How to qualify for hospice

Federal regulation (42 CFR 418.20) makes a person eligible for the Medicare hospice benefit when they are "entitled to Part A of Medicare" and "certified as being terminally ill." Terminally ill is defined next door, in 418.3: "a medical prognosis that his or her life expectancy is 6 months or less if the illness runs its normal course." The third piece is the election statement (418.24), which the person or their representative signs. It acknowledges "the palliative rather than curative nature of hospice care" and waives Medicare payment for treatment aimed at curing the terminal illness.

That is the entire list. There is no age floor beyond having Medicare. There is no requirement to be bedbound, to be in pain, to have stopped eating, or to have a specific diagnosis. Cancer is what most people picture, but heart failure, dementia, COPD, kidney failure, liver disease, stroke, and ALS all qualify on the same rule.

There is also no requirement to sign a DNR. A separate regulation (42 CFR 489.102) lists hospices among the providers that may not "condition the provision of care or otherwise discriminate against an individual based on whether or not the individual has executed an advance directive." A hospice will ask about code status, and many people on hospice choose one. But it is a conversation, not a condition. If you want the full picture of what each order does, our DNR vs DNI guide covers it.

What the person gives up is narrower than families fear. The waiver covers care "related to the terminal illness and related conditions" from providers other than the hospice. Medicare's own booklet says it will "still pay for covered benefits for any health problems that aren't part of your terminal illness and related conditions." Someone with lung cancer on hospice still gets their diabetes managed.

The hospice itself covers medication, equipment, nursing, aides, social work, chaplaincy, and four levels of care, including short inpatient stays for symptoms that cannot be managed at home. The palliative care vs hospice guide walks through what is and is not covered.

For the first benefit period, the certification needs two doctors (42 CFR 418.22(c)). One is "the medical director of the hospice, the physician designee, or the physician member of the hospice interdisciplinary group." The other is "the individual's attending physician, if the individual has an attending physician." Medicare's booklet puts it in plain words: "your hospice doctor and your regular doctor (if you have one)."

Two things in that wording matter for families. First, the hospice has its own doctor, and that doctor does half the certifying. A hospice evaluation is not a formality that rubber-stamps what the oncologist already decided. It is an independent clinical judgment by a physician who assesses people for this every week.

Second, "attending physician" is whoever the person identifies as having "the most significant role" in their care, and the regulation allows it to be a physician, a nurse practitioner, or a physician assistant. It does not have to be the specialist who has been treating the disease. If the cardiologist will not engage but the primary care doctor sees the decline, the primary care doctor can be the attending for this purpose. The attending must genuinely be the clinician with the most significant role in the person's care, not whoever is willing to sign.

After the first period, the attending physician's signature is no longer required. Recertification for every later period needs only the hospice physician (418.22(c)(2)). So the specialist's reluctance is a barrier once, at the door, not an ongoing one.

Each certification has to include "a brief narrative explanation of the clinical findings that supports a life expectancy of 6 months or less." That narrative is where the records you hold come in, and step 4 is about them.

Medicare contractors publish the criteria they use to judge whether a certification is supported, in a document called a Local Coverage Determination. The one that covers most of the country is titled "Hospice - Determining Terminal Status" (LCD L34538 and its siblings). It is written for hospice medical directors. Here is what it says, translated.

It starts with a general picture that applies to any diagnosis: a functional score below 70 percent on the Palliative Performance Scale (the person spends most of the day sitting or lying down and needs help with most activities) and dependence on help with two or more of feeding, walking, continence, transferring, bathing, and dressing. Then it lists signs of decline: "recurrent or intractable infections such as pneumonia, sepsis or upper urinary tract," "decreasing serum albumin or cholesterol," "increasing emergency room visits, hospitalizations, or physician's visits related to hospice primary diagnosis," and progressive weight loss. The document also says, in so many words, that "lab testing is not required to establish hospice eligibility."

Dementia is the diagnosis families most often ask about, and the criteria are specific. Stage 7 on the Functional Assessment Staging scale: speech limited to six or fewer intelligible words, unable to walk without help, and in the later sub-stages unable to sit up or smile. Plus one complication in the past year: aspiration pneumonia, a kidney infection, sepsis, multiple stage 3 or 4 pressure sores, fever that keeps returning after antibiotics, or "10% weight loss during the previous six months or serum albumin less than 2.5 gm/dl" from not being able to eat and drink enough.

A caution worth knowing: these dementia criteria are poor at predicting who will actually die within six months. In a study of 606 nursing-home residents with advanced dementia (Mitchell and colleagues, JAMA, 2010), the hospice criteria caught only 20 percent of those who died within six months. The criteria are a floor for eligibility, not a test of whether someone is close to the end.

Heart failure: New York Heart Association class IV, meaning symptoms at rest, after the medications have been optimized or the person is not a candidate for (or has declined) procedures. An ejection fraction of 20 percent or below supports it, "but is not required if not already available." COPD: "disabling dyspnea at rest," a "bed to chair existence," with low oxygen at rest (saturation 88 percent or below on room air) or high carbon dioxide, and unintentional weight loss over 10 percent in six months. An FEV1 under 30 percent of predicted is supporting evidence, "but is not necessary to obtain."

Cancer: metastatic disease at diagnosis, or progression to metastatic disease with either "a continued decline in spite of therapy" or the person declining further disease-directed treatment. The document adds that "certain cancers with poor prognoses (e.g. small cell lung cancer, brain cancer and pancreatic cancer) may be hospice eligible without fulfilling the other criteria." Note what is not on the list: being off chemotherapy is not required. What is required is that the plan is no longer aimed at cure.

One more line from the document that matters more than it looks: people who "stabilize or improve while receiving hospice care, yet have a reasonable expectation of continued decline for a life expectancy of less than six months, remain eligible." Getting better on hospice, which can happen once the hospital visits stop and the symptoms are managed, is not by itself a reason to lose it.

Every decline indicator in step 3 corresponds to a number or a line in a document most families already hold. Nobody tends to tell families this. Here is the map.

Weight is on every after-visit summary and most discharge summaries, and the criterion is a 10 percent loss over six months. A person who weighed 150 pounds in February and 134 in August has crossed that threshold. Pull the visit summaries from the last six months and write the weights in a row.

Albumin is a line on a comprehensive metabolic panel, and the threshold that appears in the criteria is 2.5 g/dL, with a note that it "should not be used by itself." Ejection fraction is on the echocardiogram report. Oxygen saturation is in every set of vital signs. FEV1 is on pulmonary function tests.

The count of ER visits and hospital admissions in the past six months is the one families underestimate most, because each visit felt like its own crisis. Go through the discharge papers and count. "Increasing emergency room visits, hospitalizations" is a named criterion, and three in six months reads very differently from one.

Function is in the nursing notes and physical therapy notes from any hospital stay or home health episode. Look for phrases like "requires assistance with," "max assist," "bed to chair," "two-person transfer," and the activities of daily living checklist. Two or more dependencies is the baseline. For dementia, the FAST stage is often written into the neurology or geriatrics note directly.

Put these in one place, in date order, before you ask for the evaluation. The hospice doctor will ask for exactly this, and the certification narrative is built from it. Our guide to organizing medical records covers how to get the discharge summaries and labs out of three portals and into one folder, and how to read your lab results explains the panel the albumin sits on.

You do not need a referral to call a hospice. The Hospice Foundation of America, a nonprofit that does not run hospices, says that "patients or their loved ones may self-refer." Medicare's booklet describes a "one-time only consultation with a hospice medical director or hospice doctor to talk about your care options," covered by Medicare before any election. Hospices do these visits routinely, because the evaluation is how they admit people. Call, say "I would like an evaluation for hospice eligibility," and give them the attending physician's name.

The visit itself is usually a nurse, sometimes with a social worker, at the person's home or bedside. They review the records (bring the folder from step 4), examine the person, score function, ask about weight, eating, infections, and hospital visits, and ask about goals. The hospice medical director then decides whether to certify, and the hospice contacts the attending physician for the second signature. If both agree, admission can happen the same day or the next.

The attending physician's signature is where it stalls. Ask the hospice to make the call rather than making it yourself; hospice medical directors talk to attendings about prognosis all the time and know how to frame it. If your own doctor says they will sign if the hospice thinks it is appropriate, that is a yes.

If the hospice evaluates and says the person does not yet meet criteria, ask two things: which criterion is unmet, and what would change that. Then ask for a date to re-evaluate.

Take the reluctance seriously, and take the evidence about it seriously too. In the study of 343 doctors and 468 patients referred to hospice (Christakis and Lamont, BMJ, 2000), the patients lived a median of 24 days after referral. Doctors had predicted far longer. Only 20 percent of estimates were within a third of the real survival; 63 percent were too optimistic, by a factor of 5.3 on average. This is not a criticism of any one doctor. It is the pattern for the profession, and it means that "not yet" from a treating physician is weaker evidence than it feels like.

There is a related tool clinicians use called the surprise question: "Would I be surprised if this patient died in the next 12 months?" A 2017 review of 16 studies and 11,621 patients (Downar and colleagues, CMAJ) found it had a sensitivity of 67 percent and specificity of 80 percent for death within 6 to 18 months. Modest, but it is a question you can ask the doctor directly, and the honest answer often reframes the conversation.

Three moves, in order. First, request the hospice evaluation anyway; the hospice doctor is the other half of the certification, and their judgment is independent. Second, ask about palliative care. Palliative care has no prognosis requirement at all, runs alongside curative treatment, and a palliative care team can make the referral when the time comes. Third, if the attending is the obstacle, remember that the attending can be the primary care doctor, a geriatrician, or a nurse practitioner who sees the decline. A second opinion on prognosis is reasonable and common.

And if the person is not eligible yet, the evaluation was not wasted. The hospice now has the records, the attending has been asked the question once, and the family has a list of what to watch. Ask for a re-evaluation date; 30 to 60 days out is a reasonable interval to request.

The benefit runs in periods: "an initial 90-day period," "a subsequent 90-day period," then "an unlimited number of subsequent 60-day periods" (42 CFR 418.21). Each period starts with a fresh written certification that the six-month forecast still holds (418.22(a)). The person does not re-apply. The hospice handles it, and in most cases the family never notices.

From the third period on, the rule adds one step. A hospice physician or hospice nurse practitioner "must have a face-to-face encounter" with the person, "prior to, but no more than 30 calendar days prior to, the 3rd benefit period recertification, and every benefit period recertification thereafter." In other words, once someone passes roughly six months on hospice, a hospice clinician has to lay eyes on them every 60 days to confirm continued eligibility. This is the mechanism Medicare uses to check long stays, and it is why a hospice nurse practitioner you have not met may show up around day 180.

Medicare's advisory commission reported that in 2022, 17.3 percent of hospice discharges were live discharges (MedPAC, March 2024 report to Congress). That is one person in six. It happens for three reasons: the person revokes the election, usually to pursue a treatment; the person moves or changes hospices; or the hospice "determines that the patient is no longer terminally ill" (42 CFR 418.26). The last is the one families find bewildering. They were told six months, the person got better, and now the hospice is leaving.

It is not a verdict. The regulation says a discharged person "may at any time elect to receive hospice care if he or she is again eligible," and re-enrollment is allowed. People with dementia are among the most likely to be discharged alive, because decline in dementia is slow and uneven and the criteria, as noted in step 3, are poor predictors. If a live discharge is coming, ask the hospice for a written summary of the current status and what would trigger re-evaluation, ask them to keep the person on their call-back list, and get the hospice records into the same folder as everything else. The next evaluation starts from that paperwork.

The rate varies a lot by hospice. In the national analysis that first drew attention to this (Teno and colleagues, Journal of Palliative Medicine, 2014), not-for-profit hospices discharged 14.6 percent of patients alive and for-profit hospices 22.4 percent, and the range between individual hospices was far wider. About 77 percent of hospices are now for-profit. If you are choosing between hospices, the live-discharge rate is a fair question to ask, and our questions to ask a hospice checklist includes it.

Half of people who enroll are on hospice for 18 days or fewer. A quarter are enrolled for five days or fewer, and one in ten for two days or fewer (MedPAC, 2022 data). The average is 95 days, pulled up by a small number of long stays. Half of families, in other words, get hospice in the last two and a half weeks, after the hard months are already behind them.

Three findings argue against waiting. Hospice was not associated with shorter survival: across 4,493 Medicare patients with six common terminal diagnoses, those on hospice lived on average 29 days longer than matched patients who were not, with the difference significant for heart failure, lung cancer, and pancreatic cancer (Connor and colleagues, Journal of Pain and Symptom Management, 2007).

Families rate the care higher: 70.7 percent of families whose person died with hospice rated the care excellent, against fewer than half for deaths in hospitals, nursing homes, or with home health alone (Teno and colleagues, JAMA, 2004). And earlier conversations lead to earlier enrollment: in a study of advanced cancer patients, those who had an end-of-life discussion with their doctor enrolled in hospice at 65.6 percent versus 44.5 percent, and their bereaved caregivers were less likely to develop major depression (Wright and colleagues, JAMA, 2008).

Only 11.4 percent of bereaved families in a national survey said the referral came too late (Teno and colleagues, 2007). But those who felt it did reported more unmet needs and lower satisfaction.

Hospice is paid by Original Medicare even for people in a Medicare Advantage plan. The Medicare booklet is direct: "Once you start getting hospice care, Original Medicare will cover everything you need related to your terminal illness, even if you choose to stay in a Medicare Advantage Plan." The plan "must help you locate a Medicare-approved hospice provider." Costs are the same: nothing for hospice care itself, up to $5 per prescription for symptom drugs, and 5 percent of the cost of short inpatient respite stays.

Medicaid can cover hospice, but it is an optional state benefit, so check your state. The Department of Veterans Affairs says "all enrolled Veterans are eligible" for hospice if they meet the clinical need, with no copays whether the VA or a contracted hospice provides it. For someone with no coverage at all, ask the hospice directly; ask about charity care at the evaluation visit, which is the right place to raise it.