We Need a Paradigm Shift Around End-of-Life Decision Making

Although empirical evidence has been lacking, there has been a long-held assumption that unbefriended older adults with professional guardians are at risk of burdensome high-intensity treatments near the end of life. This is because professional guardians, who are paid officials charged with making medical decisions on behalf of individuals who no longer have decisional capacity, have no prior relationship with the individual. Experts have generally believed that professional guardians, in the absence of an advance directive to guide decision making or prior knowledge to invoke substituted judgment, tend to choose high-intensity as opposed to palliative treatments for persons with dementia near the end of life.

In this issue of the Journal of the American Geriatrics Society, Cohen et al. found that veterans who were nursing home residents aged 65 and older with moderate to severe dementia and who had a professional guardian were no more likely to receive high-intensity treatments than the same population who died with decision makers who were not professional guardians (e.g., family members, friends, or an otherwise “known” surrogate).¹ The researchers also observed no significant differences or patterns in the receipt of other potentially burdensome end-of-life treatments (mechanical ventilation, cardiopulmonary resuscitation, feeding tube placement, end-of-life hospitalizations, and in-hospital deaths), a finding they attributed to the high prevalence of these treatments in the surrogate group. The study was unable to determine what proportion of individuals in either group had an advance directive and to what extent surrogates/family members were aware of patients’ wishes regarding treatments near the end of life. Nevertheless, the fact that no observed differences were found in the receipt of high-intensity treatments between individuals with surrogates/family members versus those with professional guardians begs the question as to whether our current standard of substituted judgment is an effective means of ensuring goal-concordant care.

In the United States, primacy is placed on patient autonomy. For patients lacking decisional capacity, the mechanism by which autonomy is best preserved is through advance care planning. The living will, created in the 1960s, made the rather naive assumption that patients could assure their autonomy by describing the treatments that they would want to receive in a variety of hypothetical situations. The recognition of the inability of living wills to account for the myriad of treatment decisions and scenarios that patients actually face near the end of life led to the current practice of transferring decision-making authority to a trusted surrogate or designated healthcare proxy. Surrogates are instructed to make decisions based on their knowledge of their loved one and on what they believe their loved one would want if they could speak for themselves (the substituted judgment standard).

At a fundamental level, substituted judgment first assumes that patients can somehow imagine the multiple end-of-life situations that they might face, that they can envision what they would want in each of these situations, that they can balance the benefits and burdens of treatments as their clinical condition changes, and that they can weigh the impact of their decision on the health and well-being of their loved ones. Substituted judgment further assumes that loved ones can extrapolate this information from an advance care planning conversation or more typically from distant and tangential conversations, piece together what that person would have wanted, and disentangle their own preferences, emotions, and feelings of guilt from that of the individual for whom they are making decisions. It is perhaps not surprising that having an advance directive does not guarantee goal concordance. Studies have shown that surrogates and clinicians frequently choose treatments that are not consistent with the individuals’ advance directive,^2,3 that care preferences are not durable,⁴ and that the way that choices are framed significantly influences the choices surrogates make.⁵

The finding by Cohen et al that decisions made by known surrogates are no different than those of professional guardians is yet one more reason to consider abandoning the intense focus on advance care planning in the United States and consider alternative decision-making paradigms. Whereas a few studies have demonstrated that advance care planning is associated with some reductions in high-intensity treatments, the overwhelming majority show minimal to no effect including two recent large randomized controlled trials.^6–8 Aside from the dollars spent on research to demonstrate a marginal effect (>$300 million in National Institutes of Health funds over 30 years, an average of $1 million a year) and the cost to implement advance care planning programs that to date lead to negligible outcomes, we worry that the reverence to this flawed enactment of autonomy draws attention away from other mechanisms that facilitate decision making. As one of us (R.S.M.) suggested in recent editorials, the focus on advance care planning and advance directives would be fine if it were not for a finite pot of funding and human capital available to improve end-of-life care.^9,10

Other nations provide examples of alternatives. A best interest standard of decision making for those without decisional capacity is the legal norm in countries such as the United Kingdom, where there is a process of determining what treatments and care goals are in a patient’s best interest.¹¹ If the United States was to adopt a best interest standard, it could be one where a patient’s goals and values, and cultural, religious, and personal preferences could be accounted for, but done in a way where it is evaluated along with what is medically in the patient’s best interest. Crisis communication, which supports real time in-the-moment decision making as a true active partnership between clinical teams and surrogates/family members, is another alternative model deserving of empirical research. In essence, we need to consider the ethical principles of beneficence (acting in a patient’s best interest), nonmaleficence (first do no harm), and justice as equals, and not inferiors to autonomy, as an ethical guide for decision making in those without decisional capacity to better achieve realistic goal-concordant care.

There is also a need to recognize that medical decision making does not occur in a vacuum, and broader societal and institutional cultures and the systemic defaults within which we all operate have strong influences on the care that patients receive. Nationally, our market-driven healthcare system financially incentivizes high-intensity care regardless of whether it benefits or harms patients. The systemic drive toward subspecialization, hospital-based care, and the use of life-sustaining technologies are all a product of this neo-liberal system that prioritizes a market’s best interest over the patient’s best interest.

Institutional and regional cultures and clinical practice norms are a powerful driving force that overwhelms individual efforts to affect change against a default of aggressive, high-intensity end-of-life care. It is important to note that in the study, the highest percentage of patients in each group lived in the highest spending hospital referral region (HRR) quintile. This means that a significant proportion of the individuals captured in this study lived in regions where high-intensity care was the norm. In all regions and especially in these high-intensity HRR regions, systemic defaults and norms are perpetuated by established systems, practice patterns, and individual attitudes that tend toward higher intensity care regardless of individual circumstances and patient preferences. Studies have shown significant institutional, regional, and international variability in intensity of care at the end of life that are not fully explained by patient preferences, highlighting the likelihood that hospital culture and individual practice variations contribute strongly to burdensome treatments near the end of life.^12–16 Qualitative studies by Barnato et al. and Dzeng et al. for example, among others, provide insight into the factors that drive variation by highlighting the importance of institutional norms in influencing clinician heuristics, ethical perspectives, and clinical practice patterns.^17–20

To truly mitigate goal-discordant, high-intensity care near the end of life, researchers and practitioners must consider a paradigm shift toward a wider view of what must be studied and intervened upon. We should stop forgiving the flaws of advance care planning, advance directives, and substituted judgment, and consider other strategies to improve decision making in those without decisional capacity. The palliative care and geriatrics communities need to collaborate with and adapt methodologies and knowledge from other fields that are experienced in understanding networked behaviors and social interactions such as sociology, organizational behavior, and systems design. Unless broader systemic change occurs, we will only see incremental changes that fall short of any significant progress against the systemic defaults of aggressive, high-intensity care.