Preparing for the End of Life: Medical Students Completing Their Own Advance Directives Described Increased Empathy for Patients

Discussing death is not easy. However, patients who consider end-of-life (EOL) preferences and complete advance directives (AD) before imminent death have higher quality of life,¹ reduce health care costs,^{1, 2} lessen their families’ emotional trauma,¹ suffer less, and enjoy higher satisfaction with their physician.^{1, 3} Still, many physicians avoid this discussion. Only one-third of Americans have written EOL preferences ^{2, 4}; only 10% have discussed them with their physicians. ⁴

Factors contributing to low frequency of EOL discussions may include physician discomfort, inadequate EOL discussion training,^{5, 6} and knowledge and skill gaps.¹, ⁷ Medical students and physicians-in-training feel ill prepared to address issues faced by dying patients.^8-10 Despite the importance of EOL education, most undergraduate medical programs offer students little systematic curricular exposure to advance care planning (ACP).^{11, 12} Education is typically opportunistic,¹³ occurring in hospitals in acute situations,¹⁴ or related to chronic illnesses near death,¹⁵ rather than occurring in outpatient, preventive settings appropriate for the person-centered EOL care approach recommended by the Institute of Medicine.⁵ With their focus on prevention, family physicians are well positioned to discuss ACP for inevitable death at routine visits when patients are relatively healthy. Family medicine faculty, then, could teach EOL counseling in a preventive health curriculum, which may enhance physician comfort and increase EOL discussions. For example, fourth-year medical students who completed their own AD reported greater knowledge and preparation for EOL conversations with patients.16 The most effective way to teach EOL, however, remains unclear.

To address this gap, we developed an experiential learning approach. Based on the concept of experiential empathy training,^16-18 in which students experience medical care from a patient’s perspective, we had students personally complete ADs. We wondered whether this exercise would foster empathy, which has been defined as understanding of patients’ emotions and communication of that understanding back to the patients,¹⁷ a key component of EOL conversations. Our medical students completed their own AD and reflected on this experience in writing. In this paper, we report on our analysis of evidence for empathy in these written reflections.

During the 6-week family medicine clerkship, third-year medical students each privately completed an AD using Aging with Dignity’s “Five Wishes” document,¹⁹ which covers personal, spiritual, medical, and legal wishes, including selecting a health care proxy (HCP; Table 1). Next, they submitted two-paragraph written reflections based on this prompt: “Describe how completing ‘Five Wishes’ changes the way you think about end-of-life care, and explain how going through ‘Five Wishes’ will change the way you practice medicine in the future.” Submission of the assignment with minimal reasonable effort earned full credit, which constituted 2% of the clerkship grade.

We collected all 548 student written reflections from 2016 through 2019 (100% completion rate). An administrator deidentified the reflections and securely stored the only link between student identities and reflections. Using the following methodology, we performed a qualitative analysis to evaluate and categorize what the students reported in their written reflections:

Three research team members, including two medical educators and practicing family physicians (D.E., A.L.), and a premedical student (K.W.), individually reviewed and qualitatively labeled each reflection using an open coding process,²⁰ describing each discrete concept or idea in the data and then grouping similar concepts into categories.

Next, the two physician researchers (D.E., A.L.) and a family medicine qualitative researcher (A.D.) worked together to review and clarify the code terminology and code categories and grouped these categories into overarching themes, referring back to the coded data until the researchers reached consensus on the themes. To reduce bias in the coding and theme development process, we included some researchers involved with the clerkship (D.E., A.L.) and others not involved with the clerkship (K.W., A.D.). Throughout the analysis, all researchers deliberately noted, reflected on, and discussed our potential biases as we interpreted the data. The Tufts Health Sciences Institutional Review Board granted exemption from review.

During the 3-year period, 548 student reflections were completed. Most adhered to the requested two-paragraph length; the average word count was 400.

Upon analysis of the prompted reflections, we identified 958 instances of self-reported change in students related to empathy and categorized them into five themes. After completing their own AD, students reported (1) future intention to encourage patients to complete ADs (n=312, 57%), (2) increased sensitivity for patients considering EOL choices (n=252, 46%), (3) awareness of EOL decision complexity (n=197, 36%), (4) appreciation for patient values and perspectives (n=139, 25%), and (5) appreciation for relationship-centered doctor-patient communication (n=58, 11%; Table 2)*. Additional categories, such as increased clinical knowledge about and comfort with EOL discussions, surpassed the scope of this current analysis of empathy.

In this analysis of written reflections regarding the experience of completing ADs, students expressed increased patient empathy. While we had hypothesized that students might gain knowledge about EOL discussions and greater respect for the complexity of these conversations, we were surprised by how many students eloquently expressed empathy for patients around complex aspects of planning for death. The prevalence of empathy-related reflections supported attentiveness to the values and human-centered aspects of family medicine,²¹ which are particularly important for conversation about death as an inherently vulnerable topic.

The present study adds to existing literature by describing a deliberate, rather than opportunistic, approach to teaching EOL issues with an experiential component in which students consider their own EOL wishes. Review of the literature found one other study describing learners completing ADs for themselves as part of the educational program. The University of Colorado had students complete the Conversation Project “Your Conversation Starter Kit”²² in small groups,²³ while the present study describes students completing and reflecting upon their own ACP individually. Positive learner outcomes were found in both studies. Adding the personal experiential completion of an AD could become a key piece of a comprehensive EOL curriculum when paired with skills practice and assessment, such as a faculty-observed counseling conversation with feedback. ²⁴

One study limitation is that by analyzing written reflections that were required for the clerkship, we cannot know students’ true feelings about EOL, only what they chose to write about in their mandatory assignments. The reflection prompt’s wording also suggested students would change somehow, by asking how they changed or might change. A more neutral prompt could have produced different responses. We also did not assess students’ actual skills or behaviors. Future research could incorporate Objective Structured Clinical Exams to assess students’ ability to introduce the AD process and other EOL decisions to patients. Other limitations include lack of data on who had prior EOL experiences and lack of demographic information (eg, married students may view choosing HCP differently). Future research could collect these data and compare reflective writings before and after the activity.

Just as medical curricula teach other preventive topics—nutrition, exercise, violence, cancer screening—training in EOL decision-making guidance as part of routine care should be a component of medical education and fits well with family medicine values.

*The sum of percentages exceeds 100% due to some entries discussing multiple themes.