CMS Refines Proposed Changes To SNF Reimbursement

July 26, 2018, Long-Term Care Newsletter - On May 8, 2018, in its Fiscal Year 2019 (FY 2019) Skilled Nursing Facility (SNF) Prospective Payment System (PPS) Proposed Rule, Centers for Medicare & Medicaid Services (CMS) proposed a new case-mix classification system that drives Medicare reimbursement for SNF inpatient services.  The new system would go into effect on October 1, 2019. 

CMS states that the new system, called the Patient-Driven Payment Model (PDPM), will "better account for resident characteristics and care needs while reducing both systemic and administrative complexity."  Additional refinements are expected when the final PPS rule is issued later this year.

The new proposal reflects the latest chapter in a long-running discussion of SNF case-mix reimbursement models.  In May 2017, CMS released an Advance Notice of Proposed Rulemaking (ANPRM) introducing a new approach to case-mix classification and reimbursement in SNFs.  The proposed Resident Classification System-1 (RCS-1) focused on resident characteristics to determine case-mix classification rather than the existing Resource Utilization Groups (RUGs). 

Under the RUG system, therapy minutes provided to residents are the primary driver of reimbursement rates.  Like RCS-1, the PDPM endeavors "to remove, to the extent possible, service-based metrics from the SNF PPS and derive payment from verifiable characteristics that are patient, and not facility, centered."  The PDPM includes modifications to the RCS-1 proposal in response to comments on the May ANPRM.

Key Features

The PDPM significantly restructures the reimbursement methodology for SNF services.  Nowhere will the proposed changes be more acutely felt than in the provision of therapy services.  As CMS notes in the proposed rule, more than 90 percent of covered Part A SNF days are paid under a rehabilitation RUG. 

CMS estimates that the proposed system would decrease reimbursement for patients typically receiving therapy at an "ultra-high" rehabilitation RUG by 8.4 percent while increasing reimbursement for patients typically categorized into non-rehabilitation RUGs by 50.5 percent.  Overall, CMS projects that facilities that typically bill 50 percent or more of their total patient days at an ultra-high RUG could see a decrease in total reimbursement of up to 9.8 percent (depending on volume).

Facilities will also find that, under the proposed system, assistance with activities of daily living (ADLs) are a key component of reimbursement.  Functional status, as reflected in the degree of ADL assistance required, will directly generate a component of the case-mix index (CMI) for three separate CMIs (physical therapy, occupational therapy, and nursing). 

Clear, accurate documentation of the assessment of a resident's needs with regard to ADL assistance will therefore be a critical component necessary to support reimbursement claims, as will documentation of ADL assistance actually provided.

Finally, the proposed system continues to increase pressure to shorten SNF stays.  The therapy case-mix components would be subject to a downward adjustment after a resident's 20th day in the facility; reductions would continue gradually throughout the stay. 

Because, as described in more detail below, CMS believes that therapy needs typically decrease over the course of a resident's stay, and since the per diem rate will be generated by a single MDS assessment, CMS has proposed the automatic payment adjustment factors to reflect projected decreases.

PDMP Details

Reimbursement Methodology. Reimbursement for a resident's SNF stay under the proposed system would be generated by multiplying the CMI for each case-mix group (described in more detail below) by the federal base rate for that component.  The payment rate for each group would be adjusted based on the variable per diem adjustment schedule and applicable multipliers.  The resulting amount would be added to the non-case-mix component payment rate to generate the per diem rate.

Case-Mix Components. Similar to the approach proposed under RCS-1, PDPM proposes three components of the reimbursement: therapy case-mix, nursing case-mix, and non-case-mix.  PDMP separates the therapy case-mix component into a physical therapy (PT) case-mix component, an occupational therapy (OT) case-mix component, and speech language pathology (SLP) case-mix component. 

This is a departure from RCS-1, which proposed combining PT and OT into a single case-mix component.  As proposed under RCS-1, the nursing case-mix component would consist of nursing case-mix and non-therapy ancillary (NTA) case-mix components.  As under the current RUG system, the non-case-mix component would be calculated to cover expenses that do not vary by patient, such as room and board and capital expenses.

Therapy Case-Mix. CMS identifies three primary predictors of resident PT and OT costs: (1) The clinical reason for the SNF stay; (2) The resident's functional status; and (3) The resident's cognitive status.  Only the first two are used to create case-mix classifications for the PT and OT components (which, despite generating a different CMI for payment purposes, are combined into 16 PT/OT case-mix groups).  For SLP, primary predictors are: (1) The clinical reason for the SNF stay; (2) The presence of a swallowing disorder or mechanically altered diet; and (3) The presence of an SLP-related comorbidity or cognitive impairment.  These would generate 12 proposed case-mix groups.

PT and OT Case-Mix

CMS proposes to classify SNF residents into a PDPM clinical category based on the ICD-10-CM code reported on the MDS as the primary reason for the resident's SNF stay.  The code would be reported on the first line of Section I8000 of the MDS, with the ICD-10-PCS code corresponding to a potential inpatient procedure being reported on the second line of that section (where applicable) to ensure proper resident categorization.  As an alternative, CMS proposes categorizing residents based on a primary diagnosis selected from a pull-down menu in Section I00020 of the MDS.

CMS proposes four major clinical categories for purposes of clinical classification for PT and OT: (1) Major Joint Replacement or Spinal Surgery; (2) Non-Orthopedic Surgery and Acute Neurologic; (3) Other Orthopedic; and (4) Medical Management.  This is a refinement of the five categories proposed in the ANPRM, which included separate categories for Non-Orthopedic Surgery and Acute Neurologic.  A resident would first be assigned to a clinical category and then to a case-mix group based on the resident's functional score as calculated based on the measures described below.

CMS would determine functional status based on measures of both early-loss and late-loss ADLs.  These would be captured in Section GG of the MDS (Functional Abilities and Goals).  Late-loss ADLs measured would include bed mobility, transfer, eating, and toileting, as proposed in the ANPRM as part of RCS-1.

Early-loss ADLs would include oral hygiene and walking.  CMS proposes changing scoring methodologies for these items to reflect the resident's degree of independence rather than his or her degree of dependence.  More independent residents would receive higher, not lower, functional ability scores.  Activities currently coded as "did not occur" or "occurred only once" would be scored as dependent under the proposed system.  Scores on related items (e.g., measures related to bed mobility or transfer) would be averaged.

SLP Case-Mix

To generate the SLP case-mix grouping, residents would be classified into one of two clinical categories – either Acute Neurologic or Non-Neurologic.  The Non-Neurologic group would include residents categorized into any of the other clinical categories identified for PT and OT case-mix grouping and residents classified into the Non-Orthopedic Surgery category.  Residents would also be categorized based on the presence or absence of SLP-related comorbidity or cognitive impairment.  Based on the presence or absence of these factors (none present, any one present, any two present, or any three present), residents would be further categorized based on their need for a mechanically-altered diet or the presence or absence of a swallowing disorder (neither, either, or both) to determine their SLP case-mix group and the associated CMI.

Nursing Case-Mix

  • Nursing Case-MixLike RCS-1, PDPM proposes to use a resident's nursing case-mix classification to directly generate a component of reimbursement.  Under the RUG system, nursing case-mix was only determinative of reimbursement if a resident was categorized into a non-rehabilitation RUG (i.e., was not receiving therapy services). CMS proposes to utilize nursing case-mix groups that are similar to those categories used now for resident classification.  CMS does propose to collapse certain case-mix groups based on ADL score bins, reducing the number of groups from 43 to 25.  ADL scores would be generated based on functional status measured as described above in Section GG of the MDS (with respect to late-loss ADLs only).  A multiplier would be used for residents with HIV/AIDS to reflect increased resource intensity for those patients.
  • Non-Therapy Ancillary (NTA) Case-Mix. CMS identified (1) Resident comorbidities; (2) Use of extensive services; and (3) Resident age as predictive of high NTA utilization.  Due to stakeholder concerns that including age as a factor in CMI generation could create access impediments for individuals in certain age groups, CMS did not consider age in generating case-mix groups or CMIs.  CMS proposed a list of comorbidities and services shown by its analysis to generate high NTA spending and assigned a score to each service and comorbidity.  A resident's total score would be used to classify the resident into a case-mix group.  The new approach was developed in response to ANPRM commenters concerned that NTA measures might not be sufficiently robust.

Changes to Cognitive Evaluations. CMS proposes to replace the Brief Interview for Mental Status and Cognitive Performance Scale with an integrated assessment applicable to all residents: the Cognitive Function Scale.  Scores on this scale will be a factor in generating the SLP CMI.  The presence of behavioral or cognitive issues is also measured to generate the nursing CMI.

Per Diem Adjustment Factor. To reflect changes in resident resource needs throughout a SNF stay, CMS proposes a per diem adjustment factor based on the length of a resident's stay.  The CMI for PT and OT would be adjusted downward gradually throughout a resident's stay (starting at 1.0 for days 1 – 20 and gradually decreasing to as low as 0.76 for days 98 – 100 of a SNF stay), whereas the NTA CMI component would be adjusted by 3.0 for days 1 – 3 of a resident's stay and return to 1.0 for the remainder.

MDS Assessments. Because the proposed changes to the SNF case-mix classification system tie reimbursement to factors unlikely to change throughout a resident's stay,  CMS proposes to eliminate most scheduled and unscheduled PPS assessments.  Case-mix classification would be based on the five-day assessment, and updated assessments would only be required in the event of a change in "first tier classification criteria" expected to persist for more than 14 days. 

CMS also proposes to incorporate "grace days" into the assessment window and, correspondingly, to eliminate existing "grace days."  Finally, a Discharge Assessment would also be required to allow CMS to monitor therapy utilization.

Therapy Limits. Under PDMP, CMS proposes to limit the combined amount of group and concurrent therapy provided to a resident to 25 percent of the total number of therapy minutes provided in each discipline.  However, therapy minutes would no longer be allocated across residents because therapy minutes would no longer be used to determine reimbursement rates.

The PDPM represents a total restructuring of the SNF reimbursement system.  Long term care providers and therapy providers should review the proposed rule carefully.  If you have questions, or wish to submit a comment, contact Meredith Larson, Donna Thiel or your usual Baker Donelson attorney for assistance.  Further refinements to the PDPM system are to be expected prior the October 1, 2019 implementation date, and provider input will be critical to shaping the final model that is implemented.