When I first met my patient, three years earlier, a woman in her mid-70s from Western Pennsylvania, she had just been given a diagnosis of myelofibrosis.
This disease is classified as a myeloproliferative neoplasm, in which a genetic abnormality revs up the inner machinery of cells in the bone marrow (“myeloid” cells), causing them to divide at a rate that far exceeds what is normal. As the cells proliferate, they secrete a chemical that causes the bone marrow to fill with scar tissue (“fibrosis”).
The remaining normal myeloid cells evacuate to safer territory in the spleen, and the spleen swells, squashing the nearby stomach but making the abdomen look overly full. Consequently, the person with the disease becomes progressively malnourished from an inability to take in enough calories. My patient had lost 25 pounds in the preceding six months as a result, even as her waistline expanded and made her feel uncomfortable, even unwieldy from her new girth.
I had prescribed the one drug approved by the Food and Drug Administration for such conditions, and it worked well for her: Her spleen shrank back to a normal size, she gained back some weight, and she had resumed having a good quality of life.
But the drug is not curative, and when I saw her again recently, she told me she had lost her appetite, she had dropped a few pounds, and I noticed that she was wearing pants with an elastic waistband, to accommodate her swelling abdomen.
“I figured the medicine had stopped working,” she told me. “I try to take in a few small meals over the course of the day, but I always feel full and can only eat but so much.”
I nodded sympathetically. We discussed the paucity of available treatment options and settled on a clinical trial of a new oral drug that had shown promise in preliminary studies. She asked me about the side effects of the medication.
I told her that any drug used to treat her myelofibrosis has the potential of making her blood counts worse, meaning she would be at higher risk of bleeding or infections and might need transfusions.
“It hasn’t been used in enough patients with your diagnosis yet to identify side effects that are rarer,” I said, having reviewed the clinical trial protocol before I talked with her. “For the same reason, I can’t tell you precisely how well it will work, though in previous studies between one-quarter and one-third of patients had symptoms that improved.”
She nodded. “Well, we have to do something to try and make me feel better. Let’s go for it.”
She returned to our cancer center for screening laboratory studies, and once we confirmed that she was eligible for the trial, we gave her a bottle of pills to start taking and she went back home.
A few days later, the research nurse poked her head into my office with news about our patient.
“Did you hear? She was admitted to her local hospital.”
“You’re kidding!” I said in disbelief. “What’s going on?”
“She felt lightheaded and went to the E.R. I think she had a bleed in her abdomen. You can check her labs and scans on the computer.”
Our patients’ medical records, as is the case at many hospitals, are all electronic, and have been so for about 15 years. As more and more hospitals have adopted electronic medical records, their records have become linked with ours, and we can now see results of laboratory and radiology tests performed elsewhere, enabling better continuity of care.
I accessed her medical record and reviewed her lab results in a hospital 130 miles away. She was anemic and her white blood cell count had quintupled.
“Do you think the trial drug caused this?” the research nurse asked. “Or did her myelofibrosis get worse?”
I couldn’t tell. I found a note from the E.R. doctor mentioning a CT scan that showed blood in her belly, but the actual scan results were still pending.
I logged back into the computer an hour later. The CT report, now available, showed active bleeding in her abdomen. I also saw a series of orders for platelet transfusions. Good — that’s what I would have done.
A couple of hours later, I read that she was transferred to the intensive care unit. They were giving her fluids for her low blood pressure. Then, her labs revealed a problem in her ability to form blood clots, which would make it hard to stem the bleeding. She was given clotting factors to help.
I felt as if, in starting her medication, I had launched a drone carrying aircraft ordnance and, like the military personnel who control these devices, was now remotely watching the havoc that its missiles had wrought. The substance of who my patient was had been reduced to blips on a screen detailing abnormal laboratory values, complicated medical terms like “disseminated intravascular coagulopathy,” and clipped, cautionary phrases in the doctor and nursing notes.
A nephrology consult was called for kidney failure. My patient was deteriorating in real time, on the computer.
The following day, she was transferred to our hospital. As I was about to leave my office to go see her, the research nurse stopped me. My patient had a cardiac arrest a couple of hours after her arrival and had died.
My initial instinct was to turn back to my computer and read the death note in her electronic medical record, the disembodied source of truth about her for the past 48 hours.
But it was no substitute for the person I had come to know over years.
Recognizing the foolishness of looking for the death note, I stopped myself and instead locked eyes with the research nurse for a few seconds. What I really needed was to share the loss of our patient with another human being.