CMS-1622-P: Medicare Program; Prospective Payment System and Consolidated Billing for Skilled Nursing Facilities (SNFs) for FY 2016, SNF Value-Based Purchasing Program, SNF Quality Reporting Program, and Staffing Data Collection
Submitted electronically, http://www.regulations.gov
June 19, 2015
The Center for Medicare Advocacy (Center) submits the following comments on the proposed rules, 80 Fed. Reg. 22043 (April 20, 2015).
First, we oppose increasing Medicare rates paid to skilled nursing facilities (SNFs). The Medicare Payment Advisory Commission (MedPAC) has reported that SNFs are overpaid by the Medicare program and have enjoyed Medicare margins exceeding 10% for more than 14 consecutive years. MedPAC recommends no increase in the rate for FY 2016. MedPAC, Comments on CMS-1622-P (May 19, 2015), http://medpac.gov/documents/contractor-reports/medpac-comment-on-cms's-skilled-nursing-facilities-prospective-payment-system-value-based-purchased-and-quality-reporting-proposed-rule.pdf?sfvrsn=0; MedPAC, Report to the Congress: Medicare Payment Policy, Chapter (skilled nursing facility services), p. 302, (May 2015), http://medpac.gov/documents/reports/chapter-8-skilled-nursing-facility-services-(march-2015-report).pdf?sfvrsn=0.
Second, we urge the Centers for Medicare & Medicaid Services (CMS) to develop a nursing home-specific wage index. Medicare overpays SNFs by using the hospital wage index. We have made this recommendation before.
I.SNF Quality Reporting Program
The Center is pleased that CMS is moving forward with implementation of the IMPACT Act, which calls for the collection and reporting of uniform data for various post-acute providers. Comparisons among post-acute providers – who their patients are, what their patient outcomes are – will be strengthened by the use of uniform assessment data. The measures that CMS is proposing to use – pressure ulcers (short-stay), falls (long stay), and cognitive function – are all important issues for post-acute patients. However, we are concerned that all three measures are based on self-reported data and that SNFs do not currently report discharge data for Medicare residents who remain in the facility after their Medicare coverage ends, approximately 30% of Medicare residents.
The Center understands that uniform assessment data for post-acute providers are required by the IMPACT Act. However, we are concerned that CMS is not simultaneously taking strong steps to ensure, to the extent possible, that the data submitted by nursing facilities are accurate. This industry has a very poor record of doing assessments.
Self-reported MDS data are unreliable and subject to gaming
In August 2014, the New York Times reported that 80% of facilities submitted resident assessment data to the Centers for Medicare & Medicaid (CMS) that yielded four and five stars on their Quality Measures (QMs), artificially and inaccurately boosting many facilities’ overall ratings as well as the QM ratings. Katie Thomas, Ratings Allow Nursing Homes To Game System; Medicare’s Five Stars: Data Taken at Face Value Often Fails to Reflect Real Conditions,” The New York Times (Aug. 25, 2014), http://www.nytimes.com/2014/08/25/business/medicare-star-ratings-allow-nursing-homes-to-game-the-system.html?module=Search&mabReward=relbias%3As%2C%7B%221%22%3A%22RI%3A6%22%7D&_r=0/ In response to The New York Times’s investigative report on gaming by nursing facilities, the White House and CMS announced changes to Nursing Home Compare in October 2014. The White House stressed that the new scoring methodology “will place more emphasis on data that is verified by independent sources rather than data that is self-reported by nursing homes,” White House, “FACT SHEET: Administration Announces New Executive Actions to Improve Quality of Care for Medicare Beneficiaries” (Oct. 6, 2014), http://www.whitehouse.gov/the-press-office/2014/10/06/fact-sheet-administration-announces-new-executive-actions-improve-qualit., a statement reiterated by CMS in its pledge, “Future additional measures will include claims-based data on re-hospitalization and community discharge rates.” CMS, “CMS Announces Two Medicare Quality Improvement Initiatives” (Press Release, Oct. 6, 2014), http://www.cms.gov/Newsroom/MediaReleaseDatabase/Press-releases/2014-Press-releases-items/2014-10-06.html.
Concerns about the inaccuracy of resident assessment data are not new. See Office of Inspector General, Nursing Facility Assessments and Care Plans for Residents Receiving Atypical Antipsychotic Drugs, OEI-07-08-00151 (July 2012) (finding one third of assessments reviewed failed to meet one or more requirements for resident assessments) http://oig.hhs.gov/oei/reports/oei-07-08-00151.pdf Data Assessment and Verification (DaVE) project to assess validity of resident assessment information, http://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/NursingHomeQualityInits/NHQIMDS20.html (CMS project to improve the accuracy of assessment data). GAO, Nursing Homes: Federal Efforts to Monitor Resident Assessment Data Should Complement State Activities, GAO-02-279 (Feb. 2002) (discussing federal and state efforts to improve the accuracy of assessment data), http://gao.gov/assets/240/233704.pdf.
The Center has several comments about the proposed rule:
1. The draft measures perpetuate many problems that we have been critical of before in QMs – including the exclusion of residents with missing data. Such exclusions present a perfect opportunity for facility gaming.
If a facility recognizes that a resident is declining, it can simply omit some data for that resident, ensuring that the resident is excluded from the QM. The New York Times’s investigative report cited above confirms that gaming with MDS data is a serious and widespread problem. See also Center, “The Worst-Performing Nursing Facilities Are Seldom Sanctioned; Self-Reporting is Not an Accurate Quality Measurement” (Weekly Alert, Jan. 24, 2013), http://www.medicareadvocacy.org/the-worst-performing-nursing-facilities-are-seldom-sanctioned-self-reporting-is-not-an-accurate-quality-measurement/ (documenting that Special Focus Facilities have high star ratings in self-reported QMs); Center, “Debunking Nursing Home Myths about Quality of Care and Enforcement of Federal Standards of Care” (Weekly Alert, Sep. 12, 2013), http://www.medicareadvocacy.org/debunking-nursing-home-myths-about-quality-of-care-and-enforcement-of-federal-care-standards/ (documenting high star ratings in QMs for the lowest-tier facilities in three states, Georgia, Illinois, and Oregon).
2. The Center is concerned about the limited data CMS is proposing to collect for the SNF Quality Reporting Program. CMS is proposing, at 75 Fed. Reg., 22076,:
- to require the collection of data for a single quarter in FY 2018.
- to give facilities 5½ months to submit or correct their quarterly data.
- initially, to require facilities to submit 80% of their data.
These proposals will promote gaming by nursing homes. Facilities will be found to be in compliance if they submit just 80% of their data. They will have substantial time to review the data – 5½ months – in order to decide which data to submit and which data to exclude. This plan will fail to ensure that data are accurate. CMS needs to collect more data more quickly.
3. CMS asks for the identification of elements to use in validating the data that SNFs submit. The Center agrees that validation of data is critical and that CMS should not report any data without first validating them. As discussed above, The New York Times found substantial gaming of quality measure data, which led to highly inflated quality measure reporting on Nursing Home Compare. The same mistake should not be repeated with new measures in the Quality Reporting Program.
To validate the data, the Center proposes that CMS:
- Revise and test revisions to the survey protocol to review resident assessments.
- Draft additional guidance for surveyors requiring use of the revised protocol.
- Require that all surveyors be trained in the revised protocol.
- Conduct special surveys of resident assessments both when CMS has reason to believe that data submitted by a SNF are invalid and on a random basis.
- Report on Nursing Home Compare when a facility’s assessment data are invalid and cannot be publicly reported and report, instead, that the data submitted by the facility are invalid and cannot be posted.
- Promulgate regulations to require specific penalties for violations of resident assessment requirements, including civil money penalties, denial of payment for new admissions, and temporary management, depending on the scope and severity of non-compliance.
4. The three measures require discharge data that SNFs do not currently report for residents who remain in the facility, but are not covered by Medicare Part A. CMS is proposing to add a new item set and to include its content in the 5-day PPS assessment,
http://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/NursingHomeQualityInits/Downloads/Draft-of-the-new-MDS-30-PPS-Part-A-Discharge-End-of-Stay-NPE-SPE-Version-1140.pdf. We support this proposal, but, for the reasons described above, urge CMS to validate the accuracy of discharge assessment data and to impose appropriate sanctions against facilities that submit invalid data.
II. SNF Value-Based Purchasing Program
Section 215 of the Protecting Access to Medicare Act of 2014 required establishment of a value-based purchasing program for SNFs. The provision requires the Secretary to specify an all-cause all-condition hospital readmission measure. Reducing unnecessary rehospitalization is its explicit goal and purpose.
CMS needs to avoid the pitfalls of its prior value-based purchasing initiatives in SNFs, which were not successful. An analysis of the CMS demonstration of VBP in skilled nursing facilities between 2009 and 2012 found that VBP “did not directly lower Medicare spending and improve quality for nursing home residents.” L&M Policy Research, Evaluation of the Nursing Home Value-Based Purchasing Demonstration, page 50, Contract No. HHSM-500-2006-0009i/TO 7, http://innovation.cms.gov/Files/reports/NursingHomeVBP_EvalReport.pdf.
Concern about reducing rehospitalization as an independent goal
Before discussing the specific components of the proposed rule, the Center addresses concerns about encouraging a reduction in rehospitalizations, as an independent goal, without recognizing that some hospitalizations are necessary.
Writing in The New England Journal of Medicine, Joseph G. Ouslander, M.D. and Robert A. Berenson, M.D. recognize the high cost of unnecessary hospitalization of nursing home residents and support the reduction of such hospitalization, but they add a cautionary note:
Multifaceted strategies will be needed to address the current incentives for hospitalization if we are to improve nursing home care and prevent unnecessary hospitalizations, with their related complications and costs. Two caveats are critical. First, not all hospitalizations for conditions that can theoretically be managed outside an acute care hospital are preventable. Second, given fiscal constraints and the dearth of health care professionals trained in geriatrics and long-term care, not all nursing homes have the capacity to safely evaluate and manage changes in the condition of the clinically complex nursing home population. Setting unrealistic expectations and providing incentives to poorly prepared nursing homes to manage such care rather than transferring residents to a hospital could have unintended negative effects on the quality of care and health outcomes.
Joseph G. Ouslander and Robert A. Berenson, “Reducing Unnecessary Hospitalizations of Nursing Home Residents,” N Engl J Med 2011; 365:1165-1167 (Sep. 29, 2011), http://www.nejm.org/doi/full/10.1056/NEJMp1105449.
Dr. Ouslander developed a tool, INTERACT II (Interventions to Reduce Acute Care Transfers), that can carefully and successfully reduce unnecessary hospitalization of nursing home residents. Joseph G. Ouslander, M.D., et al, “Interventions to Reduce Hospitalizations from Nursing Homes: Evaluation of the INTERACT II Collaborative Quality Improvement Project,” The Commonwealth Fund (April 26, 2011), http://www.commonwealthfund.org/Publications/In-the-Literature/2011/Apr/Reduce-Hospitalizations-Nursing-Homes.aspx. See also Ouslander, “Improving Geriatric Care by Reducing Potentially Avoidable Hospitalizations,” http://www.avoidreadmissions.com/wwwroot/userfiles/documents/43/ouslander-interact-presentation-for-ny-ipro-webcast-jan-19-2011.pdf.
However, “the goal of INTERACT is to improve care quality, not to prevent all hospital transfers” and INTERACT “can result in more rapid transfer of residents who need hospital care” (bold font and italics in original, slide 16). INTERACT’s goals are to avoid hospitalization that should be avoided and to support hospitalization that is medically necessary.
We have several specific concerns about reducing hospitalizations as a free-standing goal.
1. If nursing homes are encouraged not to hospitalize residents, many residents who need hospital care will be endangered. Under contract with CMS, the Center analyzes nursing home decisions of the Departmental Appeals Board, both the Civil Remedies Division and the Appellate Division, and prepares a monthly report and searchable database. Over the years, the Center has never seen a case involving inappropriate hospitalization of a resident, but has read many cases where a facility was sanctioned for failing to contact a resident’s physician and hospitalize a resident who needed to be hospitalized.
2. Numerous studies show that improving staffing levels in nursing homes can reduce the perceived (and actual) need to hospitalize nursing home residents. Dr. Ouslander’s INTERACT II tool demonstrates the need for better staffing as do additional studies cited in the Center’s Weekly Alerts: “Reducing Rehospitalizations . . . The Right Way” (March 1, 2012), http://www.medicareadvocacy.org/reducing-rehospitalizations%e2%80%a6-the-right-way/; “More Nurses in Nursing Homes Would Mean Fewer Patients Headed to Hospitals” (March 10, 2011), http://www.medicareadvocacy.org/more-nurses-in-nursing-homes-will-mean-fewer-patients-headed-to-hospitals/; “Preventable Emergency Department Visits by Nursing Home Residents” (Aug. 19, 2010), http://www.medicareadvocacy.org/InfoByTopic/SkilledNursingFacility/10_08.19.PreventableEmergencyVisits.htm.
3. Any effort to avoid rehospitalization must recognize observation status and other long outpatient stays as types of hospitalizations. Observation status refers to patients in acute care hospitals being called outpatients, even though, like inpatients, they receive nursing and medical care, diagnostic tests and procedures, therapy, prescription and over-the-counter medications, and food and the patients may remain in a hospital bed for multiple days. The care provided to inpatients and observation status patients is indistinguishable. Any effort to reduce hospital admissions must recognize observation status as a hospital admission by another name. See additional information on observation status at http://www.medicareadvocacy.org/medicare-info/observation-status/
A Technical Expert Panel to consider a readmission/rehospitalization measure strongly supported counting observation time. RTI wrote in an August 23, 2012 paper entitled “Key Issues for TEP Consideration,” “The TEP was definitive that the SNF HRRM should include observation stays. RTI agrees that observation stays should be included in the measure.”
Observation status should be counted because the issue, for purposes of readmission, is whether a SNF sent the resident to the hospital for care and treatment. Whether the resident is called an inpatient or an outpatient (in observation status) is not within the control of the SNF; the decision about how to classify a patient who is in the hospital is made solely by the hospital. However, what is within the control of the SNF is the decision to send the resident to the hospital in the first place. If the SNF sends a resident to the hospital, its decision to hospitalize the patient is a rehospitalization decision. Whether the hospital calls the patient an inpatient or an outpatient (observation status) is irrelevant and has no significance for purposes of the rehospitalization quality measure.
Over the past few years, hospitals have increasingly categorized patients as outpatients in observation status, largely because of their concern about the Recovery Audit Contractor (RAC) program. Under current procedures, if a RAC reviews a hospital’s decision to classify a patient as an inpatient and decides that the patient should have been classified as an outpatient, the hospital receives virtually no reimbursement from the Medicare program for whatever medically necessary services it provided. Avoiding RAC review, and the significant financial consequences of a RAC’s reversal of an inpatient decision, leads hospitals to call increasing numbers of patients “outpatients” in observation status.
Researchers have documented that hospitals’ use of outpatient observation status parallels the decline in inpatient stays. Reviewing 100% of Medicare claims data for 2007-2009, researchers found that the number of outpatient observation stays for Medicare beneficiaries increased over the three-year period, while inpatient admissions decreased, suggesting "a substitution of outpatient observation services for inpatient admissions." Zhanlian Feng, David B. Wright, and Vincent Mor, "Sharp Rise In Medicare Enrollees Being Held In Hospitals For Observation Raises Concerns About Causes And Consequences," Health Affairs 31, No. 6 (2012).
Observation status time must be recognized as rehospitalization.
For the reasons stated above, we support use of a claims-based measure, rather than a measure based on self-reported minimum data set (MDS) information. The proposed rules report that CMS proposes to use NQF #2510, a claims-based measure, but the measure itself is not available on the NQF website.
What we found on the NQF website was an analysis of possible harmonization of two measures, NQF #2510 and NQF #2375. The American Health Care Association, the nursing home trade association, developed NQF #2375 and is its steward. We oppose use of NQF #2375 in the Value-Based Purchasing Program, both because it is self-reported (MDS-based) and because it is an industry-developed and -controlled measure. Facilities should not be setting the standards of care that they must meet for participation in the Medicare and Medicaid programs.
While supporting use of a claims-based measure, we have three concerns:
1. The time period for rehospitalization should be 90 days, not 30 days.
Thirty days is too short a period for this measure. Other efforts underway to change health care delivery systems that also seek to reduce rehospitalization, such as bundling demonstrations, focus on a 90-day time period following hospitalization.
Moreover, rehospitalization within a short period of time may reflect poor care in the hospital and inappropriately early discharge. Rehospitalization later in a resident’s SNF stay is more likely to reflect problems in the care provided by the SNF.
2. The Center cannot tell from the descriptions it has seen of NQF #2510 whether it includes all hospitalizations billed to Medicare or whether it is limited to hospitalizations of residents who are in a Part A stay in a SNF. The Center supports a broad measure and suggests that the measure include residents not in a Part A stay who are inappropriately hospitalized. The purpose of the measure should be identifying inappropriate hospitalizations of residents, whatever the source of payment for their nursing home stay. Medicare claims data for dually-eligible residents not in a Part A stay or for private-pay residents could be used.
3. The Center also proposes that CMS promulgate regulations prohibiting payment of a bonus under the Value-Based Purchasing Program to any nursing facility that, regardless of its performance under VBP,
- Does not accurately report staffing data to CMS;
- Does not have sufficient nursing staff to meet each resident’s needs.
III. Staffing Data Collection
Section 6106 of the Affordable Care Act added a new requirement that SNFs and nursing facilities electronically submit to the Secretary direct care staffing information, based on payroll and other verifiable and auditable data. The information must
- specify the category of worker (registered nurse, licensed practical nurse, licensed vocational nurse, certified nursing assistant, therapist, or other health care professional),
- distinguish agency and contract staff,
- include resident census information, and
- “include information on employee turnover and tenure on the hours of care provided by each category of certified employees.”
The purpose of the section was to ensure the accurate reporting of staffing information, since the self-reported data that are now publicly reported on Nursing Home Compare are overstated and inaccurate.
Although the language of CMS’s proposed rule either quotes or paraphrases the statutory language, proposed §483.75(u), the preamble suggests that “the obligation to submit information on ‘hours of care’ is satisfied by requiring facilities to submit hours worked by staff.” 80 Fed. Reg., 22081. We could not more strongly disagree. There is a considerable difference between hours of care provided and hours of care worked. All staff, as a matter of practice and by law, have time when they are paid but are not working – meal and other mandated breaks, mandatory in-service training, etc. In an eight-hour workday, at least one hour is devoted to meal and other mandated breaks. Staff must be paid for this time, but they are not providing care to residents.
CMS attempts to justify its substitution of hours worked for hours of care provided:
One of the primary objectives of the statute is for facilities to submit staffing information that is based on payroll and other verifiable and auditable data. We believe that most payroll or employee time and attendance systems capture the hours worked by individuals, and do not typically distinguish between hours spent doing different tasks (unless the tasks require different levels of pay). If we were to assume that “hours or [sic] care” was a subset of the hours worked by individuals, we would not be able to verify or audit the data submitted. As such, we believe that requiring facilities to report data on hours worked will yield the information Congress intended regarding “hours of care provided.”
80 Fed. Reg., 22081. We are not persuaded by this rationale.
The purpose of the statutory requirement is getting accurate staffing data; the means identified by the statute is payroll validation.
Counting all hours worked plainly and significantly overstates the amount of time spent by staff on direct care.
If CMS is unwilling to require facilities to submit hours of direct care actually provided, then it must delete at least one hour from total hours worked in order to reflect the time at work that is not dedicated to resident care.
The Center endorses the comments of California Advocates for Nursing Home Reform on staffing data collection.
Toby S. Edelman
Senior Policy Attorney