“They need you in room 13,″ she said when I answered the phone and I ran back to the ICU.  The patient was coding and for each minute that felt like an hour, we tried, and failed, to save her.  She wasn’t breathing, her heart wasn’t working, and despite the 30 people gathered in the room, in the end, she died.

Running a code, as we call it, means that someone is dying or, technically, has died, and a team of doctors, nurses, respiratory therapists, technicians and others rush to the bedside and try, despite the odds, to bring her back.  Usually, we don’t succeed.  Running a code, as the code leader, means that we are emergently called to the bedside of a patient whom we often know nothing about.

We look at the patient, look at the data on the monitors, and start issuing orders.  But, not really.  We are leading a team, but each team member has been trained on what to do, has done this before, and nurses and respiratory therapists and pharmacists are bringing in supplies and getting medications ready and repositioning the patient even as the code leader asks the people around if anyone knows what happened.  How did this come to be? Who is this patient? So that we might have a clue as to what the cause is, and hopefully, what the specific treatment should be.

In the meantime, if the oxygen is low, we put in a breathing tube and deliver oxygen.  Sometimes that’s easy to do.  Sometimes it isn’t.  We check for evidence that the heart is circulating blood around the body and if it isn’t, nurses get up on the bed and start delivering chest compressions, squeezing the heart to make it push the blood around even as we try to figure out why.  Why did this patient’s heart stop working? What can we do to start it again, beside just hoping we’re delivering enough blood to her brain and the rest of her organs.

It is a harrowing, surreal experience, but also oddly inspiring.  A person has died.  She is not breathing, her heart is not beating, she is dead.  Yet a group of people are rapidly working with the singular goal of getting her back.

“The code leader doesn’t have to know everything,” a chief resident told our group of interns, as we all viewed the prospects of wearing our code pagers with trepidation.

“Identify yourself as the code leader,” we were told, and it was important.  Running into a chaotic room where people are doing things they are supposed to — putting in a breathing tube, getting IV access, holding the mask of oxygen, giving chest compressions, or setting up equipment and getting medications ready — it was intimidating to jump in and say, “I’m the code leader,” or “Who’s the code leader?”

As interns, first year doctors, we were likely to have participated in fewer codes than many of the nurses and therapists at the bedside.  And yet, we were told, it is structure that matters.  “You don’t have to know everything,” he said.  “But there has to be a flow of information and a plan.  Not just the chaos of people shouting out things to be done.”

Running a code is a team sport.  If you say the wrong thing, someone will remind you not to give this drug, or someone might suggest giving a drug or doing something else.  It isn’t a code leader saying “jump” and everyone asking “how high?”  While you’re running the code, someone else may have time to take a few minutes and read the patient’s chart and offer a key piece of history that would lead to a specific intervention.  Depending on the people at the code and their skills, sometimes the code leader leads the code at the bedside, but sometimes the code leader also has to do procedures — put in a breathing tube or a central line — while leading the code and thinking of what the heart rhythm is doing, what other drugs should be given.

And someone keeps track of time.  “Give epinephrine,” we say, and then what feels like an hour of chest compressions later we say, “give it again.”  But just to be sure we add, “How long has it been?” And someone who keeps track of what’s happening and when and writes it down, someone whose contribution may not, on the surface, seem all that important, says, “30 seconds,” so we say, wait, wait three minutes before giving it again.

People die every day, and yet, the codes where they die in the hospital and we try to reverse that process still shake us.  The people who expect to die, those who are truly terminal have decided that they do not want this final intervention at the time of death.  Through a process of grief and acceptance of reality, patients, and their families, have prepared for the end and have decided that when the patient dies, he is dead.  They have signed the DNR/DNI orders, and we do not pound on their chests or give them shocks or put in breathing tubes.  Even if we did, we would be unlikely to succeed, and in the end, we would simply be prolonging the dying process.

But it is the unexpected cases — the patient who seemed fine an hour ago, and now has a nurse doing chest compressions to get her heart to start again — those are the cases that really stay with us.   In the aftermath, we ask, “What could we have done?”  Not just, “How could we have run the code better?”

A process that is important for all team members in coping with these traumatic events and improving our chances of success.  A process that is both informal (a collective debriefing in the aftermath) and quite formal (review of all codes in the hospital by a committee and evaluation of the need for process of care improvements).  But we also ask, how could we have prevented this? How could we have foreseen that this might happen, that the patient might die, and what could we have done to prevent her from coding in the first place? Because the truth is, no matter how outstanding our code teams and how advanced our interventions, once someone has died — once they are no longer breathing and their heart is no longer beating — we’ve lost a huge battle.

“Thank you,” is not what we expect to hear after trying, and failing, to resuscitate someone who unexpectedly dies in our hands.  And yet, that is exactly what patient families say when they come, bleary eyed and devastated in the small hours of the night.  “Thank you for what you’ve done for her,” they say between sobs, and we offer condolences, and in the end, there is recognition, by us, and by them, that some things are beyond us, some things, despite the modern equipment and decades of training, some things just happen.

Denitza Blagev is a pulmonary physician who blogs at mybetterdoctor.