Executive Summary

Only half of internal medicine residents were comfortable discussing end-of-life care with patients even after observing such discussions and attending lectures, according to a recent study. Some recommended approaches from experts to teach residents this complex skill:

• Allow residents to practice discussions in a simulated environment with real-time feedback.

• Provide periodic booster sessions to keep skills current.

• Partner with residency program directors to craft curricula that address end-of-life discussions.


Internal medicine residents remained uncomfortable with end-of-life care discussions even after receiving additional training, according to a recent study.1 Researchers surveyed 83 internal medicine residents at a large academic medical center about comfort with end-of-life care discussions, and compared the findings to data from a similar survey completed by residents in the same program in 2006. The 2013 group reported having more lectures about end-of-life communication and had watched an attending have an end-of-life discussion more often than the 2006 residents. However, there were only modest improvements in their comfort level with such discussions; only about half felt strongly they were able to have open, honest discussions with patients and families.

Many practicing clinicians do not have any training in end-of-life communication. “Currently only 15 to 17 hours, on average, are spent on end-of-life and goals of care communication in medical and nursing schools,” says Salimah H. Meghani, PhD, MBE, RN, FAAN, and associate professor at Philadelphia-based University of Pennsylvania’s School of Nursing.

The most important consequence of inadequate end-of-life training is that patients receive care inconsistent with their preferences, says Meghani. “Clinicians often feel unprepared to approach the topic in the clinical setting, resulting in significant personal and financial tolls for patients and families,” she says. This includes uncertainty about care options and prognosis, multiple burdensome care transitions, care fragmentation, very late referrals to palliative and hospice care, poor quality of life and symptom control. “These are real and consequential matters for patients and families,” says Meghani.

Important deficiencies persist for training in end-of-life care, according to the September 2014 Institute of Medicine (IOM) report, “Dying in America: Honoring Individual Preferences Near the End of Life.”2 “Good communication is fundamental to good end-of-life care. All clinicians who work with these patients need communication skills training,” says James A. Tulsky, MD, one of the IOM committee members. Tulsky is professor of medicine and nursing and chief of palliative care at Duke University Medical Center in Durham, NC.

Communication about end-of-life care is an essential skill, says Tulsky; like any other complex skill, it requires training. “Unfortunately, most clinicians have never received any training in such tasks as delivering serious news or discussing goals of care,” he says. This is not routinely taught in medical school, residency, or subspecialty fellowship training; the education programs that do exist were generally created after most clinicians in practice received their training.

Simply hearing a lecture about how to communicate with seriously ill patients is unlikely to change clinician behavior. “To learn communication skills, clinicians must observe good communication, practice these skills, and then receive feedback on their behavior,” Tulsky says.

The opportunity to practice complex skills in a simulated environment, with residents getting real-time feedback on their performance, has proven effective for many other skills that residents are expected to perform competently. “The same is true for end-of-life conversations,” says Eytan Szmuilowicz, MD, assistant professor of medicine in the division of hospital medicine at Northwestern University’s Feinberg School of Medicine in Chicago. “What doesn’t work is just doing lectures — simply talking about how to have these conversations — because it omits the practice part.”

Szmuilowicz was lead author of a 2012 study that revealed a focused, multimodality curriculum can improve resident performance of simulated code status discussions; skill improvement lasted for at least two months after the intervention.3 “If residents are uncomfortable engaging in these discussions, then at the very least, they need to ask for some help — in the same way we, hopefully, work collaboratively in other areas,” notes Szmuilowicz.

Since there aren’t enough palliative medicine specialists to help every patient, however, the goal is for all clinicians to achieve a basic comfort level with such discussions in order to engage patients and families earlier in advance care planning. “To send trainees out into practice without these basic skills is unfair both to them and their patients,” says Szmuilowicz.

Here are some approaches Szmuilowicz has found to be particularly important involving effective training in end-of-life discussions:

There is a right time and place for learners to receive such training.

Trainees need to have some clinical experience on which to frame the skills. “Without that experience, we can teach them what to say and when to say it, but I don’t know how well it sticks,” Szmuilowicz explains. “Trainees might not be ready for some of the more complicated nuances early in their residency.”

Booster sessions are needed to avoid skills eroding over time.

The challenge is to find a way to dedicate the necessary resources in light of limited resident and faculty time. “The struggle we have is there is a limited amount of time to teach a lot of things,” says Szmuilowicz.

Mere observation is not enough.

Not all senior residents and attending physicians are well-trained and comfortable with end-of-life discussions themselves. “What I sometimes encounter is that people think of communication as a ‘soft’ skill, and that people should be able to pick it up by watching their senior resident or attending,” Szmuilowicz says. “That is not the best way to ensure competence.”

Trainees need to have somebody actively watching them and giving them feedback. “You can’t get good at this by just watching someone and then trying it yourself,” says Szmuilowicz.

• End-of-life discussions are a process, and need to be periodically revisited.

“These things are dynamic and they evolve over time,” says Szmuilowicz. “It’s not a one-time conversation. It’s great if one clinician can shepherd the conversation over time, but it’s not always possible.”

This means clinicians must think of this as a “team effort.” The idea may be introduced by one clinician, picked up later by another, and then revisited by a third. “We need to feel comfortable not always having a ‘conclusion,’ and we need to be better at communicating our progress with all of our colleagues,” says Szmuilowicz.

Szmuilowicz suggests that bioethicists partner with residency program directors to craft curricula that address end-of-life discussions, noting that interpersonal skills and communication are one of the six core competencies required by the Accreditation Council for Graduate Medical Education (ACGME).

“For something to occur routinely, it must be a seamless part of an existing work flow. Bioethicists may serve to normalize end-of-life conversations at the institutional and community level,” says Meghani.

Bioethicists could work with their local palliative care teams as well as their hospital leadership, suggests Tulsky, to encourage the identification of patients for advance care planning discussions. “Bioethicists can proactively identify patients for whom these discussions are relevant,” he says. “This needs to be a joint effort involving palliative specialists, other specialists, IT staff, and leadership.”

Bioethicists can then create systems interventions that ensure that identified patients have the conversation with a trained clinician. “This is a great opportunity for an electronic health records solution,” says Tulsky. “Clinicians could be flagged about appropriate patients, and perhaps receive point-of-care guidance about how to conduct the conversation.”


  1. Rhodes RL, Tindall K, Xuan L, et al. Communication about advance directives and end-of-life care options among internal medicine residents. Am J Hosp Palliat Care. 2014 Jan 12. [Epub ahead of print].
  2. Szmuilowicz E, Neely KJ, Sharma RK, et al. Improving residents’ code status discussion skills: A randomized trial. J Palliat Med 2012; 15(7):768-774.
  3. IOM (Institute of Medicine). 2014. Dying in America: Honoring individual preferences near the end of life. Washington, DC: The National Academies Press.


  • Salimah H. Meghani, PhD, MBE, RN, FAAN, Associate Professor of Nursing, Department of Biobehavioral Health Sciences, University of Pennsylvania, Philadelphia. Phone: (215) 573-7128. Fax: (215) 573-7507. Email: meghanis@nursing.upenn.edu.
  • James A. Tulsky, MD, Professor of Medicine and Nursing, Chief, Duke Palliative Care, Duke University Medical Center, Durham, NC. Phone: (919) 668-7215. Fax: (919) 684-0572. Email: james.tulsky@duke.edu.
  • Eytan Szmuilowicz, MD, Assistant Professor of Medicine, Division of Hospital Medicine, Feinberg School of Medicine, Northwestern University, Chicago. Phone: (312) 926-0001. Email: e-szmuilowicz@northwestern.edu.