Heart failure & COPD prognostic criteria — and why organ-failure trajectories are so much harder to call than cancer's steady decline
Congestive heart failure is one of the two most common non-cancer terminal diagnoses in hospice, alongside dementia. The Local Coverage Determination (LCD) framework used by Medicare Administrative Contractors — descended from the original 1996 NHPCO Medical Guidelines for Determining Prognosis — anchors heart failure eligibility on functional class at rest despite maximal guideline-directed medical therapy (GDMT), not on ejection fraction alone.
The New York Heart Association (NYHA) functional classification remains the backbone of heart-failure hospice certification. Class IV is defined as inability to carry on any physical activity without discomfort, with symptoms of heart failure — dyspnea, fatigue, or angina — present even at rest. Any physical activity increases discomfort.
Critically, the LCD requires that this Class IV status persist despite optimal, maximally tolerated medical management: diuretics titrated to euvolemia, ACE inhibitors or ARBs (or an ARNI such as sacubitril/valsartan where tolerated), and beta-blockers, unless the patient is intolerant or already on maximal doses. A patient whose NYHA IV symptoms stem from under-treatment — say, a diuretic dose that has never been optimized — has not met the terminal-trajectory bar; the point of the criterion is to identify patients who have exhausted disease-modifying options, not patients who simply need better outpatient management.
In practice this means the certifying hospice physician must review the medication list, recent echocardiogram, and hospitalization history together. A patient on a stable, maximized four-pillar regimen (diuretic, RAAS inhibitor/ARNI, beta-blocker, and often an SGLT2 inhibitor) who is still breathless getting out of a chair is a very different eligibility case than a treatment-naive patient with the same NYHA class.
Because NYHA class is somewhat subjective, LCDs list a set of supporting factors that bolster (without being individually mandatory for) a heart-failure hospice determination:
• Ejection fraction ≤20% — a severely reduced EF correlates with, but does not define, terminal status; some HFpEF (preserved-EF) patients are just as sick and just as eligible • Treated but persistently symptomatic ventricular or supraventricular arrhythmias • History of cardiac arrest or resuscitation • History of unexplained syncope • Brain embolism of cardiac origin (cardioembolic stroke) • Concomitant HIV disease
Additional corroborating evidence commonly cited in documentation includes recurrent heart-failure hospitalizations (two or more in the prior six to twelve months despite therapy), worsening renal function from low cardiac output (the cardiorenal syndrome), progressive cardiac cachexia and unintentional weight loss, hyponatremia, and an inability to walk more than a few steps or perform basic self-care without dyspnea.
None of these supporting items is individually sufficient or individually necessary — they exist to help the certifying physician build a clinical picture consistent with a six-month-or-less prognosis "if the disease runs its normal course," which is the statutory Medicare hospice benefit standard.
A common misconception — among patients, families, and even some clinicians — is that hospice enrollment means stopping all heart failure medications. In practice, most disease-directed therapies that primarily serve symptom control continue: loop diuretics (including IV diuretics in many "open access" hospice programs), low-dose beta-blockade if tolerated, and supplemental oxygen for comfort. What typically stops are therapies aimed at prolonging life or reversing the underlying disease process rather than controlling symptoms — defibrillator shock therapy is deactivated (the device usually stays in place for pacing), and escalations like inotrope infusions or advanced device therapy are generally not pursued outside of specific concurrent-care demonstration programs.
This comfort-preserving continuity matters clinically: abruptly stopping diuretics in a decompensating heart-failure patient can precipitate pulmonary edema and a distressing death, exactly the outcome hospice aims to prevent.
Chronic obstructive pulmonary disease is the other major "organ failure" non-cancer hospice diagnosis. Like heart failure, COPD eligibility is anchored on a functional, symptom-based threshold — disabling dyspnea at rest, refractory to bronchodilators — with objective pulmonary function and gas-exchange data serving a supporting, not gatekeeping, role.
The LCD standard for end-stage pulmonary disease centers on disabling dyspnea at rest, with poor or absent response to bronchodilators, functionally reducing the patient to a bed-to-chair existence with fatigue and cough. As with heart failure, this is a functional threshold, not a spirometry threshold — a patient can qualify with a moderately reduced FEV1 if the clinical picture (recurrent exacerbations, hypoxemia, weight loss, functional collapse) is severe enough, and conversely a very low FEV1 alone, in an otherwise stable and active patient, does not automatically confer eligibility.
This is intentional: post-bronchodilator FEV1 is a poor independent predictor of six-month mortality in COPD — patients with FEV1 in the teens can survive years, while patients with FEV1 in the 30–40% range can die within months if they have frequent exacerbations, cor pulmonale, and cachexia. The LCD framework therefore treats FEV1 as one supporting data point in a broader clinical gestalt.
Supporting evidence for COPD hospice eligibility includes:
• Post-bronchodilator FEV1 <30% predicted • Increasing visits to the emergency department or hospitalizations for pulmonary infection or respiratory failure • Hypoxemia at rest on room air — pO2 ≤55 mmHg by arterial blood gas, or oxygen saturation ≤88% by pulse oximetry, on supplemental oxygen or room air as clinically documented • Hypercapnia — pCO2 ≥50 mmHg, reflecting ventilatory failure • Cor pulmonale and right heart failure secondary to pulmonary disease (not left heart disease or valvulopathy) • Unintentional weight loss of more than 10% of body weight over the preceding six months, reflecting the metabolic cost of the work of breathing • Resting tachycardia greater than 100 beats per minute
As with the cardiac pathway, none of these is independently required; they cumulatively support a clinical judgment that the patient's trajectory is consistent with a life expectancy of six months or less if the disease follows its expected course. A patient with three of these features and disabling rest dyspnea makes a considerably stronger eligibility case than a patient with an isolated low FEV1 and no other findings.
Just as with cardiac patients, hospice enrollment for COPD does not mean withdrawing respiratory support. Supplemental oxygen, nebulized or inhaled bronchodilators, inhaled or systemic corticosteroids for symptom control, and low-dose opioids for refractory dyspnea (a well-established, guideline-supported palliative use with a strong safety record at appropriately titrated doses) all continue as comfort measures. What generally does not continue is escalatory, hospital-based intervention such as invasive mechanical ventilation or ICU-level non-invasive ventilation trials aimed at reversing the acute process rather than relieving symptoms — though many hospices will support time-limited BiPAP for comfort in a home or facility setting depending on the patient's goals.
The clinical continuity of inhalers and oxygen under hospice is one of the most important facts to communicate to COPD patients and families, since fear of "losing my oxygen" is a frequently cited reason patients decline a hospice referral.
Joanne Lynn's influential "three trajectories of dying" framework (JAMA, 2001) reshaped how palliative medicine thinks about prognosis. Cancer typically shows a relatively steady, high-functioning course followed by a rapid, fairly predictable terminal decline over weeks. Organ failure — heart failure and COPD chief among them — instead declines in an unpredictable staircase of exacerbations, each of which could be the last, or could resolve back toward the prior baseline.
Lynn and colleagues described three archetypal illness trajectories near the end of life. The cancer trajectory shows maintained function and quality of life for a relatively long period, followed by a fairly rapid and predictable decline in the final weeks to months — clinicians and hospice teams have decades of experience recognizing this inflection point, and objective markers (progressive tumor burden, treatment exhaustion, performance-status decline) reinforce it.
The organ-failure trajectory — characteristic of heart failure and COPD — instead shows a jagged, sawtooth pattern: a gradual overall decline in baseline function, punctuated by acute exacerbations that bring the patient close to death, followed by treatment (diuresis, antibiotics, bronchodilators, sometimes ICU care) that returns the patient to a new, slightly lower baseline. Death, when it comes, often arrives during one of these acute exacerbations rather than at a predictable point along a smooth decline curve — sometimes called "sudden death on a background of chronic disease."
The third trajectory, dwindling (typical of frailty and dementia), shows a slow, prolonged decline in function over years without discrete crises, which creates its own distinct prognostic challenge covered under the general-debility pathway.
The clinical implication of the organ-failure trajectory is stark: at any given hospitalization for a heart-failure or COPD exacerbation, a clinician cannot reliably say whether this episode will resolve back to baseline or will be the terminal event. Statistically, roughly half of heart-failure and COPD patients who die do so during an acute exacerbation rather than a recognizable, gradual terminal decline — which is precisely why fixed six-month prognostication is so much harder to apply to these diseases than to cancer.
The landmark SUPPORT study (Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments, JAMA 1995) followed roughly 9,000 seriously ill hospitalized patients and found that physicians' survival estimates were frequently inaccurate and that aggressive, often burdensome interventions continued even very close to death — findings that helped catalyze the modern hospice and palliative care movement.
Christakis and Lamont (BMJ, 2000) formally quantified physician prognostic error in terminally ill patients referred for hospice: physicians overestimated survival by a factor of roughly 5.3 on average, and accuracy was worse — not better — for physicians who knew their patients longer, suggesting an emotional or relational bias toward optimism that compounds the biological unpredictability of organ failure.
The downstream consequence is measurable in hospice utilization data: NHPCO Facts and Figures reports have repeatedly shown non-cancer diagnoses — cardiac and circulatory disease, dementia, respiratory disease, and others combined — now account for roughly 70% of hospice admissions, reflecting real progress in access. But median length of stay in hospice nationally remains strikingly short (on the order of two to three weeks), and non-cancer patients in particular are disproportionately represented among very short stays of a week or less — evidence that referral, even when it eventually happens, is often coming very late relative to the six-month benefit window Medicare intends.
Because point prognostication is so unreliable in organ failure, palliative and hospice clinicians increasingly rely on heuristics rather than a single predictive calculation. The "surprise question" — "Would I be surprised if this patient died in the next six to twelve months?" — has been validated across cardiology and pulmonology populations as a reasonably effective screening trigger for a serious-illness or hospice conversation, precisely because it recruits a clinician's pattern recognition across the whole trajectory rather than demanding a false precision the underlying disease biology cannot support.
Other practical triggers used in cardiology and pulmonology practices include: a second heart-failure or COPD hospitalization within twelve months, initiation of home oxygen for COPD, referral for a left ventricular assist device or transplant evaluation being declined or deferred due to frailty, a new diagnosis of cor pulmonale, or a clinician's own sense that the patient is "declining." Embedding these triggers into electronic health record prompts and multidisciplinary rounds has been shown in quality-improvement studies to shorten the gap between eligibility and referral — directly counteracting the prognostic-uncertainty barrier described in Stage 5.
Many seriously ill patients do not present as a clean single-organ case. An 88-year-old with mild heart failure, moderate COPD, stage 4 chronic kidney disease, and progressive frailty may not clearly meet any single disease-specific LCD — yet the cumulative burden of decline across systems is every bit as prognostically significant. The hospice field addresses this with a general-decline, multi-organ documentation pathway built around functional status scales rather than any one diagnosis.
When no single-organ LCD is clearly met, hospice teams document decline using standardized functional scales that integrate the cumulative effect of multiple comorbidities rather than any one disease process:
• The Palliative Performance Scale (PPS) rates ambulation, activity level and evidence of disease, self-care, oral intake, and level of consciousness on a 0–100% scale in 10% increments; a PPS at or below 40–50%, especially if trending downward over recent assessments, is widely used as a marker consistent with a six-month-or-less prognosis in the general-decline context • The Karnofsky Performance Status (KPS) and ECOG performance status serve a similar function, more commonly in oncology but also applied to general debility • The FAST (Functional Assessment Staging Tool) scale, originally developed for dementia, is frequently co-documented in general-decline patients who also have significant cognitive impairment
What all of these scales share is that they measure the patient's functional trajectory directly, sidestepping the need to attribute decline to any single organ system — which is exactly the point when three or four systems are failing in parallel and none alone would meet its LCD.
For many years, "debility, unspecified" and "adult failure to thrive" were commonly used as the primary hospice diagnosis for exactly this kind of multi-system decline. In 2014, CMS issued guidance effectively barring these non-specific codes as a stand-alone principal hospice diagnosis, on the rationale that a Medicare claim should reflect the actual underlying disease(s) driving the terminal prognosis rather than a vague catch-all label — a change partly motivated by program-integrity concerns about diagnosis specificity and appropriate billing.
In practice, this means the certifying hospice physician must now identify the dominant terminal diagnosis (for example, "combined cardiac and renal disease" or the most clinically significant single condition) and list contributing comorbidities separately, while the clinical narrative and supporting documentation still describe the multi-system, cumulative nature of the decline. The underlying clinical picture of general debility has not changed — the coding and documentation requirements around it have become more specific.
A well-supported general-decline hospice certification typically documents several converging lines of evidence over a period of weeks to months, rather than a single snapshot:
• A declining trend in PPS or KPS across at least two to three sequential assessments, not just a single low reading • Progressive dependence in activities of daily living — transfers, toileting, feeding — often tracked via nursing or home-health documentation • Unintentional weight loss or declining oral intake, sometimes with dysphagia • Recurrent infections (urinary tract infections, aspiration pneumonia, skin breakdown/pressure injuries) reflecting failing physiologic reserve across systems • Increasing frequency of emergency department visits or hospitalizations for any cause, not necessarily a single disease • Caregiver reports of accelerating decline, which — while subjective — are given real evidentiary weight because caregivers often detect trajectory changes before formal testing does
Because no single hard threshold governs this pathway, documentation quality and longitudinal trend matter more here than in the disease-specific LCDs — a single PPS of 40% is far less persuasive than a clear downward trend from 70% to 50% to 40% over three months, alongside two hospitalizations and 12% unintentional weight loss.
Even where a patient clearly meets heart-failure, COPD, or general-decline criteria, real-world referral to hospice lags substantially behind eligibility. The gap is driven by clinician discomfort with prognostic uncertainty, patient and family misconceptions about what hospice permits, structural barriers in how cardiology and pulmonology practices are organized, and documented racial and geographic disparities in access.
Cardiologists and pulmonologists are trained, and rewarded, for extending life and reversing disease processes — implanting devices, escalating GDMT, managing exacerbations back to baseline. Raising hospice, a framework built around a six-month prognosis, can feel to many specialists like an admission of failure, or simply an uncomfortable claim they are not confident making given the organ-failure trajectory described in Stage 3.
This discomfort is compounded by a structural disconnect: oncologists have decades of embedded palliative-care co-management culture and fairly predictable disease trajectories to lean on, while cardiology and pulmonology have historically lacked equivalent integration of palliative principles into routine practice, despite growing evidence and society guidelines (American Heart Association, American College of Cardiology, and pulmonary society statements) now explicitly endorsing earlier palliative and hospice conversations in advanced heart failure and COPD.
Patients and families frequently decline or delay hospice out of a belief that enrollment means stopping all disease-directed treatment — losing diuretics, oxygen, or inhalers, or being unable to return to the hospital. As described in Stages 1 and 2, this is largely a misconception: most hospice programs continue symptom-focused cardiac and pulmonary therapies as core comfort care, and many hospices operate "open access" models that will continue IV diuretics, occasional blood transfusions, or time-limited BiPAP when consistent with the patient's goals.
Clearly communicating this continuity — "hospice does not mean stopping your oxygen or your water pills" — is one of the highest-yield interventions available to referring clinicians, and its absence from many initial hospice conversations is a well-documented, correctable barrier.
The Medicare Care Choices Model (MCCM), a CMS demonstration that ran from 2016–2021, tested allowing eligible patients to receive concurrent hospice-like supportive services alongside continued curative-intent treatment for their terminal diagnosis. While MCCM itself has ended, it demonstrated meaningfully higher hospice election rates and lower acute care utilization among enrollees, and it has informed later concurrent-care and palliative-benefit design discussions.
One of the most consistently reproducible findings in hospice utilization research is that simply informing patients and families that most cardiac and pulmonary comfort medications continue under hospice measurably increases acceptance of a hospice referral — making accurate communication, not new eligibility criteria, one of the single most effective levers available to close the non-cancer referral gap.
Multiple studies using Medicare claims data have found that Black and Hispanic patients are less likely than white patients to enroll in hospice overall, and the gap is often more pronounced for non-cancer diagnoses than for cancer, where hospice referral pathways are more standardized and better resourced. Contributing factors identified in the literature include differential trust in the health care system rooted in historical experience, less consistent access to primary and specialty cardiology/pulmonology follow-up that would otherwise trigger a timely referral conversation, and hospice program availability that is uneven across rural and under-resourced urban areas.
Geographic access compounds the problem: rural areas often have fewer hospice agencies, longer travel distances for inpatient hospice levels of care, and fewer cardiology or pulmonology subspecialists positioned to initiate serious-illness conversations early.
Addressing these disparities requires action at multiple levels: embedding standardized eligibility-screening triggers (Stage 3) into routine cardiology and pulmonology workflows so that referral does not depend solely on an individual clinician's comfort or bias; culturally responsive serious-illness communication training; and continued investment in hospice program capacity in underserved rural and urban communities. Because non-cancer disease now represents the majority of the hospice population, closing this equity gap is central — not peripheral — to the mission of the modern hospice benefit.
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| Heart Failure | NYHA IV at rest, optimal GDMT | EF ≤20%, arrhythmia/arrest hx, cardiorenal decline supportive | Continue diuretics, beta-blockade, O2 |
| COPD | Disabling dyspnea, bronchodilator-refractory | FEV1 <30%, hypoxemia/hypercapnia, cor pulmonale supportive | Continue inhalers, O2, opioids for dyspnea |
| Multi-Organ Debility | PPS ≤40–50%, cumulative decline | No single LCD met; documented functional trend required | Whole-person, symptom-focused plan |
| Renal Failure (non-dialysis) | CrCl <10–15 mL/min, uremic symptoms | Declining function, electing to forgo/stop dialysis | Continue symptom-directed renal care |