ride
2. Extensive Services & NTA Isolation
Only 1.88% of days were classified Extensive Services in the 2019 RUG baseline (including Rehab-X/L), yet after excluding CMS’ “high COVID months,” COVID Dx & 1135 Waiver claims, ES increased to 3.15% in 2020 and 3.35% in 2021. Targeted reimbursement strategy does not explain the PDPM growth. ES is driven by high-acuity clinical qualifiers impossible to overlook and not obfuscated by Rehab under RUG-IV; in fact, adding an ES qualifier to Rehab RUGs added up to $200/day. In other words, “you don’t creep your way to Extensive Services.” Additionally, Isolation adds one point to the NTA component, overstating PDPM rates for many claims that should be removed from the recalibration benchmark.
CMS’ subset of Extensive Services benchmark claims were 80% above baseline for two reasons: COVID-related days were not removed from their calculation, and/or nonCOVID acuity increased overall; either way, SNFs are unfairly penalized. Moreover, CMS seeks input regarding expanded Isolation capture criteria. This suggests tacit acknowledgment that Medicare underpaid SNFs for patients who were Isolated per CDC guidelines but did not satisfy RAI rules for MDS capture
3. Higher Patient Acuity
As homecare adapted to new levels of medical complexity, SNF acuity increased as well, irrespective of COVID status. This elevation is difficult to quantify, so CORE created a proprietary Risk Adjustment Factor using data contained on the Medicare UB-04, which allows for more consistent measurement than the MDS. Initial analysis of “Z-RAF” scores indicates a sharp acuity increase beginning April, 2020. CMS should apply its own claims-based CMS-HCC scale (as opposed to the MDS), to quantify how patient acuity impacted PDPM reimbursement (as intended).
4. Entrapment!
CMMI gave us ACOs & BPCI. More accurately, CMS created shared-savings opportunities for health-systems at the expense of SNFs. SNF were told CMMI would reward quality providers as “valued downstream partners.” In reality, “valued downstream partner” was code for “primary target for cost reduction” and the result was shorter SNF Medicare stays. Overall ALOS may not have decreased during the pandemic, but CMMI unequivocally condenses SNF episodes. Because PDPM payment is “frontloaded” (NTA rates are tripled for days 1 – 3), average rates increase for each day removed from a stay. CMMI episodes are non-comparable against historical RUGs. Ignoring this effect inflates the PDPM benchmark and increases recalibration and will continue to distort the benchmark every year. I may have watched My Cousin Vinny a few too many times, but this feels like a set up to me.
5. The “Soft Landing”
Patients covered by Medicare on October 1, 2019 (PDPM’s start date) were treated as new admissions for purposes of day weight adjustment, regardless of how many days remained in their benefit period. This one-time allowance was a pragmatic way for CMS to offset the provider burden of system transition. As a matter of principle, transition assistance should be exempt from any comparative spending analysis.
6. Hospital TCUs & TCU Transfers
Hospital-Based Transitional Care Units are designed for relatively short stays, making TCUs disproportionate winners under PDPM. These dollars are effectively redistributed from Freestanding SNF to hospital systems. The recalibration impact is exacerbated when patients are discharged from TCUs and sent to Freestanding SNFs for continued care. “Double-dipping” by hospitals not only turbocharges NTA payments but shortens overall ALOS (increasing PDPM rates against the RUG benchmark). TCUs should therefore be removed from the recalibration calculation.
7. Savings to Other Payers
Financially, Medicare Advantage added insult to injury. Traditional Plans offered no relief as they subtly added to the Medicaid burden yet receive an 8.5% premium increase concurrent to the SNF payment reduction. Meanwhile, operator sponsored ISNPs, ideally positioned for clinical support, were excluded from CARES funding, then saw enrollment hobbled by unintended 1135 Waiver incentives.
Irrespective of how many 1135 Waiver claims CMS removed from its analysis, the allinclusive nature of Medicare Part A reimbursement mitigated some measure of Medicare Part B & D spending, but Medicaid program savings were significant. Per the 2022 MedPAC Report to Congress, “(Medicaid SNF) Spending was $39.8 billion in 2020, 3.8 percent less than in 2019.” This $1.57B decrease was caused by the pandemic, not PDPM, and the aggregate value of Medicaid dollars not spent on long-term care is eerily close to the $1.7B SNF Medicare “Parody Adjustment.”
Based on publicly available information, 27 states issued emergency Medicaid SNF funding as of October 2021, without which aggregate savings would have been even greater. So, where did this leave the 23 states that received no such relief? Despite CMS’ authority over both programs and the extreme variability across markets, all SNFs are subject to the same reimbursement policy. But that’s a story for another time.
ZHSG, CORE Analytics & z-INTEL are exploring solutions to the fractured SNF revenue cycle. As presently constructed, forthcoming reform initiatives cannot be sustained by the current SNF financial model. Even with additional funding, irrational distribution across the provider community will only serve to further destabilize the system. It’s time to have an honest discussion about SNF economics.
Skilled Nursing Facilities are relying on CMS to base policy on sound data, but we have ample reason to question the agency’s approach. z-INTEL is dedicated to balancing the SNF data equation.
Marc Zimmet is CEO of Zimmet Healthcare Services Group, Z-CORE Analytics, & Cofounder, z-INTEL.