Hospice oversight guards the door but rarely asks the patient. CMS has watched new hospices more closely since 2023. It has also reviewed their claims before paying them in the states where fraud clustered. In 2026 it suspended payments to hundreds of California hospices. On May 13, 2026, it stopped enrolling new hospices nationwide for 6 months.
Each of those tools works on the hospice. The person best placed to say whether a hospice election is real is the patient. The hospice must tell Medicare of each election within 5 days. Medicare’s own summary of claims reaches patients by mail every 6 months.
Hospice is a large benefit. Medicare spent $28.3 billion on hospice in 2024, and 52.9% of Medicare decedents used it, according to MedPAC. The fraud has clustered in a handful of states, and mostly among newly enrolled hospices. The check proposed at the end starts with the patient.
Key takeaways
- Hospice is a large benefit. Medicare spent $28.3 billion on hospice for 1.82 million people in 2024.
- Fraud clustered in 4 states. Enrolled hospices grew 126% in California and 151% in Nevada from 2019 to 2023, according to CMS.
- Targeted oversight found problems often. CMS revoked 122 of the roughly 670 new hospices it reviewed under enhanced oversight. That is 18%, against 1% to 3% for providers generally.
- The response has moved to the door. A nationwide 6-month moratorium on new hospice enrollment took effect May 13, 2026.
- A new index scores variation, not fraud. The SSVI scores every hospice on 9 claims measures. CMS says a high score does not necessarily mean a hospice is acting improperly.
- The patient is the missing check. A notice from Medicare to the patient at election would catch enrollments made without the patient’s knowledge.
Why hospice draws fraud
Hospice is for people with a terminal illness who choose comfort care over treatment aimed at a cure. A hospice physician must certify for the first benefit period that the patient’s life expectancy is 6 months or less if the illness runs its normal course. The patient’s attending physician must also certify if the patient has one. The certification must include a brief narrative of the clinical findings behind the prognosis, under 42 CFR § 418.22.
The patient then signs an election statement. It waives Medicare payment for treatment of the terminal illness outside the hospice, under 42 CFR § 418.24. The hospice receives a daily payment for every day a patient is enrolled, according to MedPAC. The payment does not depend on a visit that day.
That design rewards long stays. Stays longer than 180 days accounted for just over 60% of Medicare hospice spending in 2024, about $17.5 billion of $28.3 billion, according to MedPAC. MedPAC notes that hospices can identify and enroll patients likely to have long, more profitable stays if they choose to.
Long stays are not proof of anything. Patients admitted for neurological conditions averaged 169 days of hospice care over their lifetimes, against 51 days for cancer, according to MedPAC. Prognosis is hard for illnesses that decline slowly. Patients may also change their minds and leave hospice, which MedPAC counts among the legitimate reasons for a live discharge.
The market has grown with the incentive. There were 6,706 hospices in 2024, about 82% of them for profit, according to MedPAC. For-profit hospices served 60% of patients. Their patients’ stays averaged 120 days, against 71 days at nonprofit hospices. About 28% of hospices exceeded Medicare’s aggregate cap on average payment per patient in 2023. More than half of California’s hospices did. MedPAC writes that the patterns at hospices over the cap suggest they admit patients who do not meet the eligibility criteria.
The harm to patients
Hospice fraud is not only a billing problem. A person enrolled in hospice without being terminally ill gives up Medicare coverage for treatment of the condition the hospice listed. ProPublica reported in January 2024 on a woman with dementia in a Los Angeles County nursing home who was enrolled by a doctor her daughter had never heard of. Her physical therapy stopped. A check of her pacemaker was delayed. Medicare had paid the hospice more than $7,500, according to ProPublica.
OIG has described patients enrolled while ineligible and without their knowledge or consent, according to Hospice News. Those cases are hard to find in claims, because the paperwork can look complete. They are easy for the patient or family to recognize.
Where the fraud concentrated
CMS’s own data show how fast the market grew in 4 states. The figures below come from the notice that imposed the nationwide moratorium, published in the Federal Register on May 15, 2026.
| State | Growth in enrolled hospices from 2019 to 2023 | New enrollments in 2023 | New enrollments in 2025 |
|---|---|---|---|
| Arizona | 105% | 53 | 3 |
| California | 126% | 362 | 29 |
| Nevada | 151% | Not reported | Not reported |
| Texas | 51% | 164 | 67 |
MedPAC counted 1,046 hospices added in California from 2019 to 2023, growth of 19% a year. Entry slowed sharply in 2024. California added 38 hospices that year, and Arizona added 1.
Many of the new hospices had little behind them. In 2023, Medicare certified 15 more hospices in a 2-story Los Angeles building that already housed more than 100, according to ProPublica. In CMS data cited by ProPublica, the 4 states accounted for about two-thirds of new hospice certifications nationwide. OIG has pointed to multiple hospices operating from 1 address. It has also pointed to individuals managing several hospices at once, according to Hospice News.
Live discharges tell a similar story. Nationally, 19.1% of hospice discharges in 2024 were live discharges, according to MedPAC. A Texas Senate committee heard testimony about Texas hospices with live discharge rates of 90% or higher, according to Morgan Lewis.
Georgia shows the pattern can spread. Its hospice count rose 24% from 2019 to 2025 while its Medicare population grew about 6%, according to CMS. CMS said in January 2026 that it would add Georgia and Ohio to its enhanced oversight, according to Hospice News.
What CMS has built since 2023
CMS has added tools in stages. Most of them act on who may enroll and bill.
| Tool | Since | What it does | Where |
|---|---|---|---|
| Enhanced oversight of new hospices | July 2023 | Medical review of claims from new hospices for 30 days to 1 year | First 4 states, now 6 |
| High-risk screening | 2024 | New hospices get Medicare’s strictest enrollment screening | Nationwide |
| 36-month rule | 2024 | A majority ownership change within 36 months of enrollment requires a new enrollment | Nationwide |
| Prepayment review of existing hospices | September 2024 | Claims reviewed before payment | The 4 states |
| Special Focus Program | Finalized 2023, stopped February 2025 | Meant to single out poor performers for added oversight | Nationwide |
| Payment suspensions | April 2026 | Payments held on a credible allegation of fraud | Hundreds in California |
| Enrollment moratorium | May 13, 2026 | No new hospice enrollments for 6 months | Nationwide |
| Spending variation index | July 2026 | A public score on 9 claims measures | Every hospice |
| Election addendum for all | October 1, 2026 | A written list of what the hospice will not cover | Every new election |
The screening tier and the 36-month rule are described in MedPAC’s report.
Enhanced oversight of new hospices
Enhanced oversight has the clearest record. CMS reviews claims from newly enrolled hospices under it for 30 days to 1 year. Hospices that change ownership get the same review. About 670 hospices had been through that review as of June 2025. CMS had revoked 122 of them.
CMS called that an 18% revocation rate in the moratorium notice. Its usual rate for providers is roughly 1% to 3%. Across the original 4 states, enhanced oversight had led to more than 200 revocations by April 2026, according to CMS.
That record shows a targeted review of new entrants finds problems at a high rate. It also shows the limit of the approach. Review began only after a hospice had enrolled and started to bill.
Payment suspensions
A federal task force suspended 447 hospices in the Los Angeles area in April 2026, according to Hospice News. It also suspended 23 home health agencies. The suspected fraud was estimated at $600 million. The White House created the task force in March, and Vice President JD Vance leads it. By May 13, officials cited 800 hospice payment suspensions in California, according to Morgan Lewis.
A payment suspension is a hold, not a finding. CMS may suspend payments when it has a credible allegation of fraud. It must review each suspension every 180 days, according to an analysis by Duane Morris. The provider can submit a rebuttal under 42 CFR § 405.374. CMS reportedly relied in part on reviews of billing data assisted by artificial intelligence.
That posture fits a signal from a model. As we argued in our piece on anomaly scores, a score should start a review rather than end one. A suspension that rests on a data pattern needs a person to read the records behind it before it becomes a revocation.
The nationwide moratorium
The moratorium rests on section 1866(j)(7) of the Social Security Act. Its rules at 42 CFR § 424.570 allow a pause when CMS finds a significant potential for fraud or abuse. The notice blocks initial Medicare enrollment of hospices in every state. It also blocks ownership changes that count as new enrollments under the 36-month rule. Applications received before May 13 continue to be processed. CMS left Medicaid moratoria to each state.
CMS’s notice pointed to enrollment growth far ahead of the Medicare population. CMS Administrator Mehmet Oz warned in April that fraud could shift to other states as enforcement grew in California, according to Morgan Lewis. The agency also cited fraud cases in Oklahoma and Utah, states it had treated as lower risk. CMS wrote that it found little evidence of access problems nationwide and did not expect the moratorium to threaten access.
The industry objected. The National Alliance for Care at Home said “an enrollment moratorium does not distinguish between bad actors and compliant providers,” in a statement on May 13. It warned of longer waits in rural and underserved areas. LeadingAge asked CMS in a September letter not to renew the moratorium. It also asked for exceptions in areas with clear need.
CMS’s own figures complicate the case for a national pause. California’s licensing moratorium had been in place since 2022, yet 362 hospices enrolled in Medicare there in 2023. The steep drop to 29 came after federal review and screening tightened. Arizona fell from 53 new enrollments to 3 over the same years.
A national pause adds restriction mainly in states outside the original 4. CMS argues that fraud moves, and the Georgia figures support that worry. The question is whether a pause on every new hospice in every state works better than the targeted review that revoked 18% of the new hospices it examined.
The new index and its limits
The fiscal 2027 hospice rule created a Hospice Service and Spending Variation Index, known as the SSVI. CMS issued the rule on July 30, 2026, according to a CMS fact sheet. The index scores every hospice on 9 claims measures, listed in the proposed rule.
- The share of patients discharged after a stay of 180 days or more
- The share of all discharges that were live discharges
- The total number of discharges
- Average minutes of direct care per routine home care day
- Average visits per routine home care day
- The share of weekend routine home care days with at least 1 skilled visit
- Non-hospice spending per day
- The share of live discharges who return to the same hospice within 7 days
- Total non-hospice spending
A hospice gets 1 point for most measures if it falls in the top or bottom 25% of the distribution. Total non-hospice spending is scored in tiers, so higher spending earns more points. Scores run from 0 to 16. CMS published scores for fiscal 2024 and 2025 with the final rule. The proposed rule’s fiscal 2025 data covered 6,642 hospices. CMS plans to update the scores each year.
A high score may lead to medical review or an investigation, according to the proposed rule. Either can end in payment suspension or revocation if CMS finds fraud. CMS is careful about what the score means. A high score does not necessarily signal that a hospice is acting improperly, the agency wrote. Its April press release said the data are not a direct indicator of fraud.
The design will flag honest hospices. When a threshold sits at the top or bottom 25%, a quarter of hospices cross it by definition. A hospice that serves many patients with neurological conditions will tend to have long stays, because those patients live longer in hospice. MedPAC’s figures of 169 days for neurological conditions and 51 for cancer make the point.
CMS has been here before. It stopped its previous program for singling out hospices, the Special Focus Program, in February 2025 so it could evaluate the program further, according to the AHA. Hospices and state associations had sued in January, arguing that its data did not reliably separate fraudulent providers from good ones, according to HomeCare. The SSVI invites the same objection unless it accounts for the patients a hospice serves.
Some measures are sturdier than others. A hospice that bills every day but rarely visits will show it in minutes and visits per day, whatever its patients’ diagnoses. Those measures track the care a daily payment is meant to buy.
Medicare hears about a hospice election within 5 days. The patient can wait 6 months to hear from Medicare.
A check that starts with the patient
Every tool above works on the hospice. The patient is the one party who knows at once whether they chose hospice, and the system rarely asks.
The patient’s choice is documented, but by the hospice. The patient or a representative signs the election statement, which the hospice keeps. For elections from October 1, 2026, the hospice must also give every patient a written addendum within 5 days. It lists what the hospice will not cover because it considers it unrelated to the terminal illness, under 42 CFR § 418.24. Both documents come from the hospice, which is the party a fraudulent scheme controls.
Medicare learns of each election quickly. The hospice must file a notice of election within 5 calendar days under the same rule. Days before a late filing are not paid. Medicare’s own summary of claims is mailed every 6 months, according to Medicare.gov. Beneficiaries who choose electronic notices get an email for any month with a processed claim.
Medicare should send its own notice to the patient as soon as a hospice files an election. It would go to the address Medicare has on file rather than one the hospice supplies. A copy would go to any representative on record. In plain language it would name the hospice and the date the election took effect. It would explain that hospice care aims at comfort, and that Medicare stops paying for treatment of the terminal illness outside the hospice. It would give a number to call for anyone who did not choose hospice.
A response that says the patient did not choose hospice is a strong lead. A person should follow it up by calling the patient or family. The notice itself should never stop care.
The notice targets the schemes that do the most harm. The family in ProPublica’s report did not know how the enrollment had happened. A notice from Medicare would have told them within days that it had. The idea is the one behind a bank’s fraud alert. The customer is the fastest judge of whether a charge is theirs.
There are real limits. Many hospice patients are very ill or have dementia. The notice must reach a representative in those cases, and it must be written with care for a family at a hard time. Silence proves nothing. The notice can produce leads but cannot confirm consent. A scheme could try to change the address Medicare holds, so an address change shortly before an election should itself raise a flag.
The record check OIG proposed
The second check reads the patient’s record. OIG audited a sample of fiscal 2021 hospice stays among new enrollees who had no inpatient or emergency room claims in the 18 months before hospice. Documentation for 45 of 100 sampled periods did not meet Medicare’s requirements. In 21 of them, the clinical information did not support a terminal illness, according to OIG’s June 2026 audit. OIG estimated that Medicare could have saved $255.1 million with eligibility reviews for this group.
CMS agreed to have its contractors consider reviews of these enrollees before or after payment. OIG expects an update by December 17, 2026. This is the kind of screen we described in our piece on prepayment review. A patient with no hospital visits for 18 months may still be dying. More than half of OIG’s sample met the requirements. The pattern is a reason to read the certification narrative before paying, not a reason to deny.
Together the 2 checks cover the 2 ways an enrollment can be false. The notice asks the patient whether they chose hospice. The record check asks whether a physician’s narrative supports a prognosis of 6 months or less. Neither waits for a hospice to bill for a year before anyone looks.
Where Redlake stands
Redlake builds tools that health plans use in utilization review. We apply the same rules to fraud review that we apply to coverage decisions. A machine can say yes, but only a person can say no.
Ask the patient. Medicare should send its own notice of every hospice election within days of the filing. A response that the patient did not choose hospice should go to a person for follow-up, never to an automatic denial.
Review the records the data point to. CMS should adopt OIG’s screen for new enrollees without recent hospital or emergency claims as a targeted review before payment. More than half of OIG’s sample qualified, so the review should pay promptly when the narrative supports the prognosis.
Adjust the index before acting on it. Every recommendation should show its reasons. The SSVI should account for diagnosis before CMS acts on a score. A hospice that serves patients with dementia should not collect points for doing its job.
Target the door. Enhanced oversight of new hospices revoked 18% of those it reviewed. That record argues for extending targeted review to any state where enrollment outruns the Medicare population. CMS has already done so for Georgia and Ohio. When the nationwide moratorium comes up for renewal in November, we think targeted tools should replace it.
Keep the record. Every decision should be possible to reconstruct. A suspension or revocation should carry the evidence behind it. That lets a hospice rebut a specific finding and lets an auditor rebuild the decision later.
What to watch in the rest of 2026
- The moratorium decision. The 6-month moratorium runs to mid-November unless CMS extends it with a new notice.
- The addendum. Hospices must give the addendum to every patient who elects hospice from October 1, 2026.
- OIG’s screen. CMS owes OIG an update on reviews of new enrollees by December 17, 2026.
- Use of the SSVI. Watch whether CMS adjusts scores for diagnosis before it acts on them.
- Where fraud moves. Watch new enrollment and enforcement in Georgia and Ohio. Watch Texas too, where state senators heard testimony on very high live discharge rates.
- Suspension outcomes. Watch how many California suspensions end in revocation and how many are lifted.
About this analysis
Method. CMS’s hospice program integrity actions from 2023 through September 2026 are read against MedPAC’s data on the benefit, mainly from CMS documents and MedPAC’s March 2026 report. Reporting by Hospice News and ProPublica fills in enforcement details and patient cases.
Data notes. Enrollment growth figures are CMS’s, from the moratorium notice. Enhanced oversight figures are as of June 2025. The suspension counts come from news reports and a law firm summary of a White House briefing, not from a CMS data release. The patient notice is this analysis’s proposal.
Sources. MedPAC March 2026 report, chapter 10 (March 2026). CMS moratorium notice CMS-6102-N (May 15, 2026). CMS release on the fiscal 2027 proposed rule (April 2, 2026). Fiscal 2027 proposed rule in the Federal Register (April 6, 2026). CMS fact sheet on the fiscal 2027 final rule (July 30, 2026). OIG audit A-06-22-09003 (June 2026). ProPublica (January 25, 2024). Hospice News (April 28, 2025). Hospice News (January 12, 2026). Hospice News (April 15, 2026). Morgan Lewis (May 4, 2026). Morgan Lewis (May 14, 2026). Duane Morris (May 5, 2026). AHA (February 19, 2025). HomeCare (February 18, 2025). National Alliance for Care at Home (May 13, 2026). LeadingAge letter (September 2026). Medicare.gov page on the Medicare Summary Notice. Federal regulations at 42 CFR § 418.22 and § 418.24.
Frequently asked questions
Why is hospice a target for fraud?
Medicare pays hospices for each day a patient is enrolled, and long stays are the most profitable. Eligibility rests on a physician’s prognosis, which is hard to check from claims. A hospice that enrolls patients who are not terminally ill can bill for months.
What is the nationwide hospice moratorium?
Since May 13, 2026, CMS has not accepted new Medicare enrollment applications from hospices in any state. The pause lasts 6 months and can be extended in 6-month steps. Hospices already enrolled are not affected by it.
What is the new hospice spending index?
The Hospice Service and Spending Variation Index scores each hospice from 0 to 16 on 9 claims measures. They include the share of long stays and the number of visits per day. CMS published the first scores with its fiscal 2027 rule in July 2026.
Does a high SSVI score mean fraud?
No. CMS says a high score does not necessarily mean a hospice is acting improperly. A high score may lead to medical review or an investigation. Hospices that serve many patients with slow illnesses can score high for legitimate reasons.
How can patients spot an enrollment they did not choose?
Patients and families can read the Medicare Summary Notice, which lists hospice claims. Paper notices come every 6 months, and electronic notices come for any month with a claim. Anyone who sees a hospice they did not choose can call 1-800-MEDICARE.
What did OIG find about new hospice enrollees?
OIG looked at fiscal 2021 stays for new enrollees with no hospital or emergency room claims in the 18 months before hospice. In its sample, 45 of 100 periods did not meet Medicare’s requirements. It estimated $255.1 million in possible savings from targeted eligibility reviews.