Robert (not his real name) was an angry and dying 60-year-old man, a victim of decades of cigarette abuse that gave birth to tumors that clogged his mouth, entangled his neck, and consumed his lungs and liver. It was amazing he was alive, but somehow, he was. Because he could not swallow secondary to the large hoard of tumors in his pharyngeal cavity, a feeding tube was placed to provide basic sustenance as well as supplementary nutrition during planned chemotherapy.

But before chemotherapy could be initiated, he pulled the feeding tube from his abdomen in a moment of confusion, leaving a whale-like blow hole that was emergently plugged with a urinary catheter, a not uncommon temporizing measure in long-term care facilities, especially in situations like this. X-rays the following morning were difficult to interpret for a nonradiologist, but tube feedings were restarted once the radiologist sent his translation that seemed to indicate the urinary catheter was correctly placed.

But within 24 hours, I found myself staring at an inflamed and obviously infected abdomen surrounding the feeding tube site. The urinary catheter was immediately removed to preclude further trauma to the area. I wondered if the tube was misplaced during insertion and the X-ray misread by the radiologist, if we assumed the radiologist's words implied the tube was correctly placed but it was not, or if Robert pulled and dislodged the tube after the X-ray was taken. No matter what, I feared that the viscous nutrition had penetrated into the subcutaneous tissues and perhaps the abdominal cavity, unleashing a torrent of unfriendly microbes. Unfortunately, only the forensics of an autopsy would reveal such a sordid tale, as he was far too ill to contemplate a surgical intervention.

Delirium had settled in with a vengeance now, and I wondered if it was the underlying disease, an infection, or an assemblage of causes that tangled his synapses and made us the enemy.

Haloperidol, lorazepam (watching for a well-described paradoxical reaction), and hydromorphone were administered subcutaneously; he calmed down some, but within 15 to 20 minutes, he was up again, screaming, fighting, and cursing the demons that were taking his last breath. We gave more of the same medications, but nothing seemed to help. I called his surrogate and discussed the limited options that were available. She was quizzical as to what had happened, why his abdomen was infected, and why he was delirious; I told her I didn't know, at least not right now. After a brief conversation, we elected to try a light sedation with midazolam, as he was obviously in the throes of a terminal delirium and needed sedation. The infusion was started and soon he was calm, the evil spirits tossed aside, a therapeutic lethargy calming his soul compliments of the soporific charm of pharmaceuticals.

It was then that I returned to the possibility of an iatrogenic complication. I again wondered if the urinary catheter was correctly placed after he pulled out the original feeding tube, or if it was inadvertently inserted into a foreign cavity feeding a frenzy of bacteria. And if it was correctly placed, did Robert surreptitiously pull the tube out far enough that the tube feeding flowed into the subcutaneous tissue after the X-ray was taken? Or was his abdominal crisis the result of another, nonrelated event confounded by the display of an angry erythema that lured us to an erroneous diagnosis? Or was this simply the result of a debilitating cancer, irrespective of what we did or did not do?

I wasn't sure, but I had a sickening notion.

I tried to rationalize, to make myself feel better, but it was difficult. Again and again I returned to the various scenarios. Again and again I reviewed the X-ray. But aside from being a medical detective, I was more and more becoming obsessed with an intervening concern: What would we tell his surrogate. What would we say?

My mind wandered, and in a likely defensive reflection I questioned why we were tube feeding this man in the first place, considering the extent of his underlying disease, as it was most certainly not prolonging his life and only serving as a ready source of nutrients for the tumors rapidly growing within his body. But that wasn't the issue at hand; the issue at hand was what we would tell his surrogate. And, unfortunately, his surrogate was incensed that the oncologist had not started chemotherapy earlier in the disease process, leaving her to consider what might have been if treatment would have occurred sooner, in spite of the large tumor burden that most assuredly would have negated any real long-term benefit of chemotherapy. And now, with no avenue for nutrition short of total parenteral nutrition, which in all likelihood would not prolong his life nor be approved by the Nutrition Support Team, Robert would die from what his surrogate described as “starvation.” I assured her that patients with such extensive disease did not die of starvation in the pure sense, but rather complications of the underlying disease. I'm not sure she understood my brief tutorial, nor am I certain she wanted to. Moreover, she related that when she was 25, she had had a “bad experience” with the death of her mother and the care provided by medical professionals, further complicating the emotion surrounding Robert's care as well as the potential for any adverse consequences of disclosure.

I contemplated my options: (1) say nothing and ignore her questions (not supportive of the physician–patient relationship or acceptable or permissible under the guise of medical professionalism), (2) lie (not ethical or moral), or (3) be honest and say I don't have any concrete answers, but there was a possibility that we contributed to Roberts decline.

I was aware of the current milieu of apologizing and of states such as Pennsylvania and Colorado that had legislated protection for physicians who apologized to patients and patients' families for committed errors; however, that protection did not necessarily preclude litigation for negligence, sustaining the refrain of many hospital lawyers to never admit guilt. And although not meant to lessen the importance of physical and/or mental harm to a patient resulting from a medical error, I was also aware that physician disclosure may translate into a demoralizing lawsuit with loss of reputation, money, self-esteem, and referrals; higher insurance premiums; and, eventually, reporting to the National Practitioner Data Base. Horrible consequences resulting from the mere act of truthful disclosure, but, then again, horrible consequences balanced by unfortunate injury to a patient.

But what we would be admitting to was not guilt, but an error, or rather the possibility of an error, an action that arose from our limitations as human beings. After all, we're not mechanized robots, but rather mortal beings with faults, and, from time to time, flawed decision making.

In my mind, there was only one thing to do. I chose to tell Robert's surrogate that although we did not know what had happened, there was a distinct possibly we may have contributed to his decline. She was cordial and understanding with a slight edge of discontent, but thanked me for the wonderful care provided by the nurses and doctors. I breathed a sigh of relief—I had done the right thing, and, at least for now, there were no unwanted consequences. And as a doctor entrusted with the care of a fellow human being, I felt reassured.

As for Robert, he continued to deteriorate, the tumors eroding his hold on life; in a calming slumber, he died a peaceful death a week later.

Postscript: Postdeath review of the X-ray again suggested the urinary catheter was placed in the body of the stomach.