Should medical residents be trusted with end-of-life conversations?

I stood next to the bed of my dying patient, knowing that I couldn’t say what she needed to hear.

I had seen her several months earlier during a previous hospital admission. The stethoscope I had pressed against her back then picked up only dull silence, not the lively whooshes of breath I was accustomed to hearing during a lung exam. A tumor compressed her airway, making it impossible for air to enter her right lung. The cycles of chemotherapy and radiation that followed left her weak and with a lingering burning sensation in her toes and fingers.

Now her lungs were filled with tumors, lighting up like dying stars on an X-ray.

She looked frightened. In between labored breaths, she asked me if there would be any more chemotherapy. I knew the answer was “no.” Maybe she did, too. She was far too weak to tolerate it, and a third round of chemotherapy could kill her far faster than her cancer would. She asked if she should call her family, talk to a chaplain, arrange for hospice.

I wanted to say “yes.” Instead, I said what I was supposed to say, “I don’t know. Let’s ask your oncologist in the morning.” I left her room, hoping that death wouldn’t reach her before her longtime oncologist could explain that there would be no more treatment.

She died the next morning, before she could have that conversation.

As a resident physician-in-training in a large teaching hospital, my job is to learn from senior physicians and to help treat patients. It’s sometimes surprising how much autonomy I am given as a newly minted doctor. After appropriate training, with little-to-no supervision I can run morning rounds, perform a spinal tap, and lead a Code Blue to try to revive a patient whose heart had stopped.

But that autonomy doesn’t extend to end-of-life conversations, even though it should. Residents often assume that a physician who has known a patient for years rather than days should begin these especially difficult conversations.

I initially didn’t protest when a patient’s long-term physician would ask me not to have an end-of-life discussion without him or her present. But the problem, I’m learning, is that these conversations often need to happen right away. I’m also learning that there are consequences to disregarding the expectation that residents defer these conversations.

I once broke the news to an elderly man in the intensive care unit that his upcoming lung transplant needed to be delayed because of his current illness. He was devastated and believed he would die before getting the transplant. “I don’t want anything heroic now. Let me go in peace,” he firmly told me. We talked for a bit, and eventually came to the consensus that if his breathing declined further, we wouldn’t perform CPR or insert a breathing tube. I felt proud that I had taken an hour out of a busy day to sit down and listen to his preferences.

His long-time lung doctor didn’t feel the same way. “Why didn’t someone check with me beforehand?” he asked, frustrated. I explained that the patient had brought up his preference against CPR willingly. “You should have stopped him and paged me,” the doctor stated, “I wanted to explain to him his options. Now the horse is out of the barn.”

Preventing residents from starting end-of-life conversations with patients robs us of a critical opportunity to learn by experience. “See one, do one, teach one,” is the motto by which residents gain experience. It’s how I learned to place a catheter in a jugular vein and withdraw fluid from a swollen abdomen. No matter how many training sessions we attend beforehand, we learn with a dearth of experience but a hope of success. We’re also required to prove proficiency: Residents at many hospitals are required to successfully perform certain procedures, like a spinal tap, at least five times under supervision before being certified to independently perform or supervise the procedure.

There is no such requirement that residents have a certain number of end-of-life conversations under supervision, even though the activity, supervision, feedback, and logging of the experience would be helpful. Instead, residents automatically presume a lack of authority, a consistent need to rely on others who may know a patient better. Always needing to ask permission to have an end-of-life conversation discourages residents from having them. Despite increases in palliative care education in the past decade, this might explain why nearly half of residents report not feeling comfortable having open, honest end-of-life discussions with patients.

Discussing death is a skill honed by practice, just like putting in a catheter. My first end-of-life conversation was a disaster. I was talking with a man who had just learned that his lung cancer had spread to his brain. While he was in the throes of digesting this devastating news, I bluntly asked, “Have you thought about what you might want at the end?” He and his wife immediately burst into tears, letting out hours of stress that had built during the day. Embarrassed, I muttered that I would return later that day. It took me another three days before I had the courage to discuss advance care planning with them again, and then did so with an attending physician present to help guide the discussion.

I’ve gotten better at talking with patients about end-of-life decisions, but only through actually having such conversations — and sometimes failing at them — with and without supervision. I now feel more confident starting this conversation, even with a patient I barely know, because I have learned from my mistakes.

But perhaps I don’t need to fail to learn to succeed. If I were given more time to shadow my physicians or simulate end-of-life conversations with immediate feedback, I wouldn’t have to learn by trial and error with a devastated patient. Structured exposure to end-of-life conversations, immediate feedback from mentors, and workshop training on shared decision-making have been shown to increase confidence in end-of-life communication among residents.

If residents become comfortable with end-of-life discussions, then more of them will happen, both during and after their training. This would be good for patients, because current physicians often fail to initiate these important discussions until the end of life. And it would be equally good for residents.

In the hospital, I see patients faced with life-threatening illnesses. Some need to begin thinking about end-of-life care soon, if not immediately. Rather than shifting end-of-life conversations away from residents like me, we should be given the space and training to have these difficult discussions because, in that moment, I may be the most appropriate person to do so.

Ravi Parikh is a resident in internal medicine at Brigham and Women’s Hospital in Boston.