🏡 Hospice Eligibility Prognosis Criteria Simulator

Local Coverage Determinations — Disease-Specific Terminal Prognosis Criteria

Medicare does not certify hospice eligibility on diagnosis alone. Each Medicare Administrative Contractor (MAC) publishes a Local Coverage Determination (LCD) — most hospices reference "Hospice Determining Terminal Status" (L34538 and its regional equivalents) — that lays out clinical findings, by disease category, expected to accompany a life expectancy of six months or less if the illness follows its normal course. The criteria exist because prognostication is inherently uncertain, and CMS wants a reproducible, auditable clinical basis for the determination rather than a physician's unstructured impression.

• 10+: LCD-covered primary diagnoses (cancer, ALS, dementia, CHF, COPD, renal, liver...)
• FAST ≥7c: Dementia eligibility threshold (loss of ambulation, plus a comorbid condition)
• 2: Cancer eligibility pathways (metastatic at diagnosis OR progression to metastatic)
• ~13%: "Debility unspecified" admissions (largest single non-cancer hospice diagnosis group)

How LCDs structure the terminal-prognosis question

Local Coverage Determinations split hospice eligibility documentation into two tiers. "General guidelines" apply across all diagnoses and describe the overall pattern of decline expected in a dying patient: progressive clinical decline, decreasing Palliative Performance Scale score, increasing dependence in activities of daily living, and specific clinical variables such as weight loss and recurrent infection. "Disease-specific guidelines" then layer on findings unique to the underlying terminal condition — a falling ejection fraction and NYHA Class IV symptoms for heart disease, a FEV1 below 30% predicted with hypoxemia for pulmonary disease, or a FAST stage of 7 or beyond for dementia.

Neither tier is meant to function as a rigid checklist that must be satisfied item-by-item. CMS and the LCDs themselves emphasize that these are guidelines to support clinical judgment, not a mechanical eligibility test — a hospice medical director may certify a patient whose case does not tick every box if the clinical picture, taken as a whole, genuinely supports a six-month-or-less prognosis. Conversely, meeting the numeric thresholds does not guarantee approval if other clinical evidence points toward a longer trajectory. This is the same tension that runs through every field of prognostic medicine: population-level statistical criteria applied to an individual patient.

Cancer — the diagnosis that anchored the hospice benefit

Cancer was the dominant diagnosis when the Medicare Hospice Benefit was created in 1982, and its LCD criteria remain comparatively clean because malignant disease often (though not always) follows a more predictable terminal trajectory than organ failure or dementia. The LCD recognizes two qualifying pathways: (1) the patient has metastatic cancer at the time of initial clinical presentation, historically associated with a poor prognosis, or (2) the cancer has progressed from an earlier stage to a metastatic or widely disseminated stage despite disease-directed treatment, and the patient has declined further curative-intent treatment or continues to decline clinically despite treatment.

Supportive documentation includes a Karnofsky Performance Status or PPS score, typically 70% or below, progressive weight loss, and symptoms such as nausea, dyspnea, or edema that are refractory to management. Certain cancers with especially aggressive natural histories — small cell lung cancer, pancreatic cancer, glioblastoma multiforme, malignant hypercalcemia — are recognized as supporting a terminal determination even without lengthy documentation of progression, because their expected survival curves are steep enough that waiting for "proof" of decline would itself be clinically inappropriate.

Dementia — FAST staging and the trap of a slow, plateauing decline

Dementia is one of the hardest diagnoses to certify accurately because its terminal trajectory is long and gradual rather than sharply declining, and general debility can plateau for months or years before a final decline. The LCD requires the patient be assessed at Functional Assessment Staging Tool (FAST) stage 7 or beyond — meaning: 7a, ability to speak is limited to roughly six or fewer intelligible words during an average day; 7b, vocabulary limited to a single intelligible word; 7c, loss of ambulation, no longer able to walk without assistance; 7d, inability to sit up independently; 7e, inability to smile; 7f, inability to hold the head up.

Because FAST 7a–7b alone are compatible with survival well beyond six months, the LCD additionally requires the patient have had, within the past twelve months, at least one of: aspiration pneumonia, upper urinary tract infection (pyelonephritis), septicemia, decubitus ulcers stage 3–4, recurrent fever after antibiotics, or an inability to maintain sufficient fluid and calorie intake documented by either 10% weight loss in the preceding six months or a serum albumin below 2.5 g/dL. This dual requirement — advanced functional loss plus a marker of medical complication — exists precisely because dementia alone, even at FAST 7c (loss of ambulation), is frequently compatible with survival beyond six months in the absence of these complicating events.

Palliative Performance Scale, Karnofsky Score, and ADL Dependency

Diagnosis alone cannot establish a six-month prognosis — two patients with the same cancer stage can have wildly different trajectories depending on how much physiologic reserve remains. Functional status scales translate a patient's day-to-day capability into a standardized, reproducible number that correlates strongly with survival across almost every terminal illness, making them the connective tissue between diagnosis-specific criteria and the broader clinical picture.

• 0–100%: PPS scale range (11 levels, scored in 10% increments)
• ≤70%: PPS supportive-of-hospice level (used across debility & general decline pathways)
• ≥3 of 6: ADL dependency threshold (bathing, dressing, transfer, continence, feeding, ambulation)
• 0–100: Karnofsky (KPS) parallel scale (the original 1948 scale, still used in cancer LCDs)

The Palliative Performance Scale — five domains, eleven levels

The Palliative Performance Scale (PPS), developed by Victoria Hospice in 1996 as an extension of the Karnofsky Performance Status specifically for palliative populations, rates a patient across five domains simultaneously: ambulation, activity level and evidence of disease, ability to perform self-care, oral intake, and level of consciousness. A trained assessor reads down the standardized PPS table and matches the patient's overall pattern to one of eleven levels, in 10% increments from 100% (fully ambulatory, no evidence of disease, normal intake, full consciousness) down to 0% (death).

PPS 70% and below is typically the level at which "supportive of hospice appropriateness" language begins to apply for diagnoses without a strong disease-specific marker — the patient is reduced to normal activity with effort or unable to do normal work, though usually still largely self-caring. PPS 50% and below — mainly sitting or lying, unable to do most activities, requiring considerable assistance — correlates with markedly shorter median survival across essentially every disease category studied, often measured in weeks to a few months rather than a year. PPS 40% and below (mainly in bed) and PPS 30% or lower (totally bed-bound) are associated with median survivals frequently under 30 days in hospice-referred populations, which is why a PPS drop of two or more levels within a short interval is itself treated as a red flag for accelerating terminal decline warranting reassessment of the care plan, independent of the specific diagnosis.

Karnofsky Performance Status and its persistence in oncology LCDs

The Karnofsky Performance Status (KPS), introduced in 1948 by David Karnofsky and Joseph Burchenal to standardize the reporting of chemotherapy outcomes, remains embedded in cancer-specific hospice LCD language even though PPS has largely superseded it in general palliative practice. KPS uses the same 0–100 scale in 10-point increments, but is scored along a single continuum of general performance and care requirement rather than PPS's five separate domains — for example, KPS 70 reads "cares for self; unable to carry on normal activity or to do active work," while KPS 40 reads "disabled; requires special care and assistance."

Because KPS predates PPS by nearly five decades and was the instrument used in the oncology trials from which much of the original hospice eligibility evidence base was built, many cancer LCDs still cite a KPS or PPS of 70% or less as one of the supportive general criteria for hospice-level decline, alongside progressive disease and refractory symptoms. In practice, hospice interdisciplinary teams today score PPS almost universally because its five-domain structure is more sensitive to the specific pattern of decline (a patient may have normal intake but severely reduced ambulation, or vice versa) and because it was purpose-built for the palliative population rather than adapted from an oncology-trial context.

Activities of daily living — the functional decline that families see first

Independent of formal performance scales, hospice assessment routinely documents dependency across the six basic Activities of Daily Living (ADLs) originally defined by Katz in 1963: bathing, dressing, toileting, transferring (bed to chair), continence, and feeding. NHPCO's non-cancer clinical guidelines note that dependence in three or more of these six ADLs, combined with an appropriate diagnosis and evidence of disease progression, meaningfully strengthens the case for a six-month-or-less prognosis — particularly for the "debility unspecified" and "adult failure to thrive" pathways where no single dominant terminal diagnosis exists to anchor the LCD-specific criteria.

ADL decline is also the domain most visible to family caregivers and often the first concrete trigger for a hospice referral conversation, even before a physician has formally reassessed PPS or reviewed imaging. A patient who could dress and transfer independently three months ago but now requires two-person assistance for both has experienced a functional trajectory that is, in itself, a form of prognostic data — and hospice intake clinicians are trained to elicit this history systematically rather than relying solely on the most recent clinic note, since functional decline at home frequently outpaces what is captured in intermittent outpatient visits.

PPS and KPS are strong population-level predictors of survival, but both were validated on cohorts of patients already identified as seriously ill or hospice-eligible — applying a single PPS cutoff to the general population would produce enormous numbers of false positives. Functional scores are necessary but not sufficient; they are always interpreted alongside diagnosis-specific criteria and trajectory, never in isolation.

Illness Trajectory, Weight Loss, Recurrent Hospitalization, and Comorbidity Burden

A single snapshot of functional status cannot distinguish a patient who has been stable at a reduced baseline for years from one who is actively, rapidly declining toward death. Hospice eligibility determination therefore places heavy weight on trajectory — the shape of decline over time — and on specific clinical events, like weight loss and recurrent hospitalization, that mark the transition from chronic illness into a terminal phase.

• >10%: Weight loss criterion (unintentional, in the preceding 6 months)
• 3: Classic illness trajectory archetypes (cancer / organ failure / frailty & dementia)
• ≥2: Recurrent acute-event flag (hospitalizations or ED visits within 6 months)
• <2.5 g/dL: Serum albumin supportive marker (nutritional decline indicator across LCDs)

Three archetypal trajectories of dying — Lynn and Lunney's framework

Joanne Lynn and June Lunney's widely cited framework, developed from Medicare claims and SUPPORT study data in the late 1990s and early 2000s, describes three broadly distinct trajectories of functional decline toward death, and hospice clinicians still use this framework to interpret where a given patient sits.

Trajectory one — short decline with a clear terminal phase — is classically associated with cancer: relatively high, stable function is maintained for a prolonged period, followed by a comparatively rapid, predictable decline in the final weeks to months. This trajectory is the one the hospice benefit was originally designed around, and it is why cancer prognostication, while still imperfect, tends to be somewhat more reliable than for other diagnoses.

Trajectory two — long-term decline punctuated by episodes of acute, severe exacerbation with partial recovery, and eventually a sudden or unpredictable death — is typical of organ failure: congestive heart failure, COPD, end-stage liver or renal disease. A patient may be hospitalized in florid decompensated heart failure, recover to near-baseline function within days on IV diuresis, then repeat this cycle several times before an eventual admission from which they do not recover. This sawtooth pattern makes six-month prognostication for organ failure notoriously difficult, since any single hospitalization could be "the one," or could be followed by months of relative stability.

Trajectory three — prolonged, gradual dwindling with a low, slowly eroding baseline and no single dramatic terminal event — characterizes dementia and generalized frailty. Because the decline is so gradual, it is easy for clinicians and families alike to underestimate how close a patient actually is to death, which is precisely why the LCDs impose specific complicating-event requirements (recurrent infection, weight loss, pressure injuries) on top of functional staging for these patients.

Weight loss, nutritional decline, and recurrent acute events as trajectory markers

Across nearly every disease-specific LCD, unintentional weight loss exceeding 10% of body weight over the preceding six months (or a comparable decline reflected in serum albumin below roughly 2.5 g/dL, when weight measurements are unavailable or unreliable due to edema or ascites) functions as one of the most consistently cited supportive findings for a terminal trajectory. Weight loss of this magnitude reflects a catabolic state that oral nutritional supplementation and appetite stimulants typically cannot reverse once the underlying disease has progressed to this point, distinguishing it from reversible causes of weight loss such as depression or medication side effects that a clinician should rule out before treating it as prognostic.

Recurrent hospitalizations or emergency department visits — generally, two or more within a six-month window for a condition related to the terminal diagnosis — serve a similar function, marking the point at which a chronic disease has become unstable enough that outpatient management alone can no longer maintain the patient's baseline. Each hospitalization for decompensated heart failure, aspiration pneumonia, or recurrent infection represents both a physiologic insult from which the patient may not fully recover and a data point suggesting the current care setting or treatment intensity is no longer adequately controlling the disease trajectory — which is often the clinical moment at which a hospice referral conversation becomes appropriate, independent of whichever specific LCD criteria are ultimately documented.

Comorbidity burden — why the sum is often worse than any single diagnosis

Most hospice-eligible patients, particularly in the non-cancer population, carry multiple concurrent chronic conditions rather than a single clean terminal diagnosis, and comorbidity burden independently shortens survival beyond what any one disease would predict alone. A patient with moderate COPD and moderate chronic kidney disease and mild heart failure may have a shorter expected survival than a patient with severe COPD alone, because each additional organ system with reduced reserve narrows the physiologic margin available to tolerate any acute stressor — an infection, a fall, a medication interaction.

This is part of why the "debility unspecified" and "adult failure to thrive" LCD pathways exist: for a meaningful fraction of hospice-eligible patients, no single diagnosis dominates the clinical picture enough to anchor a disease-specific LCD, and the terminal trajectory is better explained by the aggregate burden of several moderate conditions compounding each other than by any one severe disease. Tools like the Charlson Comorbidity Index, while more commonly used in research than in individual hospice certification, formalize this intuition by assigning weighted points to each comorbid condition and summing them into a composite score that correlates with mortality risk across diverse populations.

Two-Physician Certification, Documentation Standards, and the Limits of Prognostic Accuracy

However strong the diagnosis-specific, functional, and trajectory evidence, hospice eligibility ultimately requires a formal act of physician judgment: certification, in writing, that the patient's prognosis is six months or less if the illness runs its normal course. This step exists precisely because prognostication is a probabilistic clinical skill, not a deterministic calculation — and the literature on physician prognostic accuracy shows that skill is systematically and predictably biased.

• 2: Certifying physicians required (initial) (hospice medical director + attending physician)
• ~20%: Physician prognostic accuracy (estimates within 33% of actual observed survival)
• ~5.3×: Median overestimation factor (Christakis & Lamont, BMJ 2000, terminally ill cohort)
• 30 days: Face-to-face encounter window (prior to every recertification after the first 180 days)

Who must certify, and when

For the initial certification of terminal illness, Medicare requires two physicians to independently attest, in writing, that they believe the patient's prognosis is six months or less should the illness run its normal course: the hospice medical director (or a hospice physician member of the interdisciplinary group) and, if the patient has one, their attending physician — the physician the patient identifies as having primary responsibility for their care at the time of hospice election. If the patient has no attending physician, the hospice medical director's certification alone suffices.

For every subsequent recertification, only the hospice physician's certification is required — the attending physician's independent sign-off is not mandated again, though many hospices continue to involve the attending physician in recertification discussions as a matter of clinical continuity and shared decision-making. This asymmetry — two physicians at the start, one thereafter — reflects Medicare's judgment that the initial determination carries the highest risk of error and benefits most from an independent, first-time second opinion.

The certification periods and the face-to-face encounter requirement

The Medicare Hospice Benefit is structured into a defined sequence of certification periods: two initial 90-day periods, followed by an unlimited number of subsequent 60-day periods, for as long as the patient continues to meet eligibility criteria and elects to remain on the benefit. Each period requires a fresh certification of terminal prognosis before it begins.

A reform introduced by the Affordable Care Act in 2010 added a face-to-face encounter requirement starting with the third benefit period (that is, beginning around the 180-day mark) and continuing for every recertification thereafter: a hospice physician or nurse practitioner must see the patient in person within the 30 days prior to the start of that recertification period, and must attest in writing that the visit occurred and describe the clinical findings supporting continued eligibility. This requirement was added specifically in response to CMS data showing that patients with unusually long hospice stays were disproportionately associated with diagnoses — dementia and debility chief among them — where prognostic uncertainty is highest, and was intended to add a periodic, in-person clinical check against a slow drift toward certifying patients whose prognosis has, in practice, extended well beyond six months.

What the documentation must actually contain

A certification that simply states "prognosis six months or less" without supporting clinical detail does not meet Medicare's documentation standard and is a common target of post-payment audits and program integrity reviews. Acceptable certification documentation is expected to describe the clinical findings that support the determination in language specific to the patient — the relevant disease-specific LCD criteria met, the current PPS or KPS score and its trend over recent assessments, weight change, recent hospitalizations, and any other trajectory evidence discussed in the prior stages of this determination — rather than boilerplate language copied across patients.

The narrative requirement exists in part because hospice fraud enforcement actions have repeatedly identified cases where hospices admitted or retained patients using generic, unsupported certification language for individuals who were, on independent chart review, clearly not terminally ill by any reasonable clinical standard. Well-documented, patient-specific certifications protect both the patient — by ensuring the clinical determination was genuinely individualized — and the hospice, by creating an auditable record that the eligibility determination reflected real clinical reasoning rather than a rubber stamp.

In a landmark study of 468 terminally ill patients (Christakis & Lamont, BMJ 2000), physicians' survival estimates were accurate — within 33% of actual observed survival — only about 20% of the time, and the dominant error was systematic overestimation, by a median factor of roughly 5.3, with longer doctor-patient relationships associated with even greater optimism bias. This is a large part of why hospice policy leans on structured, disease-specific criteria and functional scoring rather than unstructured physician estimation alone: the clinical intuition that "the patient still has time" is, on average, measurably unreliable in exactly the direction that delays appropriate hospice referral.

Electing the Medicare Hospice Benefit — Certification Periods, Revocation, and Live Discharge

Once diagnosis-specific criteria, functional decline, trajectory, and physician certification align, eligibility determination culminates in a concrete administrative and clinical event: the patient (or their representative) formally elects the Medicare Hospice Benefit, choosing palliative, comfort-focused care over continued Medicare coverage of curative treatment for the terminal illness. That election is never irreversible — patients may revoke it at any time, and a meaningful share of hospice patients are ultimately discharged alive.

• 2×90d → 60d: Certification period structure (two 90-day periods, then unlimited 60-day periods)
• 1982: Medicare Hospice Benefit established (Tax Equity and Fiscal Responsibility Act (TEFRA))
• ~17%: National live discharge rate (patients discharged alive from hospice care)
• Any time: Revocation availability (patient may resume standard Medicare coverage)

The election statement — what the patient agrees to

Electing hospice requires the patient or their authorized representative to sign a formal election statement acknowledging that they understand the nature of hospice care — an interdisciplinary approach centered on comfort and quality of life rather than cure — and that, for the duration of the election, standard Medicare Part A and Part B coverage is waived for treatments related to the terminal illness and related conditions, in favor of the palliative approach delivered under the per-diem hospice benefit. Coverage for conditions unrelated to the terminal diagnosis continues under standard Medicare as before.

The hospice interdisciplinary team — physician, nurse, social worker, chaplain, home health aide, and volunteers — then develops an individualized plan of care addressing physical symptoms, psychosocial needs, and, for many patients and families, spiritual and existential concerns, delivered wherever the patient calls home: a private residence, assisted living facility, nursing home, or dedicated hospice inpatient unit for patients requiring more intensive symptom management than the home setting can provide.

Certification periods, recertification, and the rhythm of ongoing eligibility review

Once elected, the benefit proceeds through the certification-period structure described in the previous stage: an initial 90-day period, a second 90-day period, and then successive 60-day periods without an upper limit, each requiring a fresh certification of continued terminal prognosis before the next period begins. This structure means hospice eligibility is never a one-time determination — a patient who is appropriately certified today must be periodically reassessed, and it is entirely appropriate, and clinically expected, for a meaningful fraction of patients to no longer meet criteria at some later recertification point if their disease trajectory stabilizes or improves.

This periodic reassessment is a feature of the benefit's design, not a flaw: because prognostication carries genuine uncertainty, particularly for non-cancer diagnoses with slow or fluctuating trajectories, the recurring certification cycle is Medicare's built-in mechanism for catching and correcting cases where the original six-month estimate no longer holds, rather than leaving patients enrolled indefinitely on the basis of an initial assessment that may have been reasonable at the time but is no longer accurate.

Revocation, discharge for extended prognosis, and the live discharge rate

A patient may revoke the hospice election at any time, for any reason, restoring standard Medicare coverage immediately — most commonly because they wish to pursue disease-directed treatment they had previously declined, or because their condition has improved enough that palliative-only care no longer feels appropriate to them. A patient who revokes remains eligible to re-elect hospice again later if their condition subsequently declines.

Separately, a hospice may discharge a patient for several defined reasons: the patient moves out of the hospice's service area or transfers to another hospice; the patient or family requests discharge; the hospice determines the patient is no longer terminally ill, meaning their prognosis has, on reassessment, extended beyond six months; or, rarely, discharge for cause in situations involving safety concerns. Nationally, roughly one in six hospice patients — commonly cited around 17% — are discharged alive rather than dying while enrolled, and the largest single reason is an extended prognosis on reassessment, particularly common among dementia and debility patients whose slow, plateauing trajectories make the initial six-month estimate the hardest of any diagnosis category to get right.

CMS and the HHS Office of Inspector General monitor live discharge rates as a program-integrity signal: hospices with unusually high live discharge rates, especially concentrated in long-stay non-cancer patients, have been flagged in oversight reports as a marker potentially indicating overly liberal admission practices rather than genuinely uncertain prognostication, and have been subject to targeted audits under CMS's hospice special focus initiatives. For an individual patient, however, a live discharge is not a failure of the system — it reflects exactly what the recertification structure is designed to catch: a patient whose condition, happily, did not follow the expected terminal trajectory.