1	This is the title page for the presentation titled "Utilization and Vulnerabilities in the Medicare Hospice Benefit" Presented at the Gerontological Society of America Meeting.
2	This slide has background material.  It states, Medicare has covered hospice services since early 1980s. Beneficiaries who elect hospice forgo curative treatment for their terminal condition.  Medicare continues to cover items for conditions unrelated to terminal illness.  Hospice offers terminally ill patients (< 6 month life expectancy) important access to palliative care services meant to maximize patient comfort and quality of life in the days before death.
3	This slide has background material.  It states, core benefit features largely unchanged over the last 30 years, but last decade has shown notable changes.The number of beneficiaries using hospice more than doubled (from 534,000 in 2000 to 1.3 million in 2012). The total number of providers has also increased (from 2,255 in 2000 to 3,727 in 2012, mostly among for-profits).  Medicare payment for hospice services grown substantially from $2.9 billion in 2000 to $15 billion in 2012.
4	This slide motivates the reason for the presentation.  CMS and industry stakeholders recognize that some changes may be driven by vulnerabilities in the current payment system.   As a result, researchers and policymakers examining changes in hospice utilization over time and identifying the sources and motivations for these changes, particularly as they relate to the behaviors and practices of hospice providers. 
5	This slide presents the focus of the panel.  The presentation seeks to have a better understanding vulnerabilities related to the Medicare Hospice benefit.  These include:  a descriptive summary of provider cost reports identifying the distribution of key components of the cost report. An analysis of live discharges in the hospice system. Current utilization patterns of general inpatient care while in hospice.  A Descriptive examination of utilization of the Part D drug benefit while a beneficiary is enrolled in hospice. 
6	This slide presents the objectives of the presentation.  The first objective is to provide the research community an understanding of the current Medicare hospice payment systems vulnerabilities. The second objective is to provide an understanding of the sources of variation in costs, practice patterns, and utilization of the Medicare hospice benefit.
7	This slide acknowledges research funders.   The analyses upon which this presentation is based were performed under Contract HHSM-500-2005-00018I, entitled, Hospice Study and Report, funded by the Centers for Medicare & Medicaid Services, Department of Health and Human Services. The content of this presentation does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does the mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government. The authors assume full responsibility for the accuracy and completeness of the ideas presented.  
8	This slide is a title slide for the presentation "Hospice Cost Reports: Benchmarks and Trends, 2004-2011".
9	This slide presents background information.  Cost reports reflect fiscal year costs and payment information.  They are required annually from Medicare-certified providers. They are a publically available source for costs of services from hospice providers.Other sources of descriptive information on costs to hospices are not easily found.
10	This slide presents objectives of the presentation.  The presentation will examine, descriptively, cost reports for freestanding hospice providers from 2004-2012 to do the following.  First, present an easy-to-apply, uniform trimming methodology that eliminates extreme-value cost reports from multiple years of data.  Second, examine three measures of central tendency to determine whether an average experience exists within the data.  Third, highlight trends over the 2004-2012 time period and cost centers where data are bounded by zero or potential misreporting of information is more common.
11	This slide previews results.  First we find a simple methodology to provide consistent measures over the varying cost report years.  Second, reports of $0 costs may influence results.  Third, Average per person drug costs have decreased. Fourth, visiting services are the largest and increasing proportion of total costs.  Finally, total costs per patient have not significantly increased from 2004-2012.
12	This slide discusses the data.  The presentation looks at cost Reports for Freestanding Hospice Providers using data from FY 2004-2012. The data is cleaned to eliminate extreme values.  The following types of cost reports are dropped.  Short or long cost report periods, that is,  cost reports with a period less than 10 months or greater than 14 months. Missing or negative value costs or payments.  That is, cost reports with missing information or negative reported values for total costs or payments. Top and bottom 1% of cost per day. That is, providers in the highest and lowest percentile in costs per days across all levels of care. Top and bottom 5% of provider margins.  Aggregate of cost centers does not equal total costs as reported.
13	This table describes the descriptive statistics before and after the sample trim from 2009 cost reports.  
14	This slide reviews methods used to create the results.  Cost centers are grouped into four broad categories.  This includes Inpatient Care; Visiting Services; Other Hospice Services; and Non-reimbursable Services.  Include shared service costs are allocated to each cost center.  We identify measures that describe the average in two ways.  The first way is representative of overall industry costs.  The second way is the average of provider-level means.  The final way is the median of provider-level means.
15	This table describes inpatient care costs.  Inpatient care costs differ by the method of averaging and have varying trends over time.
16	This slide describes results of inpatient care costs.  First, it is difficult to describe average inpatient costs.  The measures of central tendency disagree, even among providers reporting inpatient costs.  Second, one third of providers report $0 in inpatient costs. However, significant numbers of cost reports  list a non-zero number of days but zero costs for inpatient care.  Inpatient costs constitute ~12% of all costs across all freestanding providers and ~14% of all costs for providers who report some inpatient costs.
17	This table describes visiting service costs from 2004 - 2012.  The mean over all providers is slightly lower than the mean at the provider level.  The trends across all measures are rising over time.
18	This table describes other hospice services from 2004 - 2012.  Most "other hospice service" costs come from Drugs and biologicals, DME and Oxygen, and medical supplies.  The proportion of costs which are other hospice service costs are falling over time.
19	This table describes drug costs per patient day by year using data from 2004 - 2012.  Drug costs have fallen from roughly $20 per day in 2004 to $11 per day in 2012.
20	This table describes total costs using data from 2004 - 2012.  Costs have risen only slightly between 2004 to 2012.
21	This slide contains conclusions.  First, a simple methodology provides consistent measures over the varying cost report years.  Second, roughly one-third of providers report zero inpatient costs, causing skewed average costs.  Third, the high number of providers reporting zero is unexpected as these costs should include contractual costs for inpatient care.  However, we find evidence of misreportingproviders  report zero inpatient costs with non-zero inpatient days.Fourth, per person drug costs have decreased, on average, over the 2004-2011 time period.  Fifth, up to 25% of providers report $0 in non-reimbursable costs, including required bereavement costs. Sixth, visiting services are the largest and increasing proportion of total costs.  Seventh, total costs per patient have not significantly increased from 2004-2011.
22	This is a title slide for the presentation "Facility Problematic Live Hospice Discharges".
23	This slide states this research was funded by a CMS contract to Abt Associates and Brown University School of Public Health.
24	This slide provides an introduction.  First, Live discharges can occur for variety of reasons:  1) patient choice; 2) medical  condition improves; 3) hospice may be avoiding costly medical care related to terminal illness.  Second, Kutner, 2004  provides reasons for discharge 1) 79% improve or stabilized; 2) 7% pursue aggressive treatment; 3) 12% patient/family decision.  Transfers from one hospice to another are a rare reason. 
25	This slide provides an introduction.  First, MedPAC has stated  the live discharges rate in above CAP  hospices is 48% compared to 16% in below CAP hospices. Second,  above cap hospices substantially higher live discharges combined with longer length of stay raise questions about whether above cap hospices are admitting patients before they meet hospice eligibility. 
26	This slide describes methods.  These analyses examined hospice patients discharged between January 1, 2010 and December 31, 2010 with all discharges having six month follow up in 2011.  Analysis at the facility level is among those facilities with at least 30 discharges.
27	This slide describes methods.  The authors document the rates of live discharges and six months outcomes after hospice discharge. They then examine patterns of 4 problematic live discharges.  First, High provider rate of complicated transitions.  Second, high provider rate of long stay patients and low rate of long stay patients and low rate of live discharges.  Third, high provider rate of discharges of long stay patients (>210 days).  Fourth, high provider rate of live discharges in the first 7 days of hospice stay.
28	This slide describes methods.  The facility based analysis examined a weighted index of the 4 problematic discharge patterns.Organizational characteristics examined include: 1) For profit vs. not for profit ; 2) Chain status operationalized as national, regional, state, or a hospice provider without affiliation with other providers.  With the facility as unit of analysis, a multivariate poison model examined whether type of hospice providers classified by ownership and chain affiliation  differed in the rate of problematic live discharges. 
29	This map shows the state average live discharge rate in the United States.  The mean rate of live discharges was 18.1.
30	This slide describes the results.  There were 956,497 discharges among 3,489 hospice providers.  There were 173,409 (18.1%) live discharges in 2010.  49.2% persons died six months after discharge.
31	This diagram shows how many live discharges were or were not hospitalized within 30 days and their eventual outcomes.
32	This slide shows the results of variation in the rate of live discharge by hospice programs.
33	This slide shows multivariable results at the provider level.
34	This slide describes the conclusions of the presentation.  First, there is substantial variation in the rate of live discharge by state and hospice facility raises concerns that need further research. Second, Problematic Live Discharges are higher among for-profit providers with slightly higher rates among those hospices not in national or regional chains. 
35	This slide describes implications for policy.  The rate of live discharge is a potential vulnerability that should be monitored as part of hospice payment reform.
36	This is the title slide for the presentation "Understanding Variation in Utilization of Hospice Inpatient Care".
37	This slide provides background.  First, General inpatient care, or GIP, is short-term inpatient care provided in a hospice facility, hospital or SNF for pain control for acute or chronic symptom management which cannot be managed in other settings.  Second, extant research suggests differences in Medicare spending and length of stays across different sites of service, providers, and geographic location for GIP.
38	This slide shows a table with the FY 2014 daily payment rates for the four levels of care in hospice.
39	This slide presents the objectives of the presentation.  The first objective is to examine utilization of GIP services.  The second objective is to better understand characteristics of hospice providers who provide GIP services as well as those who do not provide any GIP services.
40	This slide describes the data used.  All 2012 CMS hospice claims were used. A GIP stay defined as consecutive GIP days in hospice claims file.
41	This slide has a table presenting how many GIP stays a beneficiary has.  93.1% of beneficiaries with a GIP stay had only 1 GIP stay.
42	This slide has a pie chart showing how long the GIP stay was.  The average stay was 5.5 days.
43	This slide has a graph showing how the GIP length of stay varied by site of service.
44	This slide describers the number and percent of GIP days by site of service.
45	This slide has a graph describing the site of service a beneficiary has before the start of their GIP stay and after their GIP stay is finished.
46	This slide has a table summarizing the percent of GIP days each provider in the data had in 2012.
47	This slide has a graph showing that older hospices are more likely to provide GIP compared to newer hospices.
48	This slide has a graph showing that small hospices are less likely to provide GIP compared to medium and large hospices.
49	This slide has a graph showing that providers in the south and the west census regions are less likely to provide GIP compared to hospices in the New England census region.
50	This slide had concluding remarks.  First, among the quarter of hospice beneficiaries who had a GIP day, most had just 1 GIP stay.  Second, most GIP stays are short (5.5 days), but LOS varied by site of service.  Third, over half of beneficiaries were not in hospice the day immediately before their GIP stay.  Fourth, there is considerable variation in provider characteristics and provision of GIP.
51	This is the title slide for the presentation "Characteristics of Hospices Serving Beneficiaries with Medicare Part D Claims".
52	This slide contains background on Medicare's prescription drug benefit.  Medicare Part D is a federal program subsidizing prescription drug costs.From 2006-2010, Part D expenditures increased 30% from $43 billion to $56 billion.  All beneficiaries entitled to Medicare Part A benefits are also eligible for Part D coverage.
53	This slide describes how hospices should cover medication under Medicare.  The Medicare Hospice Benefit fully covers medications related to patients terminal condition.  Covered medications costs are structured into the hospice benefits per diem payment rate.
54	This slide contains background on the OIG report in 2012.  It states the OIG reported Part D billed $33.6 million in 2009 for standard hospice medication.
55	This slide discusses the purpose of the presentation.  We describe Part D utilization for palliative medicines or drugs related to the primary diagnosis.We characterize hospices associated with high rates of these medications billed through Part D.  We use Medicare Hospice and Part D Claims.  We match beneficiaries prescription fill dates to hospice admission/discharge timeframe.
56	This slide presents results.  In 2010, we identified 750,590 Medicare hospice beneficiaries enrolled in Part D. Among these individuals, 208,468 received 1,875,065 palliative and related medicines through Part D during hospice enrollment totaling $99,148,348.  In total, hospice beneficiaries received 5,852,399 prescriptions of any type totaling $350,250,118 during 2010.
57	This table describes the drugs received through Part D by hospices beneficiaries in 2010.  Total drugs billed to Part D equaled $350 million dollars.
58	This table is the same as the table in the previous slide but highlights debility patients had $49.2 million in drug costs and failure to thrive patients had $25 million in drug costs.
59	This table is the same as the table in the previous slide but highlights 14.9% of patients in the sample received analgesics under Part D.
60	This map shows Part D utilization per hospice admission for palliative and related Medicines in 2010 by state.
61	This chart shows that Part D claims for Palliative and Related medicines are concentrated among a few hospices.
62	This chart shows palliative and related medicine concentrated is related to service in the nursing home.  
63	This table indicates hospice characteristics associated with high rates of billing for palliative and related drugs.  Newer hospices have higher rates of billing compared to older hospices.  For profit  facilities have higher rates compared to non-profit or government owned facilities.
64	This slide presents limitations.  First, our analysis uses data from a time period (2010) prior to the 2012 OIG report on Part D and hospice, and hospices may have since corrected their behavior.  Second, we cannot identify patients exact clinical condition and therefore cannot adequately determine the appropriateness of their drug usage.
65	This slide contains final remarks.  Part D was billed almost $100 million in 2010 for medications hospices likely should have covered per Medicare rules.Over half of this amount is associated with just ten percent of hospices.  There is an association between nursing home service and Part D billing. Newer, for-profit, and free-standing hospices had higher billing rates.  These findings highlight a payment vulnerability which should receive ongoing regulatory oversight.