There are important medical studies, and then there are landmark studies–the kind of science that disrupts the entire medical community. The most recent game-changer was published recently in the Lancet.

Well known surgeon and author, Dr. Atul Gawande and colleagues published this important look-back study on the intensity and variation of surgical care of 1.8 million elderly patients in the US. They put numbers to the well-known and ever-expanding problem of excessive care of those nearing end of life.

The Harvard researchers report that 1 in 3 elderly patients have surgery in the last year of life; nearly 1 in 5 have surgery in their last month, and 1 in 12 in their last week. Not surprisingly, regions with more hospital beds and high total Medicare spending did more surgery on the elderly.

Though politically and emotionally charged, I am here to report from the real world this truth. What gets done to the elderly in the name of “do everything” can be tragic. And sadly, it’s a growing problem.

You simply cannot find a more challenging part of being a doctor than the conundrum of how to apply the fury of modern medicine to the elderly. Nothing is more complex, more emotional, and more human. (Caring for AF pales in comparison to making the right decision for an acutely-ill elderly person.)

Let me tell you two stories:

My grandmother, Nellie, (we called her ‘Non’) lived next door to us growing up. To this day, I, nor my wife Staci, have ever met a kinder, gentler, more loving human. It was as if she was blessed before coming to Earth.

The call came during my cardiac fellowship almost twenty years ago. Non had abruptly developed severe belly pain. It was a terrible pain; I could hear her crying in the background as my Dad reported the news over the phone. And lousy news it was: she had progressive and diffuse blood vessel disease that had acutely obstructed an artery to the bowel. At age 84, with blockages in the heart, brain, kidney and carotids this was a terrible sign. There were only two choices: let Non die with comfort measures or surgery.

In the best case, which was unlikely, surgery would have meant a prolonged ICU stay, an extended and painfully long rehab, indwelling rubber tubes and possibly more surgery. She hated pain, and feared doctors. She was fiercely independent.

My wife Staci and I convinced my grandfather to let her go.

My only regret was that she died before we could say goodbye. For years after, my grandfather felt guilty. We reassured him that it was the right decision. As he got older and sicker, I think he understood the rightness of his choice.

The second story occurred many years ago. I was young then, but could make the same mistake again.

He was still a vigorous 92 year-old man, As a a multi-decade business owner that catered to the young, he had become a beloved figure in his closely-knit community. Like most his age, he lived in equilibrium with a ‘touch’ of heart disease. That was until sustained Ventricular Tachycardia delivered him to the Emergency Room. He was dying in front of our eyes; his family gathered at the bedside, watching him fade slowly away while his heart struggled from the rate of 180 bpm. Simple medicines had failed. He was well when he awoke and fading away in an ER bay in the afternoon.

There were again two choices: let him die, or intervene. Here, things were a little different. I thought; it’s just a shock. A little sedative and a millisecond-length shock. We cardiovert people every day. He’d be fine.

I cannot remember what I said to him, but the gist of it was … “let me try and save you.”  He didn’t want to be on machines or have prolonged life-sustaining care at the end of his life. But he sounded sad about dying that afternoon. It was just one shock, I had thought.

The problem occurred after the shock. His heart stopped. I was prepared to let him go, but the ER staff ran in and started what they do: Chest compressions, tubes and more shocks. I nearly passed out in anguish over making such an error. I went out to tell the family he was likely to die. But then he didn’t die; his heart rhythm came back. But then it stopped again an hour later. And then it came back. He did this for another day or so before dying.

That day still haunts me. I’m not sure what would happen today. It isn’t easy letting someone die in the afternoon, when they woke up well.

These two cases only scratch the surface of the complexity of dealing with acute illness in the elderly. These are crazy-hard concepts:

Futility. You can read the Skeptical Scalpel’s thoughts. He writes about the difficult situation surgeons face with unrealistic families. He’s right.

Expectations of the general public. I’m not sure the public understands the gravity of the situation with excessive care. Until one personally experiences the consequences (let’s call them downsides) of the fury of medicine it’s hard to know. People may fear the loss of a loved one, but they should also fear what medicine can do to their grandparents.

The issue of death. Accepting death simply doesn’t come easy. And it’s not just patients that struggle with mortality; it’s us doctors too. Cardiologists, in particular, frequently view death as failure. It’s how we are trained. Cynics say heart doctors do too many procedures because of greed. That’s hard to deny in some cases, but in most, we just don’t accept death. One more stent, a pacemaker, valve surgery–otherwise the patient could die.

Cost. Most think dollars. That’s true, but I’d also consider the human costs. Things like chronic pain, loss of dignity, and the incredible stress of family members charged with the care of an elderly parent kept alive by procedures. Would any of us in middle-age desire that in our last days? (Of course, knowing when your last days are near isn’t always clear.)

The lack of training and expertise of doctors in discussing end of life. Nobody taught me how to speak with patients and families about end of life issues, or goals of care. I learned myself that stopping life-prolonging care isn’t the same as stopping altogether. I leaned that at the end of life, patients can choose a strategy that emphasizes comfort, dignity and a better quality of what time remains. Do doctors and patients know the benefits of a multidisciplinary palliative care team?

Finally, at this moment, our healthcare system de-incentivizes end of life discussions. It’s so much easier to just do the pacemaker than it is to discuss the notion of a palliative care score. Or that fixing the heart rhythm won’t make a grandmother continent, or lucid or mobile. One hour to emotionally discuss these issues: no compensation—20 minutes to put in the pacemaker, a couple hundred dollars. Those are the facts.

It’s complicated. And a blogger that makes a living moving catheters in the left atrium surely does not have the answer.

Aggressive care at the end of life continues to grow. It’s crucial that we have more conversation and more education. This is why I believe Dr. Gawande’s paper holds such promise. Maybe, just maybe, people will start talking about how best to treat the elderly.

We can, but should we?

John Mandrola is a cardiologist who blogs at Dr John M.

Submit a guest post and be heard on social media’s leading physician voice.