The 2026 National Hospice Study

Where the averages break

The national hospice numbers describe two industries averaged into one.

This is the Astrelis State of the Industry series: federal data, read straight, with the distortions named. This edition covers hospice. Every figure in it comes from federal data and independent analysis: the MedPAC March 2026 and March 2025 reports, the CMS FY2026 hospice final rule and FY2027 proposed rule, the CMS Hospice Monitoring Reports of April 2025 and April 2026, CMS public quality and provider data files, the CMS nationwide hospice enrollment moratorium notice, and HHS OIG. The full bibliography is in Appendix A: Source Overview; the methods, vintages, and killed claims are in Appendix B: Methods.

The thesis of this edition is one sentence. The national hospice numbers, the provider count, the Medicare margin, the quality averages, describe two industries averaged into one: a nonprofit sector and a for-profit sector that differ in margin, length of stay, live discharge, breadth of service, and geography, blended into single medians that describe neither. Every operator who benchmarks against a national hospice average is comparing itself to a number that conceals the very differences that decide performance. Here is where the averages break, and, new this cycle, here is the federal machinery that has stopped averaging.

The Quick Read

The nine numbers from this study worth carrying into your next board meeting:

The growth is all one kind

The hospice count grew about 64 percent in a decade. The decade's growth was almost entirely for-profit while nonprofit and government supply contracted.

An industry whose front door is closed nationwide is an industry whose regulator stopped treating the problem as local.

The count is a real growth story: 6,706 hospices cared for Medicare beneficiaries in 2024, up from about 4,092 in 2014, roughly 64 percent more providers in ten years. Underneath it is a composition shift. For-profit hospices were about 63 percent of providers in 2014 and about 82 percent in 2024, and they now serve 60 percent of Medicare hospice patients. In 2024, nonprofit provider counts fell 7.4 percent and government counts fell 4.4 percent while for-profits grew, so the decade's net growth was almost entirely for-profit while nonprofit and government supply contracted. And it happened in a few places. Between 2019 and 2023, four states ran far ahead of the 7.8 percent national growth rate: California added 1,046 hospices, Texas 257, Arizona 96, and Nevada 60.

Those four states were the origin of the oversight story; the oversight is now national. It began as geography: the California State Auditor found in 2022 that hospice growth in Los Angeles County had vastly outpaced the need, with numerous indicators of fraud and abuse, and since July 13, 2023 CMS has held newly enrolling hospices in Arizona, California, Nevada, and Texas under a provisional period of enhanced oversight with prepayment claims review, expanded to Georgia and Ohio in 2025. As of June 2025 that program had reviewed 668 hospices in the original four states and revoked the Medicare enrollment of 122. Then CMS closed the front door everywhere: effective May 13, 2026, a nationwide temporary moratorium halts enrollment of new hospices and hospice practice locations in every state, territory, and the District of Columbia, for six months, extendable in six-month increments. The moratorium also catches a hospice sold in a non-exempt change of majority ownership, because re-enrolling after such a sale counts as a new enrollment. The concern remains specific, not categorical, about certain providers in certain geographies rather than for-profit hospice as a class. But an industry whose front door is closed nationwide is an industry whose regulator stopped treating the problem as local.

There is also, for the first time in years, a question about the direction of the count itself. CMS's monitoring file, on preliminary fiscal 2025 claims, counts 6,642 hospices billing Medicare in fiscal 2025, down from 6,732 in fiscal 2024, the first decline in the report's fiscal 2021 to 2025 window. The figure is preliminary, extracted January 15, 2026, and CMS cautions that the last half of fiscal 2025 may be incomplete, so the decline may soften as claims run out. Read alongside the moratorium, though, it suggests the decade-long expansion may have reached an inflection point.

What it means: any provider-weighted national hospice average carries that four-state, for-profit-entry composition inside it. Read national figures with the geography and the ownership disclosed. Which providers make up the denominator is the first question, not the last.

The average margin belongs to no one

For-profit hospices earned 13.7 percent on Medicare in 2023. Nonprofit hospices earned negative 1.3 percent. The 8.0 percent aggregate is a policy-level measure, not the margin of a typical hospice.

The margin question is not just who owns the hospice; it is what the hospice is attached to and how many patients it serves.

MedPAC put the aggregate Medicare hospice margin at 8.0 percent in 2023, down from 9.8 percent in 2022. That is the number policy debates cite, and it is a policy-level measure, not the margin of a typical hospice. Ownership is the loudest divide: for-profit hospices earned 13.7 percent, nonprofit hospices negative 1.3 percent, with freestanding nonprofits at positive 2.6 percent. But structure and scale cut deeper. By structure, freestanding hospices earned 10.2 percent while hospital-based hospices ran negative 25.6 percent. By scale, the lowest-volume fifth of hospices, which together served just 2 percent of patients, ran negative 20.2 percent, while the highest-volume fifth, serving 67 percent of patients, earned 9.5 percent. The margin question is not just who owns the hospice; it is what the hospice is attached to and how many patients it serves.

There is a margin this study cannot give you, and the reason matters. In home health the honest split is Medicare against all-payer. Hospice has no published all-payer margin: MedPAC says irregularities in how hospices report total revenue and expense prevent a reliable estimate, and Medicare is about 91 percent of hospice days anyway. For hospice, the splits that carry the weight are ownership, structure, and scale. And the stakes are live: MedPAC projects the aggregate near 9 percent for 2026 and, on that basis, recommended Congress eliminate the fiscal 2027 update entirely, a freeze, on a 17 to 0 vote, while CMS has proposed a 2.4 percent update. A freeze written to a 9 percent aggregate lands very differently on a large for-profit at 13.7, a small rural hospice in the negative-20 quintile, and a hospital-based program at negative 25.6.

What it means: when someone quotes a hospice margin, ask whose. The aggregate conceals a spread of more than 35 points between the structures at its two ends, and a hospice that benchmarks against it is measuring itself against a blend of its opposite.

The only benefit with a ceiling

Hospice is the only Medicare fee-for-service benefit with a provider-level aggregate payment cap calculated from a per-beneficiary amount, and 28 percent of hospices went through it in 2023.

Take those four states out and the national rate falls from 28 percent to about 6 percent.

Every other Medicare fee-for-service benefit pays for what is delivered. Hospice alone carries a provider-level aggregate payment cap calculated from a per-beneficiary amount: $35,361.44 in fiscal 2026, multiplied by the hospice's beneficiary count, with payments above that ceiling repaid to Medicare. The cap turns long stays from a clinical style into a solvency question. In 2023, 28 percent of hospices exceeded it, up from 23 percent a year earlier and 19 percent in 2020. Payments over the cap equaled 2.9 percent of all Medicare hospice spending, and the average above-cap hospice owed back about $410,000.

The hospices that hit the ceiling are a type: longer stays, higher live-discharge rates, more facility patients, more likely for-profit, freestanding, and recently enrolled, and concentrated in the four states from Finding 1. More than half of California hospices exceeded the cap in 2023, and roughly 30 to 40 percent did in Texas, Nevada, and Arizona. Take those four states out and the national rate falls from 28 percent to about 6 percent. The cap is also where Finding 2's margin gets real: above-cap hospices posted a 17.8 percent Medicare margin before returning their overpayments and negative 1.4 percent after.

There is a second, quieter ceiling an experienced operator will look for. Inpatient days, general inpatient plus respite, may not exceed 20 percent of a hospice's total Medicare hospice days; days above the limit are paid at the routine home care rate rather than the inpatient rate, with the difference refunded. MedPAC notes this inpatient cap is rarely exceeded, which is itself a datum: the binding constraint in this benefit is not too much inpatient care, it is too many long routine-home-care stays.

What it means: the caps sort the industry into populations that differ on stay length, discharge behavior, geography, ownership, and post-repayment margin. The national exceed-rate, 28 percent or 6 percent depending on which four states you count, conceals which population a given hospice belongs to. Any hospice within reach of the ceiling has to model its own cap position.

The crisis-capacity question

Routine home care is 98.8 percent of hospice days. One patient in six still touches a higher level of care.

The day shares say the benefit runs on one level; the patient shares say escalation is a normal part of one stay in six.

Medicare's hospice benefit pays four levels of care: routine home care, continuous home care for a patient in crisis at home, inpatient respite to relieve a family, and general inpatient care for symptoms that cannot be managed at home. Two numbers describe how the benefit actually runs, and they use different denominators. By days, routine home care was 98.8 percent of all Medicare hospice days in 2023 and again in 2024; general inpatient care was 0.8 percent, respite 0.3 percent, and continuous home care one-tenth of one percent. By patients, the picture is meaningfully different: in both years, 16 percent of Medicare hospice patients received at least one general inpatient day, 4 percent received respite, and 2 percent received continuous home care. The specialized levels are a sliver of the days and a meaningful minority of the patients, brief, high-acuity episodes inside stays that are otherwise routine.

The provider file adds the third cut: over 2022 to 2024, about 1 in 9 hospices with reportable data provided routine home care and nothing else, never billing a crisis, respite, or inpatient day, while the other 88 percent reached a higher level at least once; about a quarter of providers are suppressed for too little data and sit in neither share. Low utilization alone is never evidence of absent capacity: a small hospice with a stable census can go a year without a legitimate continuous-care day. But the operator question the numbers pose is about capability, not statistics: when a patient crashes at two in the morning on a Saturday, can this hospice actually escalate, to continuous care at home or a general inpatient bed, and does it happen reliably enough to show up in claims at all.

What it means: the level-of-care mix is a fingerprint the national totals conceal. The day shares say the benefit runs on one level; the patient shares say escalation is a normal part of one stay in six. Where a specific hospice sits between those two truths is a question about its crisis capacity, including nights and weekends.

The last three days

Fewer than half of hospice patients get a nurse or social-worker visit in their final days.

Death does not keep business hours, and the staffing pattern behind those numbers sometimes does.

CMS publishes a measure built for the moment hospice exists to handle. Hospice Visits in the Last Days of Life, HVLDL, is the share of a hospice's patients who received an in-person visit from a registered nurse or a medical social worker on at least two of the final three days of life. Nationally, from claims across 2023 and 2024, it is 48.3 percent. Fewer than half of American hospice deaths included that visit pattern. And the average conceals the point: across the roughly 4,400 hospices with a reportable score, the bottom tenth delivered these visits to about 21 percent of their dying patients and the top tenth to about 81 percent, a 60 percentage-point gap on the most basic promise hospice makes. About 2,450 hospices had too few deaths to score, and a suppressed score is not a low score, it is no score.

A second CMS measure agrees from another angle. The Hospice Care Index folds ten claims-based indicators into one score from zero to ten; the national value is 8.8, and about 39 percent of scored hospices earn a perfect 10. But two of its indicators track the same end-of-life staffing gap: nationally, gaps in nursing visits appear in 52.4 percent of hospice elections, and only 9.6 percent of skilled-nursing minutes are delivered on weekends. Death does not keep business hours, and the staffing pattern behind those numbers sometimes does.

What it means: HVLDL is the measure a family would ask for if they knew it existed, and the one a national average conceals most completely. The distance between a 21 percent hospice and an 81 percent hospice is the distance between two different promises.

The discharge that is not a death

Almost one in five hospice discharges is a patient leaving alive, and the rate is climbing.

The public data accuses no one. It raises the question OIG raised.

Hospice is for the last months of life, so most stays end in death. A growing share end another way. In 2023, 18.5 percent of hospice discharges were live discharges by MedPAC's count, up from 17.3 percent a year earlier, and CMS's monitoring file puts the rate at 19.1 percent in preliminary fiscal 2025 on its broader basis, which also counts transfers between hospices and moves out of the service area. Some of this is what the benefit intends: patients revoke to seek treatment, move, or stabilize and no longer qualify, and the two most common recorded reasons are beneficiary revocation, 6.7 percent of discharges, and no longer terminally ill, 6.2 percent. The average again sits on a wide distribution. Among hospices with more than 30 discharges, the median live-discharge rate was about 21 percent, and the top tenth discharged at least 56 percent of their patients alive. CMS splits the pattern further: 7.3 percent of live discharges are early, 37.4 percent late, and burdensome transitions that route a patient through a hospitalization and back run 8.5 percent of one type and 2.1 percent of another.

A high rate is not proof of anything by itself. MedPAC says only that it may signal quality or program-integrity issues, and a genuinely improving patient population would show one too. But it is the pattern a decade of federal scrutiny has circled. HHS OIG, in its 2018 hospice portfolio, wrote that the per-diem system creates incentives for hospices to minimize their services and seek beneficiaries who have uncomplicated needs, and documented enrollment of beneficiaries who were not terminally ill. The public data accuses no one. It raises the question OIG raised, and shows which hospices sit at the far end. It is also, as Finding 9 details, now part of a published federal score: two of the nine SSVI measures are live-discharge measures.

What it means: live discharge is an association, not a verdict. A national rate near one in five conceals whether a given hospice's live discharges are the benefit working or the incentive working. Only its own number, split by reason and stay length, against true peers, answers that.

The two-patient problem

A cancer patient stays 51 days. A dementia patient stays 164. The mix decides the economics.

Diagnosis mix, setting mix, length of stay, margin, and cap exposure are one chain.

Hospice used to be, in the public imagination, a cancer benefit. It is now split down the middle between two clinical populations. In fiscal 2024, cancers were the largest principal-diagnosis group at 22.2 percent of Medicare hospice patients with the Alzheimer's, dementia, or Parkinson's group at 21.6; on preliminary fiscal 2025 claims the order flips, 21.9 percent for the dementia group against 21.7 for cancers, with cardiac diagnoses third at about 19 percent. The two populations are clinically and financially different, and the difference is length of stay. MedPAC's claims-based figures put the average lifetime stay at 51 days for a cancer decedent and 164 for a neurological one, with COPD at 131. Setting moves it the same way: 97 days at home, 113 in a nursing facility, 169 in assisted living. The national median stay is 18 days and the average 96, but that average blends a two-week cancer death and a five-month dementia decline.

Length of stay is where the mix becomes margin. A hospice concentrated in nursing and assisted-living facilities, serving long-stay dementia patients, collects the routine-home-care rate for many more days at lower daily cost, which is why MedPAC found facility-heavy hospices running Medicare margins near 15 percent against about 2 percent for facility-light ones in 2022, and why they are the most likely to hit the cap in Finding 3. Diagnosis mix, setting mix, length of stay, margin, and cap exposure are one chain, and a national average cuts it in the middle.

What it means: the setting and diagnosis mix is the master variable, and it is invisible in any national total. A short-stay, home-based, cancer-heavy hospice and a long-stay, facility-based, dementia-heavy one share a per-day rate and little else that matters. Benchmark either against the blend and the comparison conceals exactly the differences that drive the result.

The two-tail problem

One hospice stay in five lasts four days or less. One in six lasts more than half a year.

Length of stay is not a number, it is a shape.

Hospice does not have one length-of-stay problem; it has two, and they live at opposite ends of the same distribution. On preliminary fiscal 2025 claims, 20.4 percent of beneficiaries leaving hospice had lifetime stays of 1 to 4 days, while 17.6 percent had stays of 181 days or longer. The national conversation, and most federal scrutiny, attends to the long tail: eligibility, live discharge, cap exposure, the patterns of Findings 3 and 6. The short tail is quieter and at least as costly in human terms. A four-day stay means the referral came days from death: too late for the interdisciplinary team to do most of what hospice is for, after weeks or months in which the patient qualified for a benefit nobody started. It concentrates the most intense, most expensive days of care, admission and death, into a stay with no middle, and it points upstream, at referral timing, physician comfort with the conversation, and hospital relationships, more than at anything the hospice controls alone.

The two tails are not the same problem wearing two sizes; they pull operators in opposite directions. Chasing earlier referrals lengthens stays and, pushed far enough in the wrong mix, walks a hospice toward the cap; managing cap exposure by admitting late starves patients of the benefit's value and swells the short tail. A hospice sitting at the national averages, median stay 18 days, average 96, could be balanced, or could be running both failure modes at once, one tail concealing the other. The distribution, not the average, is the diagnosis.

What it means: length of stay is not a number, it is a shape. Two hospices with identical average stays can serve their communities in opposite ways, and every national length-of-stay statistic conceals the shape. The operator question is which tail is yours, and whose behavior, yours or your referral sources', is feeding it.

CMS already has a score on you

In April 2026, CMS published a program-integrity score for every hospice, by name, built from nine claims measures.

In 2026 CMS stopped looking only at the average and started scoring the individual hospice.

This study has argued that the national averages conceal the industry's real variation. In April 2026, CMS said, in effect, the same thing, and stopped looking only at the average. Alongside the fiscal 2027 proposed rule, CMS introduced the Service and Spending Variation Index, SSVI: a zero-to-16 score computed for every hospice from nine claims-based measures, published at the provider level, by CCN, for fiscal 2024 and fiscal 2025, covering more than 6,600 hospices in each year. Half the score is non-hospice spending, dollars Medicare pays outside the hospice benefit for enrolled patients, care the hospice was arguably paid to cover. The other half is one point per flag across eight utilization measures, and an operator should read the list slowly, because it is this study's findings restated as a scoring rubric: no continuous home care and no general inpatient provided in the year; 40 percent or more of routine home care days in a nursing facility; a bottom-quartile skilled-visit rate in the last two days of life; a top-quartile live-discharge rate; a top-quartile share of discharges past 180 days; bottom-quartile skilled nursing minutes per routine day; a bottom-quartile share of weekend days with a skilled visit; and live discharges returning to the same hospice within seven days.

CMS states the score's purpose plainly: to give beneficiaries a provider ranking and to target program-integrity work, medical review, education, and investigations that can end in payment suspension or revocation. The SSVI is proposed, its scores may be revised in the final rule, and it moves no payment rate today. But the scores are already public, refreshed annually with each final rule, and the direction is unmistakable. Level-of-care breadth, last-days visits, weekend staffing, live discharge, long stays, the whole terrain of Findings 4 through 8, now compiles into a single number attached to each hospice's name.

What it means: the era of hiding in the average is ending on the regulator's side too. National averages have always concealed the differences between hospices; in 2026 CMS stopped looking only at the average and started scoring the individual hospice. Every operator should know their score before a surveyor, a referrer, or a journalist knows it first.

The 2027 radar

Six items are in motion, each with one operator implication. The moratorium: nationwide, effective May 13, 2026 for six months, extendable, covering new enrollments and non-exempt majority-ownership changes; if your growth plan involves a new location or an acquisition structured as a change of majority ownership, it is frozen until CMS says otherwise. The SSVI: proposed with the fiscal 2027 rule, scores already published for every hospice; know yours before the final rule hardens the machinery. Rates: CMS finalized 2.6 percent for fiscal 2026 and proposes 2.4 percent for 2027 while MedPAC recommends zero, on a 17 to 0 vote; budget for the low end. The caps: $35,361.44 in fiscal 2026, proposed $36,210.11 for 2027, moving 2.4 percent while long-stay pressure compounds faster; a cap position that is tight this year gets tighter. HOPE: the assessment instrument replaced the HIS for admissions on or after October 1, 2025, and a hospice that misses the 90 percent on-time reporting threshold loses 4 percentage points of its update, turning 2026's raise into a 1.4 percent cut; reporting compliance is now a margin line. The Special Focus Program: suspended since February 2025 and still paused, which means the loudest enforcement tool is dormant while the quietest one, the SSVI, is already publishing; do not mistake the pause for a retreat.

The hospices that come through 2027 well will be the ones that can prove, with their own numbers, which side of each of these findings they are on: which margin, which cap position, which crisis capacity, which last-three-days number, which stay distribution, and which score.

The operator scorecard

Eight questions for your next leadership meeting, each answerable from your own data and the public files:

1. What is our SSVI score for fiscal 2024 and 2025, and which of the nine measures put points on it?

2. What is our cap utilization right now, estimated Medicare payments per beneficiary against $35,361.44, and how concentrated is our census in stays past 180 days?

3. What share of our patients had a nurse or social-worker visit on at least two of their final three days, and what is our weekend skilled-visit coverage?

4. If a patient crashes at 2 a.m. on a Saturday, can we escalate to continuous home care or a general inpatient bed tonight, and who owns that call?

5. What are our live discharges by reason code, and how many of those patients came back to us within days?

6. What is our short-stay share, admissions who die within a week, and which referral sources drive it?

7. Which margin are we benchmarking against, and is the cohort actually built like us on size, ownership, rurality, and payer mix?

8. Are our HOPE submissions complete and on time against the 90 percent threshold that protects our full payment update?

Where you actually stand

This study is national by design, and that is exactly its limit. Finding 1 shows why the national provider count is a four-state, for-profit-entry composite. Finding 2 shows why the aggregate margin is a policy-level measure and not yours. Finding 8 shows why even an honest average conceals the shape that matters. And Finding 9 shows that the federal government has itself moved from the average to the individual hospice. A national median cannot tell you whether your margin, your cap position, your level-of-care mix, your last-three-days rate, or your stay distribution is a problem, because it compares you to hospices that are nothing like yours.

The Astrelis Hospice Benchmark Report answers that for a single hospice: your filed cost report and your public quality measures, benchmarked against structurally similar hospices matched on size, ownership, rurality, and payer mix, with material operating signals quantified where the public data support the calculation, and every conclusion traceable to its source.

Source Overview

This edition is built on published federal data and independent analysis. Figures are attributed in the text to their originating source, and where sources measure different years or populations, the year and population are stated. The web edition links each entry to the verified source document.

1. MedPAC, Report to the Congress: Medicare Payment Policy, March 2026, hospice chapter (Chapter 10; 2023 and 2024 data; margins, cap, volume quintiles, level-of-care shares). medpac.gov

2. MedPAC, Report to the Congress: Medicare Payment Policy, March 2025, hospice chapter (Chapter 9; length of stay, live discharge, and setting). medpac.gov

3. MedPAC, Report to the Congress: Medicare Payment Policy, March 2016, hospice chapter (the decade-earlier provider count). medpac.gov

4. CMS, FY2026 Hospice Wage Index and Payment Rate Update final rule (CMS-1835-F), Federal Register, August 5, 2025. govinfo.gov

5. CMS, FY2027 Hospice Wage Index and Payment Rate Update proposed rule (CMS-1851-P), Federal Register, April 6, 2026, including the Service and Spending Variation Index (Section III.B.2). govinfo.gov

6. CMS, Hospice Service and Spending Variation Index: methodology overview and provider-level FY2024 and FY2025 score files. cms.gov

7. CMS, Announcement of Nationwide Temporary Moratorium on Enrollment of Hospices (CMS-6102-N), Federal Register, May 15, 2026, effective May 13, 2026. govinfo.gov

8. CMS and Abt Global, Hospice Monitoring Report, April 2026 (preliminary fiscal 2025 claims, extracted January 15, 2026). cms.gov

9. CMS and Abt Global, Hospice Monitoring Report, April 2025 (fiscal 2024 diagnosis and level-of-care detail). cms.gov

10. CMS Provider Data Catalog, Hospice provider and national quality files: HVLDL, the Hospice Care Index, level-of-care and site-of-service fields (May 2026 refresh; 2023 to 2024 claims window). data.cms.gov

11. CMS, Hospice Quality Reporting Program current measures. cms.gov

12. CMS, Period of Enhanced Oversight for New Hospices in Arizona, California, Nevada, Texas, Georgia and Ohio (MLN7867599, February 2026). cms.gov

13. 42 CFR 418.302 (payment procedures including the inpatient cap) and 42 U.S.C. 1395x(dd) (the 20 percent inpatient limit). ecfr.gov; uscode.house.gov

14. HHS OIG, Vulnerabilities in the Medicare Hospice Program Affect Quality Care and Program Integrity (OEI-02-16-00570, July 2018). oig.hhs.gov

15. California State Auditor, report on the state's oversight of hospice agencies, 2022, cited via MedPAC. auditor.ca.gov

Methods

One methods note in the spirit of this series. National hospice medians carry a four-state, for-profit-entry composition effect; the clearest instance is the cap, exceeded by 28 percent of hospices nationally in 2023 but only about 6 percent outside Arizona, California, Nevada, and Texas. Wherever this study cites a provider-weighted national figure, that effect applies, so treat it as context rather than a target. On scope, Medicare hospice cost reports cover freestanding hospices only, since provider-based hospices file inside their host hospital or nursing-facility reports; MedPAC's published margins span all hospice types and are labeled here by ownership, structure, and volume. No length-of-stay figure here is derived from public-use-file service days; every one comes from MedPAC or a CMS claims-based measure, because CMS states its hospice public-use file does not support a length-of-service calculation. CMS suppresses scores for hospices with too few cases, and a suppressed value is neither a zero nor a low score; suppressed hospices are excluded from the distributions above.

The layers do not share a year, and each figure names its own where it is used: provider counts are 2024 from MedPAC, with the fiscal 2021 to 2025 series from the CMS monitoring file; margins, cap, and quintile figures are 2023, held there by the lag to settle cap overpayments; HVLDL and the Hospice Care Index are the May 2026 refresh of a 2023 to 2024 claims window; level-of-care day and patient shares are 2023 and 2024 from MedPAC; diagnosis detail is fiscal 2024 and preliminary fiscal 2025 from the monitoring reports; policy is current fiscal 2026, with fiscal 2027 figures marked proposed. Every figure labeled preliminary comes from fiscal 2025 claims extracted January 15, 2026, which CMS cautions may be incomplete for the last half of the fiscal year; preliminary figures are updates inside findings, never a re-basing of the study, whose data line remains 2024 federal data and current 2026 policy. The two live-discharge rates cited use different bases and are labeled: MedPAC's 18.5 percent for 2023, and the monitoring file's 19.1 percent for preliminary fiscal 2025, which also counts transfers between hospices and moves out of the service area.

Tested and killed or corrected in this edition. From version 1: the claim that most hospices provide only one level of care did not survive, and the finding is reported on the two-denominator basis above; a single days-weighted national care-setting mix could not be sourced to a current vintage and setting is reported through MedPAC's length-of-stay figures instead. From the version 2 review: the reviewer's framing that fiscal 2025 marked the first decline after a decade could not be sourced, because the monitoring report's window begins in fiscal 2021, and the finding states the decline within that window only; the reviewer's description of the moratorium was verified against the hospice notice itself, which does not mention reactivations, so no reactivation claim renders; and version 1's benchmark-cohort description overstated the matching dimensions, corrected here to the four the shipped product uses: size, ownership, rurality, and payer mix.

About the author

Matt Borchardt is the founder of Astrelis. He has spent roughly twenty years in healthcare finance, about half of it inside critical access hospitals, working directly with Medicare cost reports and reimbursement. Astrelis builds national benchmarking and data products for rural and post-acute providers. Corrections and inquiries: astrelis.co/about

Published July 2026. Version 2.0. Data centered on the 2023 and 2024 federal reporting cycles and current fiscal 2026 policy, with preliminary fiscal 2025 figures labeled where they appear.

The Astrelis State of the Industry series is published annually. © Astrelis. Built in rural Minnesota on federal data and a working knowledge of what the numbers feel like from inside the building.