Medical Decision Making, 705
Risk Assessment and Expectations for the Future, 706
Categories of Major Treatment Decisions, 707
Low-Intensity Interventions That May Improve Quantity and Quality of Life, 707
High-Intensity Interventions That May Improve Quantity and Quality of Life, 709
High-Intensity Interventions That May Improve Quantity But Not Quality of Life, 710
Temporary Therapies with Potential Dependence, 710
Noncardiac Procedures in Patients with Advanced Heart Failure, 711
Palliative Care, 711
Types of Palliative Care, 711
Primary Palliative Care, 711
Secondary (Specialty) Palliative Care, 711
Indications for a Palliative Care Consultation, 711
Symptom Palliation, 713
End-of-Life Care and Device Deactivation, 714
Decision-Making Approaches and Communication Skills, 714
Timing of Discussions, 714
Optimal Communication Techniques, 714
Overcoming Barriers to Optimal Communication, 716
Decision Support to Assist With Particularly Difficult Conversations, 716
Unmet Needs and Directions for the Future, 716
Existing therapies slow, but rarely reverse, heart failure disease progression. As a result, the prevalence of symptomatic heart failure has increased, as has the length of time that people spend in later stages of the disease. At the far end of the heart failure spectrum are a group of patients with advanced (Stage D) heart failure for whom symptoms limit daily life despite the usual recommended therapies, and lasting remission into less symptomatic disease is unlikely. Once hospitalized, regardless of etiology or left ventricular ejection fraction (LVEF), outcomes are poor, with frequent readmissions and median survival worse than for most cancers. This is further complicated by multimorbidity (the average patient has more than four other diagnoses), frailty (the median age of a patient hospitalized with heart failure is now over 77 years), and psychosocial limitations (costs of care are not only increasing, but increasingly being shifted to patients). This increasing prevalence and high symptom burden of advanced heart failure have broadened the need for palliative care services. Simultaneously, the possibility of high-intensity therapies with complex trade-offs mandate a systematic and thoughtful approach to medical decision making. This mandate is increasingly taking the form of requirements for shared decision making and palliative care involvement as part of payment policies for common heart failure devices: implantable cardioverter-defibrillators (ICDs), left ventricular assist devices (LVAD), and left atrial appendage occlusion.
This chapter highlights how decision making and palliative care play a central role in the care of patients with advanced heart failure. Specifically, the chapter aims to describe theoretical foundations of medical decision making, to summarize issues around risk assessment, to outline a framework for major medical decisions faced by advanced heart failure patients and their clinicians, and to detail how palliative measures and communication can be better integrated into the care of these complex patients. The goals are to both emphasize the inclusion of formal palliative care services into the management of patients with symptomatic heart failure, and to help health care providers of all types incorporate these concepts into routine practice. The chapter draws from the prior work of researchers, clinicians, and policy experts in the fields of heart failure, palliative care, and medical decision making, with particular benefit from the Writing Group that crafted the American Heart Association’s Scientific Statement on “Decision Making in Advanced Heart Failure.”
Medical Decision Making
Health care decisions nearly always involve uncertainty and are made within the context of incomplete knowledge of the future. Generally, we compare reasonable options to make treatment decisions. However, how potential benefits, risks, burdens, and costs of various options are weighed depends on the perspective of the decision maker. With growing emphasis on evidence-based medicine, shared decision making, and patient-centered care, improving processes for making complex decisions with difficult tradeoffs has garnered increasing attention, particularly in disease states like advanced heart failure.
Health care providers have an ethical and legal mandate to involve patients in medical decisions. Judicial decisions (e.g., Cruzan v Missouri Department of Health ) and legislative actions (e.g., the Patient Self-Determination Act) have repeatedly affirmed the rights of patients, or duly-appointed surrogates, to choose their medical therapy from among reasonable options. The process of informed consent prior to procedural interventions is an embodiment of the ethical principal of autonomy in that it underscores the clinician’s obligation to ensure that the patient is aware of the diagnosis and prognosis, the nature of the proposed intervention, the risks and benefits of that intervention, and all reasonable alternatives and their associated risks and benefits. A gold standard for informed consent would entail a high-functioning health care system that is able to provide the resources with which an activated, informed patient can engage in productive discussions with a proactive, prepared health care team.
Shared decision making builds upon the principles that guide informed consent. It asks that clinicians and patients share information with each other and work toward patient-centered decisions about treatment. Shared decision making incorporates the perspective of the patient, who is responsible for articulating goals, values, and preferences as they relate to his or her health care. Shared decision making also incorporates the perspective of the clinician, who is responsible for narrowing the diagnostic and treatment options to those that are medically reasonable and then communicating expected outcomes. Shared decision making aims to uphold the ideal that patients’ values, goals, and preferences should guide the medical decision-making process. It should be assumed that discussions and decision making with patients also include, when appropriate, the family and other individuals involved, such as caregivers and companions. Thus, shared decision making puts into practice the principle of “patient-centered care,” which the United States Institute of Medicine has identified as one of the six pillars of quality.
Shared decision making is most naturally applied to preference-sensitive decisions, where both clinicians and patients generally agree that equipoise exists, and decision support helps patients think through, forecast, and deliberate their options; however, even seemingly one-sided medical decisions, such as tobacco cessation and medication adherence, remain dominated by patient choice. In situations where clinicians hold the view that scientific evidence for benefit strongly outweighs harm, behavioral support designed to describe, justify, and recommend may be appropriate. Separately, certain therapeutic options may be considered unreasonable (transplantation or permanent mechanical circulatory support above a certain age or comorbidity burden) and therefore independent of patient demands. Situations of medical futility can be difficult to define but have standardized definitions and legal precedent.
Finally, it should also be noted that health policy makers and societal considerations also play a role in medical decision making. Rules and regulations often help promote distributive justice and optimal resource allocation. Patient-clinician discussions regarding treatment options occur within the context of these societal rules and regulations. With the costs of care rising, these policy discussions will almost certainly become more common. While clinicians should play a crucial role in these broader policy decisions, clinicians should simultaneously avoid allowing their own interpretation of societal considerations around cost in the absence of policy to dictate individual decisions at the bedside.
Risk Assessment and Expectations for the Future
Medical therapies tend to derive their indications and risk-benefit ratios from the nature and severity of the disease for which they are being considered. Therefore, assessment of prognosis provides the context for selecting among therapies for life-threatening disease. Frequent reappraisal of a patient’s clinical trajectory helps to calibrate future expectations, to guide communication, and to inform rational decisions. Recognizing when patients have entered into a state of advanced heart failure—American Heart Association Stage D “refractory end-stage heart failure” ( Table 50.1 ) —is important for considering and applying advanced therapies in an appropriate and timely manner. The INTERMACS profiles were designed to help clinicians quickly map disease severity with timely treatment approaches.
Unfortunately, estimating prognosis is particularly challenging for heart failure. The clinical course varies dramatically across the spectrum of disease severity and is relatively unpredictable for individual patients ( Fig. 50.1 ). This contrasts with the more linear decline of patients with advanced cancer, which has traditionally been the model for decision making and palliative care in end-stage disease. Even late in heart failure, patients often enjoy “good days” and brief interludes of apparent stability, which can lull them and their care providers into postponing vital decisions. Prognosis is further clouded by the unique contrast between unexpected sudden death (i.e., lethal arrhythmia) and lingering death with congestive symptoms (i.e., progressive pump failure).
Hundreds of factors have been shown to predict mortality and hospitalization in heart failure, and multivariable risk models have been developed to integrate risk factors into useful prognostic tools for both ambulatory and hospitalized settings ( Table 50.2 ). Such models can objectively ground prognostic estimates, which is important as physicians, nurses, and patients tend to significantly over-estimate survival. However, such models rarely identify individual patients who will die in the next year and, in the advanced heart failure population models, often underestimate absolute risk. As a practical alternative for routine prognosis in day-to-day care, a number of clinical events are known to herald worsening disease ( Table 50.3 ), which should prompt reassessment of patient goals of care as well as consideration of therapy options for advanced heart failure
|Heart Failure Survival Score||Peak VO 2 , LVEF, serum sodium, mean BP, HR, ischemic etiology, QRS duration/morphology||All-cause mortality|
|Seattle Heart Failure Model (depts.washington.edu/shfm)||NYHA class, ischemic etiology, diuretic dose, LVEF, SBP, sodium, hemoglobin, percent lymphocytes, uric acid, and cholesterol||All-cause mortality, urgent transplantation, or LVAD implantation|
|MAGGIC Heart Failure Risk Score ( www.heartfailurerisk.org/ )||Age, LVEF, sBP, BMI, creatinine, NYHA class, gender, smoking, diabetes, COPD, HF diagnosis >18 months prior, BB, ACEI||All-cause mortality|
|ADHERE Risk Model||BUN, SBP, serum creatinine||In-hospital mortality|
|ESCAPE Discharge Score||BNP, cardiopulmonary resuscitation or mechanical ventilation during hospitalization, BUN, sodium, age >70 years, daily loop diuretic dose, lack of beta-blocker, 6-minute walk distance||6-month mortality|
|EVEREST Risk Model||Age, diabetes, h/o stroke, h/o arrhythmia, beta-blocker use, BUN, sodium, BNP, KCCQ scores||Mortality or persistently poor quality of life (KCCQ <45) over the 6 months after discharge|
|I||Inotropes: Previous or ongoing requirement for dobutamine, milrinone, dopamine, etc.|
|N||NYHA Class/Natriuretic Peptides: Persisting NYHA IIIB or IV and/or persistently high BNP or NT-proBNP.|
|E||End-Organ Dysfunction: Worsening renal or liver dysfunction in the setting of heart failure.|
|E||Ejection Fraction: LVEF <25%.|
|D||Defibrillator Shocks: Recurrent appropriate defibrillator shocks for VT/VF.|
|H||Hospitalizations: >1 Hospitalizations for worsening heart failure in last 12 months.|
|E||Edema/Escalating Diuretics: Persistent fluid overload, increasing diuretics, or need for ultrafiltration.|
|L||Low Blood Pressure: Consistently low BP with systolic <90–100 mm Hg and symptoms.|
|P||Prognostic Medication: Inability to up-titrate (or need to decrease/cease) ACEI, BB, ARNI, MRA|
a The I-NEED-HELP mnemonic provides a useful guide to remember events that herald worsening disease, which should prompt reassessment of patient goals of care as well as consideration of therapy options for advanced heart failure.
Prognostic considerations tend to focus on the singular outcome of mortality. However, other clinical outcomes also rank high in importance to individual patients ( Fig. 50.2 ). Multiple studies have documented patients’ willingness to sacrifice survival in exchange for symptom relief, a trade-off that varies between patients and even within the same patient over time, is correlated loosely with disease severity but strongly with do not resuscitate (DNR) status. A full discussion of prognosis therefore includes not only the risks of death but also the patient’s goals and values including potential burdens of worsening symptoms, limited functional capacity, loss of independence, reduced social functioning, decreased quality of life, and increased caregiver commitment. Unfortunately, even less is known about risk prediction for these latter outcomes.
Ultimately the stochastic nature of heart failure conveys a high level of prognostic uncertainty for most individual patients. It is therefore vital to acknowledge and incorporate uncertainty in discussions about future care.
Categories of Major Treatment Decisions
In the face of complex treatment options for heart failure, a framework for classifying various medical decisions can help clinicians and their patients better anticipate and manage those decisions most likely to occur as the disease progresses to an advanced stage ( Table 50.4 ). Clinicians should take an active role in defining the set of interventions that are medically reasonable, and from these, clinicians and patients can work together to determine which option is most consistent with patient values, goals, and preferences. The more that clinicians can anticipate medical decisions in advance (see the section “Risk Assessment” earlier), the more they can help patients have adequate time to consider their likely options.
|Types of Options||Examples of Interventions||Examples of Adverse Outcomes that Should Be Anticipated in High-Risk Patients|
|Low-Intensity Interventions That Might Improve Quantity and Quality of Life|
|Medical therapy||Beta-blocker, ACEI/ARB, MRA, diuretics, control of hypertension and other comorbidities|
|High-Intensity Interventions That Might Improve Quantity and Quality of Life|
|Procedures with the potential to improve cardiac function||CABG|
|Percutaneous valve intervention|
|Replacement of cardiac function||Transplantation|
|High-Intensity Interventions That Might Improve Quantity But Not Quality of Life|
|Procedures to reduce the risk of sudden cardiac death||ICD|
|Temporary Therapies to Stabilize Patients That Can Lead to Dependence|
|Adjunctive therapies instituted during acute decompensation with potential chronic dependence||Temporary mechanical circulatory support devices (IABP, percutaneous VAD, ECMO)|
|Renal replacement therapy (dialysis or ultrafiltration)|
|Noncardiac Procedures With Increased Risk and Potentially Decreased Benefit|
|Joint replacement |
|Resection of pulmonary nodule, routine screening colonoscopy|
Low-Intensity Interventions That May Improve Quantity and Quality of Life
While anticipating and addressing new management options that accompany progressive heart failure, clinicians should work to optimize background medical therapy for heart failure and other comorbidities. In order for patients to be defined as advanced Stage D heart failure, patients must have failed optimal medical therapy (see Table 50.1 ). The need to decrease or discontinue neurohormonal antagonists (e.g., beta-blockers; Table 50.3 ) generally signals worsening disease.
High-Intensity Interventions That May Improve Quantity and Quality of Life
A variety of invasive procedures exist that have the potential to improve cardiac function, thereby increasing both survival and quality of life. However, they also have the potential to cause harm and patient burden, particularly among patients with advanced heart failure who are older and have greater multimorbidity than patients studied in randomized trials. Therefore, decisions around such therapies should include detailed informed consent and shared decision making, even if the therapies are generally supported by a strong evidence base. Additionally, careful consideration of a range of potential complications should be addressed with patients pre-procedurally (“What if?”) so that contingency plans are in place in case rare but serious events do occur.
Patients with heart failure may be considered for cardiac surgery for coronary, valvular, and pericardial disease. Cardiac surgeries involving general anesthesia, thoracic access, and cardiopulmonary bypass are higher risk as a consequence of underlying severe cardiac dysfunction. In order to consider cardiac surgery as an option, the surgery should be expected to convey significant long-term benefit. However, the benefit of many cardiac procedures in advanced chronic heart failure has been difficult to study. The potential for residual cardiac dysfunction, perioperative death, protracted postoperative rehabilitation, and loss of independence must be considered and included thoughtfully in the shared decision, as surgery inherently increases short-term risk for the prospect of longer term benefit.
Less invasive percutaneous approaches for the treatment of coronary and valvular disease may be appealing in advanced heart failure. Catheter approaches to aortic, mitral, and even tricuspid disease have now been shown to be reasonable alternatives to surgery in certain populations ( see also Chapter 26 ). However, the benefits of valve repair or replacement are less well established in patients with significant heart failure, especially when treating functional (secondary) mitral regurgitation for patients with a dilated left ventricle. Additionally, potential benefits may be diminished by noncardiac comorbidities, while risks of contrast-induced nephropathy, stroke, and 30-day mortality are increased in the advanced heart failure population. Patient-clinician discussions regarding percutaneous interventions should also include preprocedural consideration of whether emergency surgery would be appropriate and feasible.
Cardiac implantable electronic devices for bradycardia and cardiac resynchronization therapy (CRT) can improve cardiac function but require an invasive procedure, include the risk of infection, and necessitate long-term follow-up ( see also Chapter 38 ). Additionally, patients with advanced heart failure (New York Heart Association [NYHA] functional class IV) have represented a small fraction of patients included in randomized trials of CRT, creating uncertainty about the benefit in this population. Regardless, the care team should plan for contingencies, such as consideration of an open thoracotomy for perforation or unsuccessful coronary sinus lead placement. Factors likely to modify the risk–benefit ratio of device implantation, such as noncardiovascular morbidity and acute decompensation, should also be recognized and incorporated into these discussions. Although CRT and ICDs are often packaged together, their purposes are quite different; CRT, like neurohormonal antagonist therapy, is designed to improve cardiac performance and can improve quality of life; ICDs do not. Therefore, the recommendation for CRT combined with ICD should prompt separate discussions around the indications for defibrillation versus cardiac resynchronization, as well as differences in the need for monitoring, the chances for inappropriate shocks and worsening heart failure, the risks for infection and lead malfunction, and the options for deactivation.
Cardiac transplantation and mechanical circulatory support offer the potential to fundamentally change the clinical course of heart failure by exchanging it for surgical therapy and a different set of benefits, risks, and burdens. In the case of transplantation, patients are asked to weigh their current clinical course against a posttransplant estimated mean survival of approximately 12 years as well as the risks of surgery, graft rejection, infection, and the other side effects of immunosuppression. For permanently implanted LVADs in inotrope-dependent patients, 2-year survival is increased from less than 10% without LVAD to greater than 70% with LVAD, and a near doubling in quality of life measures among survivors, but at the risk of (1) reoperation to replace a malfunctioning pump (10%); (2) disabling stroke (11%); (3) device infection; (4) bleeding; and (5) recurrent hospitalizations. Even in the setting of successful LVAD, patients must maintain a constant power source, perform vigilant driveline care, and manage other chronic diseases—and these issues spill over to caregivers. Thus, for eligible patients, whether to pursue transplantation and/or mechanical circulatory support involves complex tradeoffs and major uncertainties. For young patients dying from heart failure without other comorbidities and with adequate socioeconomic support, transplantation and LVAD are rarely declined. For older patients with significant noncardiac disease, declining an LVAD is not unusual. For the majority of patients with heart failure these advanced therapies are not an option due to the predominance of heart failure with normal ejection fraction, multiple comorbidities, or very advanced age. Detailed clinical practice guidelines for transplantation and mechanical circulatory support—including detailed discussions of candidacy—are available, and are covered in other chapters ( see also Chapters 44 and 45 ). Medicare’s 2013 decision memo requires the inclusion of palliative care service for programs to maintain certification for destination therapy implantation.
High-Intensity Interventions That May Improve Quantity But Not Quality of Life
ICDs are fundamentally different than many life-saving therapies for patients with chronic heart failure with reduced ejection fraction. ICDs improve survival by aborting lethal arrhythmias, but do not improve cardiac function or heart failure symptoms. At the same time, ICDs create an additional burden for patients including inappropriate shocks, increased risk of hospitalization, potential decrease in quality of life, and higher likelihood of death from progressive pump failure, meticulous discussion of absolute risks with and without ICDs are particularly important for informed consent and shared decision making. Medicare’s proposed 2018 changes to the national coverage determination for ICDs include “Requiring a patient-shared decision-making interaction prior to ICD implantation for certain patients.”
Temporary Therapies with Potential Dependence
Some therapies are intended for short-term use to stabilize patients, thereby allowing for recovery from potentially reversible insults or transition to more definitive therapy. Although initially intended as a temporary intervention, such stabilizing therapies can create indefinite dependence if the patient does not improve as hoped or develops an adverse event (e.g., stroke, progressive renal failure) that compromises previously anticipated options.
Short-term circulatory support with intra-aortic balloon pumps or percutaneous mechanical circulatory support devices may be initiated when acute hemodynamic instability requires urgent intervention to avoid permanent end-organ dysfunction and/or death. It may be instituted with the hope of supporting a reversible underlying condition, such as fulminate myocarditis or cardiogenic shock after acute myocardial infarction. It may also be initiated in patients who might be potential candidates for transplantation or permanent circulatory support in whom (1) there has not been opportunity to appropriately evaluate a patient’s candidacy or preferences for more definitive high-dependence therapies, (2) reversibility of end-organ dysfunction is uncertain (INTERMACS profile 1), or (3) socioeconomic contraindications to more definitive therapies may resolve in the near future. If device dependency persists and contraindications to definitive therapies do not resolve, a decision will need to be made to discontinue therapy, which can be particularly challenging when patients or families are not in agreement with withdrawal. To whatever degree possible, these issues should be addressed prior to the initiation of short-term support.
Intravenous inotropic agents are commonly initiated in the acute setting for hemodynamic stabilization and to improve end-organ perfusion. Their use is most often anticipated to be temporary, with the hope of either clinical improvement or eligibility for more definitive therapies as above. Regardless of intent, the initiation of inotropic support for exacerbation of chronic heart failure should be considered a harbinger of severe heart failure (see Table 50.3 ). When patients fail to wean from intravenous inotropic support, decisions arise around continued chronic use. Therefore the goals of temporary inotrope use should be clearly established prior to initiation, and unexpected dependence on this therapy should prompt direct discussions about overall goals of care. High-quality evidence about the risks and benefits of ambulatory inotrope infusions in patients with advanced heart failure is limited, particularly when used for palliation (as opposed to bridge to transplantation). Available data suggest that outpatient inotrope infusions are associated with early improvement in NYHA functional class, fewer hospital days, and no major change in survival. In a contemporary single-center experience of 197 patients discharged with continuous inotropes, 17% had ICD shocks, 29% had infections, and 57% were rehospitalized. Thus inotropic infusion may offer a medical option to help end-stage patients leave the hospital and return home, but it must be recognized that their use is relatively expensive and only about half the hospice agencies will take patients on inotropes. The decision to arrange for chronic continuous infusions should be offered only after checking availability and cost considerations, and should be subsequently guided by the patient’s goals and preferences, including the need for symptom relief.
The prevalence of advanced kidney disease increases dramatically with worsening heart failure, and measures of renal dysfunction are strong predictors of adverse outcomes in patients with heart failure (see Tables 50.2 and 50.3 ). Dialysis in the setting of older age has been shown to add to patient burden, and in high-risk patients (e.g., those with heart failure) may not extend life. Not surprisingly, ICDs may be less effective in heart failure patients on dialysis. LVAD use in dialysis patients has a median survival of less than 3 weeks. Therefore, the initiation of renal replacement therapies (e.g., hemodialysis, ultrafiltration) in patients with advanced heart failure or, conversely, the initiation of mechanical circulatory support in patients with severe chronic kidney disease should only be made after carefully assessing medical feasibility followed by a clear discussion with the patient about the risks and benefits of dialysis on the patient’s quality of life and prognosis.
Feeding tubes, placement of permanent peritoneal and pleural catheters for the control of volume status, and intensive care provide additional examples of therapies with questionable long-term value in patients with irreversible underlying disease. When such situations are anticipated, these types of near end-of-life interventions can be considered as part of a healthy discussion about end-of-life care, ideally prior to the occurrence of a near-terminal event.
Noncardiac Procedures in Patients with Advanced Heart Failure
The risks and benefits of interventions for noncardiac conditions may be significantly altered in patients with advanced heart failure. When the likelihood of meaningful recovery without the procedure is small, the increase in procedural risk in patients with advanced heart failure may be considered acceptable. Examples include both emergent (e.g., laparotomy for perforated viscous) and urgent (e.g., hip arthroplasty for fracture) surgical procedures. However, the majority of noncardiac procedures, such as knee replacement for degenerative joint disease, must be carefully considered in the context of patient preferences, as complications of the procedure may or may not outweigh the potential benefit. Major procedures should generally be discouraged when they do not offer a tangible improvement in quality of life (e.g., repair of asymptomatic abdominal aortic aneurysm). Routine screening tests (e.g., mammography, colonoscopy) are generally not appropriate in the context of a significant competing risk of mortality from advanced heart failure, yet such tests are frequently ordered at the end of life.