What a MELD Score Says About Hospice Timing

What MELD score qualifies for hospice?

No number does. The paper that introduced MELD presented it as a continuous severity index for prioritising organ allocation, not as a thresholded classifier, and it defined no cutoffs 1Ref 1Kamath PS, Wiesner RH, Malinchoc M, et al. (2001).A model to predict survival in patients with end-stage liver disease.MELD's origin and composition as originally published (serum bilirubin, serum creatinine, INR for prothrombin time, and etiology of liver disease); its direction (higher score means greater short-term mortality risk); its discrimination, with a c-statistic of 0.78 to 0.87 for three-month mortality across four independent validation cohorts; and that it was presented as a continuous severity index for prioritising organ allocation with no cutoffs defined.. Hospice eligibility works differently: it requires a physician to certify that a person's illness, if it runs its normal course, is likely to end life within six months 2Ref 2Centers for Medicare & Medicaid Services (2024).Medicare Hospice Benefits (CMS Product No. 02154).That hospice eligibility requires certification of a terminal prognosis of six months or less if the illness runs its normal course, that the benefit is structured as two 90-day periods followed by unlimited 60-day periods, and that a person may stop hospice at any time..

Those are two different kinds of statement. One is a risk estimate on a sliding scale. The other is a clinical judgment about a particular person, made by someone who has examined them and knows what the last several months have looked like.

A MELD score is evidence a clinician weighs. It is not a gate, and there is no figure at which a door opens.

This matters in practice, because families arrive at hospice eligibility for liver disease expecting a test that can be looked up and either passed or failed. Some coverage rules in other conditions do read a little that way. Liver disease does not, and the reasons are worth understanding, because they change what is useful to ask the hepatology team.

The benefit itself is built as a series of judgments rather than one verdict: two 90-day periods followed by unlimited 60-day periods, each requiring recertification, with the option to stop hospice at any point and elect it again later 2Ref 2Centers for Medicare & Medicaid Services (2024).Medicare Hospice Benefits (CMS Product No. 02154).That hospice eligibility requires certification of a terminal prognosis of six months or less if the illness runs its normal course, that the benefit is structured as two 90-day periods followed by unlimited 60-day periods, and that a person may stop hospice at any time..

What does a MELD score actually measure?

MELD — the Model for End-Stage Liver Disease — was published in 2001 as a way to predict survival in people with end-stage liver disease. As originally described, it combined serum bilirubin, serum creatinine, the international normalised ratio for prothrombin time, and the cause of the liver disease 1Ref 1Kamath PS, Wiesner RH, Malinchoc M, et al. (2001).A model to predict survival in patients with end-stage liver disease.MELD's origin and composition as originally published (serum bilirubin, serum creatinine, INR for prothrombin time, and etiology of liver disease); its direction (higher score means greater short-term mortality risk); its discrimination, with a c-statistic of 0.78 to 0.87 for three-month mortality across four independent validation cohorts; and that it was presented as a continuous severity index for prioritising organ allocation with no cutoffs defined.. Three of those four are laboratory values from an ordinary blood draw.

INR, the international normalised ratio, measures how long blood takes to clot. A failing liver makes fewer clotting proteins, so the INR rises as liver function falls.

Direction is the simple part: higher is worse, meaning greater short-term mortality risk 1Ref 1Kamath PS, Wiesner RH, Malinchoc M, et al. (2001).A model to predict survival in patients with end-stage liver disease.MELD's origin and composition as originally published (serum bilirubin, serum creatinine, INR for prothrombin time, and etiology of liver disease); its direction (higher score means greater short-term mortality risk); its discrimination, with a c-statistic of 0.78 to 0.87 for three-month mortality across four independent validation cohorts; and that it was presented as a continuous severity index for prioritising organ allocation with no cutoffs defined.. What the score does well is discriminate between people. Across four independent validation cohorts — hospitalised patients with decompensated disease, ambulatory noncholestatic cirrhosis, primary biliary cirrhosis, and a historical cirrhosis cohort — its c-statistic for three-month mortality ran from 0.78 to 0.87 1Ref 1Kamath PS, Wiesner RH, Malinchoc M, et al. (2001).A model to predict survival in patients with end-stage liver disease.MELD's origin and composition as originally published (serum bilirubin, serum creatinine, INR for prothrombin time, and etiology of liver disease); its direction (higher score means greater short-term mortality risk); its discrimination, with a c-statistic of 0.78 to 0.87 for three-month mortality across four independent validation cohorts; and that it was presented as a continuous severity index for prioritising organ allocation with no cutoffs defined..

MELD's discrimination for three-month mortality was 0.78 to 0.87 across four separate validation cohorts 1Ref 1Kamath PS, Wiesner RH, Malinchoc M, et al. (2001).A model to predict survival in patients with end-stage liver disease.MELD's origin and composition as originally published (serum bilirubin, serum creatinine, INR for prothrombin time, and etiology of liver disease); its direction (higher score means greater short-term mortality risk); its discrimination, with a c-statistic of 0.78 to 0.87 for three-month mortality across four independent validation cohorts; and that it was presented as a continuous severity index for prioritising organ allocation with no cutoffs defined..

That is strong performance for a prognostic model, and it is worth being precise about what it means. The score is good at ranking a group of people by risk. Sorting a group correctly is a different achievement from telling one person in that group what is going to happen to them.

MELD has been adapted more than once since 2001, including for organ allocation, so a calculator found online may not be the formulation the original paper described. The direction has not changed.

Why there is no MELD cutoff for hospice

There are two reasons, and they compound. The first is that the instrument was never built with one. MELD was designed as a continuous index for deciding who should be prioritised for an organ, not as a classifier separating eligible from ineligible, and its defining publication reports no threshold anywhere 1Ref 1Kamath PS, Wiesner RH, Malinchoc M, et al. (2001).A model to predict survival in patients with end-stage liver disease.MELD's origin and composition as originally published (serum bilirubin, serum creatinine, INR for prothrombin time, and etiology of liver disease); its direction (higher score means greater short-term mortality risk); its discrimination, with a c-statistic of 0.78 to 0.87 for three-month mortality across four independent validation cohorts; and that it was presented as a continuous severity index for prioritising organ allocation with no cutoffs defined..

The second is that Medicare's guidance does not work from a lab value alone. The Local Coverage Determination on determining terminal status lays out non-disease-specific criteria — functional decline, nutritional decline, the accumulated weight of other conditions — alongside disease-specific ones, and frames all of it as documentation supporting a prognosis rather than as absolute cutoffs 3Ref 3Centers for Medicare & Medicaid Services (Medicare Administrative Contractor LCD) (2023).Local Coverage Determination (LCD): Hospice - Determining Terminal Status (L33393).The framework of Medicare hospice eligibility documentation: non-disease-specific criteria (functional decline, nutritional decline, comorbidities) alongside disease-specific criteria, including the end-stage liver disease pattern of impaired synthetic function with complications and the person's transplant status, all treated as guidance supporting a prognosis rather than absolute cutoffs..

So even where a specific figure appears in coverage guidance, it functions as one element in a picture. A person can sit past a quoted value and not be certified, because the rest of the picture does not support a six-month prognosis. A person can sit short of it and be certified, because everything else does.

If a hepatology team quoted a number and a hospice said something that did not match it, nobody made a mistake. The two are answering different questions under different rules.

What Medicare's liver criteria actually look at

Instead of a single score, the liver-disease criteria describe a pattern: blood tests showing the liver is no longer making what the body needs, the complications that mark end-stage disease, and the person's transplant status 3Ref 3Centers for Medicare & Medicaid Services (Medicare Administrative Contractor LCD) (2023).Local Coverage Determination (LCD): Hospice - Determining Terminal Status (L33393).The framework of Medicare hospice eligibility documentation: non-disease-specific criteria (functional decline, nutritional decline, comorbidities) alongside disease-specific criteria, including the end-stage liver disease pattern of impaired synthetic function with complications and the person's transplant status, all treated as guidance supporting a prognosis rather than absolute cutoffs.. The complications on that list are ones most families already recognise, because they are usually why the past year has been so hard.

What tends to appear in the documentation:

• Ascites that keeps returning despite treatment, or that no longer responds to the medications which used to control it.
• Spontaneous bacterial peritonitis — infection of that abdominal fluid.
• Hepatorenal syndrome — kidney failure arising out of the liver disease itself.
• Hepatic encephalopathy — confusion, reversed sleep, or unresponsiveness caused by toxins a failing liver no longer clears.
• Recurrent variceal bleeding that continues despite treatment.
• Impaired clotting and a low albumin on blood tests, describing lost synthetic function.

Transplant status sits alongside all of it, because hospice and an active transplant listing point in opposite directions. When transplant is off the table — someone was evaluated and turned down, was never a candidate, or has decided not to pursue it — that single fact reshapes the eligibility conversation more than any laboratory value in the chart 3Ref 3Centers for Medicare & Medicaid Services (Medicare Administrative Contractor LCD) (2023).Local Coverage Determination (LCD): Hospice - Determining Terminal Status (L33393).The framework of Medicare hospice eligibility documentation: non-disease-specific criteria (functional decline, nutritional decline, comorbidities) alongside disease-specific criteria, including the end-stage liver disease pattern of impaired synthetic function with complications and the person's transplant status, all treated as guidance supporting a prognosis rather than absolute cutoffs..

Where a MELD score genuinely helps

Its real use here is timing rather than qualification. MELD and Child-Pugh are both used prognostically in decompensated cirrhosis, and a review of palliative and end-of-life care in that population describes a symptom burden that is heavy and frequently undertreated 4Ref 4Peer-reviewed review (see article) (2023).Palliative Care and End of Life Care in Decompensated Cirrhosis.The high and often undertreated symptom burden of decompensated cirrhosis and the palliative-care needs it creates, and the prognostic use of MELD and Child-Pugh in this population.. A worsening score is one of the few concrete prompts that the conversation about goals ought to be happening now instead of later.

The symptoms are the real argument. Ascites that needs repeated draining, itch that keeps someone awake, muscle wasting, exhaustion, the disorientation of encephalopathy, and the fear that comes with each of them — this is precisely the territory a palliative or hospice team is built for, and it is treatable territory 4Ref 4Peer-reviewed review (see article) (2023).Palliative Care and End of Life Care in Decompensated Cirrhosis.The high and often undertreated symptom burden of decompensated cirrhosis and the palliative-care needs it creates, and the prognostic use of MELD and Child-Pugh in this population..

A rising MELD is a prompt to talk, not a countdown. The question it should trigger is what the next six months are meant to look like, not how many of them there are.

Worth putting to a hepatology team directly: what does this trend suggest about the coming months, what would change it, and what would it take to have a palliative-care referral now regardless of how the transplant question resolves.

Hospice usually comes late in liver disease

Later than it needs to, and the pattern is documented rather than anecdotal: people with end-stage liver disease in the United States are frequently referred to hospice late in the course of illness, at a point where much of what the benefit offers has little time left to work 5Ref 5Peer-reviewed study (see article) (2021).Hospice Care for End Stage Liver Disease in the United States.That patients with end-stage liver disease in the United States are frequently referred to hospice late in the course of illness, and the access and utilisation barriers behind that pattern.. Some of the reasons are structural — a disease trajectory that lurches rather than declines steadily, the hope a transplant evaluation holds open, and a benefit that does not sit comfortably beside an active listing.

One way out of that bind is worth knowing. Palliative care is not hospice: it can be given alongside treatment intended to cure or prolong life, at any stage of a serious illness 6Ref 6National Institute on Aging (NIH) (2024).Frequently Asked Questions About Palliative Care.That palliative care differs from hospice in that it can be provided alongside treatment intended to cure or prolong life, at any stage of a serious illness.. Someone waiting on a transplant list can have a palliative team managing symptoms and holding the difficult conversations without surrendering anything and without any decision being forced early.

Deciding when to choose a hospice, in liver disease, usually turns on whether treatment aimed at the liver is still doing more good than harm, and on whether the person — asked plainly, while they can still answer — would trade some possible time for more comfortable time. Those are answerable questions. A MELD score helps a team ask them at the right moment, which is a better use of it than searching it for a line it does not have.