Medicare Hospice Payment Reform: Analysis of Medicare Hospice Benefit Utilization
January 7, 2016 09:59 AM
The Centers for Medicare & Medicaid Services (CMS) recently posted what it calls the December 2015 Technical Report. This report, Medicare Hospice Payment Reform:Analysis of How the Medicare Hospice Benefit is Used, contains details of hospice utilization and payment used by CMS for hospice payment reform decisions and was referenced in the preamble to the FY2016 payment rule.Much of the information is from FY2013 data.Items covered include live discharges, drug reporting on hospice claims, a comparison of visits received under hospice services and home health services, and, of course, financial analysis. Information is compared to that of the 2014 Technical Report. Both reports were prepared by and analyses completed by Abt Associates.Below is a summary of the findings in the 2015 Technical Report.
Total Medicare Utilization by Hospice Beneficiaries
These expenditures include payments made to hospices for services provided through the hospice benefit in addition to utilization of non-hospice Medicare services by beneficiaries on days in which they elected hospice. The individual non-hospice services utilized were not examined. Abt could not observe individual patients’ charts nor could it know the particulars of their medical conditions and needs. Therefore, Abt was unable to ascertain whether the non-hospice utilization observed is related or unrelated to the terminal illness and any related conditions of the beneficiary. The original intent of the hospice benefit was that the relatedness of non-hospice care would be determined on a case-by-case basis and CMS has never issued guidance, though the original Hospice Final Rule stated CMS’s belief that “hospices are required to provide virtually all care that is needed by terminally ill patients.”
Total Medicare hospice expenditures in FY2013 totaled $15.1B over approximately 92 million hospice days compared to $15.0B in calendar year 2012.It is difficult to do a direct comparison as data is reported slightly differently in the two technical reports.However, the fact remains that data is being gathered regarding total Medicare expenditures for the Medicare hospice benefit and the total expenditures outside of the hospice benefit that Medicare is making under Part A and Part B while the beneficiary is also receiving hospice care.These expenditures include DME, home health, inpatient and Part B and SNF services. The majority of these expenditures were for the Physician/Supplier and Other Part B category services.The Report delves into details regarding these expenditures and indicates that there are greater expenditures outside of the hospice benefit when the hospice patient is in an inpatient setting.The services in this category were primarily determined by the hospice to be unrelated to the patient’s terminal prognosis.Non-hospice spending is also compared across geographic regions and by facility type. Additional expenditure data by patient diagnosis is available.Not surprisingly, the greatest expenditures were for non-Alzheimer’s dementia.
Also in the financial analysis is a chapter dedicated to data of pre-hospice Medicare spending.This data is gathered as an initial step toward analyzing data for a case-mix payment system for hospice. The FY2014 Hospice Final Rulementioned that CMS would, upon collecting more accurate diagnosis data, evaluate whether a case-mix system could be created as part of future reform efforts. The objective of introducing a case-mix system into the hospice benefit would be to differentiate hospice payments by patient characteristics and thereby better align expenditures with individuals’ resource use. Given the information that is currently available on the hospice claim, it is not yet possible at this time to conduct a thorough assessment of whether a case-mix system is appropriate. However, as an initial step to begin thinking about a case mix system, pre- hospice Medicare spending was examined. Across the diagnoses, there is an inverse relationship between pre-hospice spending rates and lifetime hospice utilization.It does not necessarily follow that because there is variation in expenditures across diagnoses prior to hospice that there must also be variation in expenditures across those same diagnoses during hospice. The results indicate that hospice patients with the longest length of stay—patients with Alzheimer’s dementia, other non-Alzheimer’s dementias, and Parkinson’s disease (neurological conditions which are associated with more difficult prognostication)—have the least Medicare utilization prior to starting hospice, at daily rates which are on average less than the current per diem hospice payments. Abt notes that this finding is inconsistent with the concept of hospice as a cost-saver as it was promoted during the benefit’s creation.
Hospice Cost Reports – Benchmarks and Trends
Data from freestanding hospice cost reports were utilized.Throughout this chapter means are calculated two ways: over all hospices, and at the individual hospice provider level. If a mean is calculated over all hospices (weighted), then it is defined using the totals across hospices in a given year. For instance, the mean cost per patient calculated over all hospices is defined as the sum of costs across all hospices divided by the sum of patients across all hospices. When the mean is calculated in this manner, larger hospices influence the mean to a greater degree than smaller hospices and may be more representative of the industry as a whole.Alternatively, when the mean is calculated at the hospice level, it is calculated for each hospice; then a mean of those hospice means is calculated. When calculated in this manner, smaller hospices and larger hospices have an equal weight in the calculation.
Inpatient costs were calculated in this chapter and it was discovered that there is a discrepancy with how hospices are reporting this information.Specifically, some hospices show inpatient days but no inpatient costs and some show inpatient costs but no inpatient days.
Relative to labor costs, the data shows that smaller hospices have slightly lower labor costs per patient than larger hospices.
Other hospice service costs were also calculated, as was the total cost per patient.Other Hospice Services include the following 10 cost centers: drugs, biologicals, and infusion; durable medical equipment/oxygen; patient transportation; imaging services; labs and diagnostics; medical supplies; outpatient services (incl. E/R dept.); radiation therapy; chemotherapy; and “other”. For the drugs, biological, and infusion cost center, Abt also aggregated the sub-lines (i.e. analgesics and sedatives/hypnotics) up to this center. Three costs centers—drugs, DME, and medical supplies—account for the majority of the “Other Hospice Service” costs.
The measures of average cost per patient when measured in constant dollars have remained fairly flat over time, trending upward until 2007 and downward after this time. The 2013 average costs per patient was roughly equivalent to the 2004 average costs per patient. Note that the mean costs per patient reflect costs associated with the mean length of election, which is significantly longer than the median length of election.
Trend Estimates of the Aggregate Cap and Seasonality Analyses
In its 2012 Report to Congress,The Medicare Payment Advisory Commission (MedPAC) analyzed 2009 hospice claims and determined that above- cap hospices tended to be “for profit, freestanding hospices and to have smaller patient loads” (p. 294). Abt similarly reported finding that above-cap hospices are smaller (there were 133 beneficiaries on average in above-cap hospices vs. 93 in below-cap hospices).Additionally, this table shows hospices are more likely to be above-cap if they are newer (17.1% certified since 2000, 3.3% certified in the 1990s, and 0.7% certified in the 1980s), urban (12.4% urban vs. 5.3% rural), and in the West or South (17.5% West, 13.5% South, 2.9% Midwest, 2.7% Northeast). These patterns seemed to remain fairly consistent across time—the same characteristics were associated with higher rates of above-cap status in 2006 as in 2014.Abt identified large regional variation in the prevalence of hospices exceeding the cap: hospices in the West and South exceeded the aggregate cap at a rate 5–6 times that of hospices in the Midwest and Northeast.
CMS has indicated in its comments in proposed and final rules for hospices that it may consider rebasing the hospice aggregate cap.The objective for the per-person cap amount—that would be multiplied by the number of beneficiaries the hospice treated during the year to determine the aggregate cap amount (or total allowable limit on payments—would be that it equal 40% of the average costs of traditional, curative treatment for cancer patients in the last six months of life). At the time, hospice patients were predominantly those with cancer. The original base reimbursement limit (per-person) amount was set for $6,500, for which there was a consensus would be sufficient to provide compassionate, high- quality care.
Any changes to the hospice per-person cap amount would require a change to the existing statute and are not within the scope of current hospice payment reform efforts.Therefore, Abt stated that the analysis here is exploratory and compares what the cap amount would be using CY2013 data and the original aggregate cap methodology to the actual 2013 cap amount.Findings are that the cap amount would be lower.
Seasonality patterns:The objectives of this section are to examine whether the timing of live discharges from hospices relate to the cap year or risk of exceeding the cap. In the 2014 Technical Report it was reported that data showed that hospices exceeding the cap tended to have higher live discharge rates.This seems to continue to be the case.The 2015 Technical Report shows that live discharges increase the last four months of the cap year for all hospices with the rate of live discharges being higher for above-cap hospices than below-cap hospices.Note that as hospices near (and exceed) the cap limit, the adjusted rates of overall live discharge increase (from 13.0% to 27.8% over the range of the figure), no longer terminally ill increase (from 7.1% to 19.3% overall), and revocations increase (from 5.0% to 13.4% overall). This finding should be investigated further to determine whether hospice preference for the anticipated risk of exceeding the cap contributes to the rates of potentially inappropriate live discharges or hospice-encouraged revocations.
Trends in Live Discharge
The live discharge rate varies substantially across hospices.In their calculations Abt excluded hospices with fewer than 50 discharges (live or deceased discharges).The median live discharge rate is 18.3 percent.The top 10 percent of hospices (with 50 or more discharges) in terms of the live discharge rate represented 5.9% of all hospice payments allocated to hospices with 50 or more discharges.There is large variation among states.Additionally, hospices above the 90thpercentile of live discharge provided on average 3.97 visits per week when looking at all six disciplines of care reported on the hospice claim. When focusing only on visits classified as skilled nursing or medical social services, these hospices provided on average 1.91 visits perweek. The hospices below the 90thpercentile for live discharge meanwhile provide more visits on average with 4.48 visits per week looking at all six disciplines of care and 2.35 visits per week when only looking at skilled nursing or medical social services.
Trends in Drug Reporting on Hospice Claims
Claims from calendar year 2014 where the drugs were reported were used for this analysis.Hospices began reporting this information on claims April 1, 2014.Abt looked at the total expenditure on drugs as well as expenditures based on the patient’s level of care.Not surprisingly, the largest percentage of injectable drugs were provided to patients receiving the general inpatient (GIP) level of care and the largest percentage of non-injectable drugs were provided to patients receiving the routine home care (RHC) level of care.Abt also looked at the types of drugs being billed.There are over 1,600 unique non-injectable drug names reported on the claims.
Evaluation and Management Visits (E&M) Prior to Starting Hospice Services
Most beneficiaries had these visits immediately prior to starting hospice. Again, looking at just the beneficiaries with an E&M visit in the last 30 days (and dropping the non-standard providers), 77.6% of beneficiaries had an E&M visit in the seven days before starting hospice. Future research should examine if the presence of E&M visits are related to any quality outcomes. The presence or absence of an E&M visit could impact whether entry into hospice is appropriate for the beneficiary and could also impact live discharge rates.
Comparison of Visits Provided Between the Medicare Hospice Benefit and Home Health Medicare Benefit
Fewer visits and minutes of service were provided to beneficiaries under home health than hospice.The data also shows that hospices provide more visits and minutes of care at the beginning of hospice care, consistent with the change in hospice payment for a greater daily routine home care rate for the first 60 days of hospice care.
Skilled Visits During the Last Two Days of Life
The number of skilled visits during the last days of life has been a concern of CMS for some time, hence the implementation of the service intensity add-on (SIA).The data in the 2015 Technical Report indicates 27.4% of beneficiaries receiving RHC level of care at the end of life did not receive a skilled visit during the last two days of life.Abt delved further into this data looking at characteristics of hospices and site of service.Excluding the very short stay beneficiaries, it does not appear that length of stay has a large impact on the probability of receiving a visit at the end of life.Beneficiaries in small hospices were more likely to go without visits than those in large hospices.
Overall, the data in the 2015 Technical Report mirrors that of the 2014 Technical Report.The National Association for Home Care & Hospice (NAHC) will continue to monitor and report how the information from this Report is being referenced and utilized by CMS, other oversight bodies such as the OIG, and policymakers.