Clinical research on Alzheimer’s disease is much needed. The development of effective treatments for Alzheimer’s disease depends on clinical trials involving human subjects, as the disease’s complex pathology cannot be fully replicated in animal or laboratory models. The United States’ National Institute on Aging puts this succinctly: “No major advance in Alzheimer’s and related dementias treatment, prevention, or care will be possible without robust clinical research.” ¹ This research is critical not only for testing the safety and efficacy of new drugs but also for understanding the disease’s progression, identifying biomarkers for earlier and more accurate diagnosis, and determining the most effective windows for therapeutic intervention.Reference Zhang ² For decades, the field was defined by clinical failures, but recent breakthroughs underscore the need for human trials.Reference Kim ³ The recent approvals of amyloid-beta-directed monoclonal antibodies, such as lecanemab and the promising results for donanemab, represent a positive shift, demonstrating that disease modification is possible.Reference van Dyck ⁴ These successes were built on decades of research involving thousands of patient volunteers, whose participation was essential. Hence, it is apparent that the need for clinical trials with Alzheimer’s disease patients is present.

One glaring ethical problem facing the field is that patients with Alzheimer’s who are needed for research studies have, at times, significant cognitive impairment, which hinders the informed consent process. Research advance directives (RADs), a type of advance directive in which a capacitated participant documents their wishes for research participation at a future point when they no longer retain decision-making capacity, have been discussed in the literature for decades.Reference Muthappan, Forster and Wendler ⁵ As Hart notes in his article, “Research Advance Directives: Ethical Implications for Persons with Alzheimer’s Disease, and for the Families of Elderly Dementia Patients,” objections have been raised at various points to the validity and usefulness of RADs based on faulty conceptions of personhood.Reference Hart ⁶

Hart argues, convincingly, that RADs are not only ethically permissible but offer a superior mechanism for knowing a person’s preferences about research participation compared to surrogate decision-making. This claim is strengthened by dismantling what Hart views as the most significant philosophical objection to RADs: that an individual with advanced dementia is a fundamentally “different person” from the one who completed the directive, thereby rendering the RAD invalid. By arguing for a more holistic and enduring model of personhood, he makes a strong case for the expanded use of RADs to advance critical research while respecting the long-term values of patients.

While we agree that RADs should play a role in lowering the barriers to participation, we believe the case for them cannot come at the expense of the roles that capacity assessment and surrogate decision-making play in the decision-making process. An advance directive is invalidated if the principal who issued the directive rescinds it and has the requisite decision-making capacity to do so. Furthermore, no RAD can take effect unless the participant is found to lack decision-making capacity. Hence, capacity assessment, despite its flaws, is essential to the validity of RADs; criticizing capacity assessment as unreliable risks undermining the only clinical tool we have for determining whether someone’s RAD takes effect or is rescinded.

As for surrogate decision-making, Hart rightly notes that it is an inadequate solution to the barriers facing trial participation. While it is a cornerstone in clinical ethics, its application both in clinical and research settings is fraught with challenges. The standard of “substituted judgment,” by which the surrogate is expected to decide as the patient would do so if that patient were able to, is difficult to meet. For research decisions, it requires the surrogate to accurately infer the patient’s wishes from hypothetical deliberation about complex research protocols the patient may never have considered. Faced with this uncertainty and burdened by the emotional weight of their loved one’s decline, surrogates will likely default to a more protective “best interests” standard. This can lead them to decline research participation, prioritizing the patient’s immediate comfort and avoidance of any risk faced in the research, however minimal. While understandable, this protective instinct may inadvertently override the patient’s previously stated values, such as a deep-seated altruism or a desire to contribute to a cure for future generations — the very “legacy” interests a research advance directive is designed to protect. Documenting that interest in a RAD at least offers a direct, unmediated expression of the patient’s own voice, values, and considered choices that a surrogate has a prima facie duty to follow.

This benefit of RADs may be more limited than Hart suggests. While we agree that persons remain numerically identical before and after developing Alzheimer’s disease, their desires and beliefs may nonetheless change so substantially as to become quite distinct from what they once were. For instance, a person with early-onset Alzheimer’s may initially state that they would want to donate their brain to research upon death. Yet, once decisionally incapacitated, their current desires and beliefs might diverge sharply, and they may come to be insistent that their brain not be used for research upon death (similar problems occur with advance directives about withholding feeds). How to weigh a decisionally incapacitated person’s present preferences against their decisionally capacitated past ones is far less clear than simply appealing to the person’s RAD. So, at the end of the day, an important role for surrogates seems needed.

Of note, there is an absence of recommended language that would constitute an ethically sound RAD in Hart’s paper. Given that Hart says RADs should “never be used to compel unwilling patients to participate in research,” we surmise that the participant’s assent to participate is nonnegotiable and that the duty to respect the “precedent autonomy” of the participant must be weighed against the duty to act in the participant’s best interest in their day-to-day circumstances. Settling the questions of whether the participant is able to assent and whether the research is consistent with the participant’s best interests will depend on the clinical facts and communication with the designated surrogate. Language that discloses these limits on what a RAD can accomplish in the research setting should be included in the directive.

Given these qualifications, we agree that RADs have a role to play alongside surrogate decision-making and capacity assessment in permitting Alzheimer’s disease patients to participate in trials. Because one of us has a family member that would benefit from improved medical interventions for this neurodegenerative disease, we support using every available tool we can to help further biomedical research in the most humane and ethical way we can.