The ongoing discourse surrounding Advance Care Planning (ACP) has been reignited by a recent study published in JAMA Internal Medicine, offering a meticulously designed randomized controlled trial (RCT) that challenges long-held assumptions about its efficacy in improving patient-centered outcomes. This analysis, building upon a previous examination of ACP’s perceived limitations, delves into the study’s innovative methodology and its potentially paradigm-shifting implications for healthcare policy and practice.
Groundbreaking Study Design Addresses Long-Standing Research Conundrums
For years, researchers have grappled with significant confounding factors in observational studies attempting to quantify the benefits of Advance Care Planning. A common observation has been that individuals who complete healthcare directives (HCDs) often experience longer hospice stays. However, the causal relationship has remained elusive. It could be that ACP directly leads to longer hospice stays, or conversely, that individuals predisposed to ACP completion are also more inclined towards less invasive end-of-life (EOL) care. This inclination stems from underlying personal and social characteristics that prioritize planning, acceptance, and specific notions of dignity at EOL, often emphasizing home-based care and prioritizing quality of life over mere quantity of life.
Even randomized controlled trials (RCTs) in ACP have faced a unique challenge: participant selection bias. Critics argue that individuals who are already receptive to ACP are more likely to volunteer for such studies. This inherent enrichment of "pro-ACP" participants could skew results, making it difficult to demonstrate significant benefits from an ACP intervention compared to a control group that may already lean towards less aggressive EOL treatments. The JAMA Internal Medicine study, however, employed a sophisticated design intended to mitigate this very issue.
The Zelen Design: A Novel Approach to Real-World ACP Intervention
The study utilized what is known as a Zelen design, a methodology that prerenalizes subjects before obtaining their consent for the intervention. This means participants are assigned to either the intervention or control group at the outset, and the analysis then compares these groups based on this initial randomization, irrespective of whether individuals within the intervention arm ultimately consent to and receive the ACP services.
This approach aims to accurately reflect the real-world impact of implementing an ACP program within a healthcare system, particularly for high-risk patient populations. When patients are approached for consent for an ACP study, a substantial number may decline participation for various reasons, including a reluctance towards ACP itself. This selective dropout leaves a cohort of predominantly "pro-ACP" individuals, hindering the ability to assess the intervention’s effect on the broader at-risk population.
The Zelen design circumvents this by randomizing half of the at-risk population to the active intervention group. Subsequently, only this group is approached for consent. Within this active treatment group, some will consent, and some will decline the ACP intervention. However, for the purpose of analysis, the entire active treatment group—both those who consented and those who declined—is compared against the control group, which received routine care. This methodology mirrors the actual operational impact of an ACP program in a healthcare setting, where not all offered services are utilized. Ethically, this is considered sound because all data is derived from administrative healthcare records, and individuals receive standard care unless they explicitly opt for the ACP intervention.
Study Details: Targeting Frail Elders in North Carolina
The study focused on community-dwelling older adults in a five-county region of North Carolina, specifically enrolling individuals with significant comorbidities and frailty. The ACP intervention itself was structured around the Respecting Choices program. It commenced with a nurse-led educational visit delivered via telephone. Following this, the nurse accompanied the patient and a designated caregiver to a primary care visit. During this visit, the primary care provider (PCP) engaged in an ACP discussion, guided by prompts to explore prognosis, disease understanding, unacceptable EOL states, and preferred treatment limitations, such as the desire for a Do Not Resuscitate (DNR) order. An Electronic Health Record (EHR) template and platform were provided to facilitate documentation of these discussions by the providers. The control group received usual care, with their providers having access to the same EHR ACP platform but without any specific encouragement or enablement to utilize it.
The researchers lauded this study design as exceptionally robust, particularly for its integration of a nurse-led educational component with a structured prognosis and goals-of-care discussion facilitated by a physician.
Key Outcomes and Findings
The primary outcome measured was the new documentation of ACP discussions within the EHR following randomization. The study also tracked a variety of secondary outcomes, including healthcare utilization.
A total of 765 patients were randomized. Within the intervention group (379 patients), approximately 90 individuals could not be approached for consent due to logistical challenges such as inability to locate them. Of the 294 patients who were approached, 146 consented to the intervention, and 139 successfully completed it. The participant demographic data revealed a mean age of 78 years, with 60% being female, 17% Black, and 80% White.
The intervention demonstrated significant success in its primary objective: 42% of patients randomized to the intervention group had ACP documentation in their charts, compared to only 4% in the usual care group. Patient-reported feedback on the quality of ACP communication was overwhelmingly positive, with participants generally rating it very highly.
However, beyond the increased documentation of ACP discussions, no other measured outcomes showed significant differences between the groups. Approximately 10% of the total study population died during the follow-up period (median follow-up of 304 days). Analysis of healthcare utilization metrics, including emergency department (ED) visits and hospitalizations, revealed no statistically significant disparities between the intervention and control groups.
It is important to note that the study did not explicitly investigate outcomes related to suffering reduction, an area the author of the original article had previously hypothesized as a potential avenue for ACP research.
Analysis: The "So What?" Question for Advance Care Planning
The findings of this well-executed RCT present a critical juncture in the ACP discourse. At a high level, the study confirms that a comprehensive, multi-component ACP intervention successfully increases the documentation of ACP discussions. For proponents who view ACP as an intrinsic good for patients, this outcome is unequivocally positive.
However, the absence of demonstrable impact on other key patient-centered outcomes—such as reduced suffering or avoidance of unwanted, burdensome EOL interventions—raises a pertinent question: "So what?" The author argues that despite the study’s methodological rigor and its success in increasing ACP documentation, it does not provide compelling evidence that ACP fundamentally alters patient experiences or outcomes in ways that matter most to them.
The study’s strength lies in its ability to capture a population that is already predisposed towards less invasive care. The implication is that the act of completing ACP documents may be a proxy for a patient’s pre-existing preferences rather than a catalyst for changing those preferences or their healthcare trajectory.
The Damning Exploratory Analysis
An exploratory analysis within the study, which compared patients within the intervention group who consented to the ACP intervention versus those who declined it, offered particularly striking insights. Patients who agreed to participate in the ACP intervention exhibited fewer hospitalizations and ED visits than those who refused. This might initially suggest that the ACP intervention itself led to these reductions.
However, the author’s analysis posits a more nuanced interpretation. When the entire randomized group (including those who consented and those who declined) is compared to the entire control group, no overall differences in outcomes emerge. This suggests that the control group, by its very nature, likely contained a similar proportion of ACP-interested individuals as the randomized group. Therefore, the observed differences within the intervention arm might be attributed to the pre-existing preferences of those who opted into ACP, rather than a direct effect of the intervention itself.
This leads to a critical conclusion: ACP interventions may not be the causal agent of improved outcomes, but rather a reflection of patients’ existing desires for less invasive EOL care. The millions of dollars invested annually in ACP initiatives, therefore, might represent a significant opportunity cost. The author suggests that had these resources been directed towards investigating alternative methods for mitigating suffering and providing care aligned with patient preferences, the impact on patient well-being could have been far greater.
Broader Implications and Future Directions
The implications of this study extend beyond academic discourse. If ACP interventions do not inherently change patient outcomes, but merely document pre-existing preferences, then their widespread implementation as a standard of care warrants re-evaluation. The author suggests that ACP should be treated as an investigational intervention, not yet ready for broad-scale adoption without further evidence of tangible patient benefit.
The comparison to medical devices or pharmaceuticals is apt: if a well-designed trial successfully increased the adoption of a medical device but failed to demonstrate any improvement in patient well-being, the justification for widespread implementation would be questioned. Similarly, the author contends, ACP should not be embraced as a standard of care in the absence of proven patient benefit beyond documentation.
This research highlights the need for a paradigm shift in how ACP is conceptualized and evaluated. Instead of focusing solely on the completion of directives, future research should prioritize understanding and addressing the underlying factors that contribute to patient suffering and the desire for specific types of EOL care. The substantial financial and human resources currently allocated to ACP could potentially be redirected towards developing and implementing more effective strategies for palliative and EOL care that demonstrably improve the quality of life for patients with advanced illness.
The study, while potentially disheartening for some in the ACP field, offers a crucial opportunity for evidence-based refinement of healthcare practices. It underscores the importance of rigorous research methodologies, such as the Zelen design, in untangling complex causal relationships and ensuring that healthcare interventions are truly beneficial to the patients they aim to serve. The opportunity cost argument, in particular, compels a critical examination of resource allocation and a renewed focus on interventions that demonstrably alleviate suffering and honor patient wishes.
