This article is more than 5 years old. Information may no longer be current.
Few textbook answers when making decisions amid significant uncertainty
Click Here to Manage Email Alerts
“You’re the doctor. What would you do if you were me?”
Many of us hear this sentiment from our patients who are confronting difficult health care decisions. This is especially true in my field, stem cell transplantation, in which the implications of the decisions being made have profound consequences, up to and including a near-term risk for death.
Earlier in my training, I found that these questions were easier to answer, probably because I failed to appreciate the extent of the impact of patients’ individualized values and preferences upon these decisions, and also because I hadn’t come to appreciate my own ambivalence about how I might respond under similar circumstances.
William Wood
I had previously assumed a “right” or “textbook” answer to these questions, based on evidence in the literature or local consensus. As time has gone on, though, I find these questions increasingly difficult to answer.
Sometimes I’m tempted to respond that I have no idea what I would do, but this would hardly be reassuring to someone turning to me for help. Instead, I try to outline the choices as completely as I can, acknowledge my own uncertainty and the impossibility of substituting my own values and preferences for someone else’s, and end with a tentative recommendation.
The conclusions to these conversations have become less and less satisfying to me over time. What would I really do? Given biologic uncertainty, incomplete data and a lack of precision in my predictions, how useful are my answers?
Informed decisions
In clinic a few weeks ago, I met with a young man in his late 30s who had a diagnosis of B-cell acute lymphoblastic leukemia. He was there with his wife and young son. He’d been treated with hyper-CVAD therapy and had entered a morphologic and molecular remission.
At diagnosis, he was noted to have a specific molecular translocation associated with higher risk. After 21 days of treatment, he had persistence of minimal residual disease by flow cytometry (subsequently eradicated). We discussed the role of consolidative myeloablative, total body irradiation (TBI)-based allogeneic transplantation in first complete remission, as he had a fully HLA-matched sibling donor available.
In general terms, we reviewed the ECOG/Medical Research Council and Dutch-Belgian Cooperative Trial Group for Hemato-Oncology (HOVON) data. We discussed expected toxicities of TBI-based myeloablative transplantation.
We came to a pause in the conversation, and he asked, “You’re the doctor. What would you do if you were me?”
I was struck at the time that a simple answer to this question — “I would choose transplant” — would be very misleading. Having practiced in this field for a few years, and having seen the trajectories of treatment and response for many patients — some positive and others negative — I knew there was so much that would need to be factored into my own decision as a patient, the volume of which seemed nearly impossible to try to convey during the time constraints of the clinic visit.
How could I try to explain the concept that there were long-term survivors in the large ALL study who never underwent transplantation? How could I try to explain the myriad predictable and unpredictable effects of myeloablative TBI-based conditioning upon symptom burden, quality of life, out-of-pocket cost, effects on caregivers, and risk for permanent disability or mortality? How could I try to accurately discern the probability that this particular patient, relative to others with this disease in first remission, might experience short- and long-term positive or negative outcomes from transplant?
As a clinician, this is the information I wanted to feel more confident in my recommendation and, as a patient, this is the type of information I would have needed to make my decision — even if I was able to guess with reasonable certainty what my ultimate decision would have been.
Comfort care vs. surgery
I recently attended on the inpatient clinical service and found myself in the middle of another very trying management decision.
I was in the room of a patient in his late 50s who had a reduced-intensity transplant for myelofibrosis 3 months earlier. Although never engrafting robustly, his peripheral blood counts had declined during the previous weeks and he was now severely pancytopenic, with no neutrophils and a platelet count in the 20s.
His marrow and peripheral blood showed mixed chimerism. His most immediate threat, however, was his profound anemia. He had developed a warm autoantibody and multiple allo-antibodies, as well as a blood type (O positive) as a mix of his native type (A positive) and donor type (O negative).
There was limited availability of phenotypically matched red blood cells for transfusion. His hemoglobin was 3.7 g/dL. He hadn’t responded to 2 mg/kg of steroids or rituximab (Rituxan, Genentech/Idec Pharmaceuticals) and derived no improvement in his hemoglobin value with red cell transfusion.
By laboratory parameters, his ongoing hemolysis did not appear to be brisk. His spleen was more than 20 cm in size, not appreciably different than its estimated size before transplantation.
With no guarantee of benefit or success, the only realistic therapeutic maneuver left was splenectomy. I sat in his room as he lay in bed, unable to move and unable to speak more than a few sentences without becoming winded. We discussed transitioning to comfort care vs. reversing code status and attempting surgery in 48 hours.
I had a hard time visualizing a long-term successful outcome. An operation and the immediate postoperative period would be exceptionally high risk under the circumstances — no neutrophils, a hemoglobin less than 4 g/dL, difficulty obtaining blood and no real response to transfusions. Even if successful, there was still the issue of declining chimerism, profoundly poor hematopoietic function and no short-term availability of additional cells for rescue. Onset of infection seemed inevitable.
When he asked me, “You’re the doctor, what would you do if you were me?” I truthfully answered that I didn’t know what I would do under the circumstances. I worried that proceeding aggressively would lead to suffering and a poor outcome, with missed opportunities for him to say goodbyes and to find peace.
‘Staggering’ complexities
As it turns out, my clinic patient elected to proceed with transplantation. This is the decision that I would have made, too, and from what I can tell, it was the right decision for him — despite the limitations of my ability to provide perfect transparency and personalized prediction about his expected short- and long-term outcomes with the path that he chose.
The patient in the hospital elected to proceed with splenectomy. He survived the operation, and his hemoglobin rose to 8 g/dL, where it stayed. Within about a week, his white blood count rose to 5,000 cells/mcL, with an absolute neutrophil count of more than 2,000 cells/mm3, and his platelets improved to more than 80,000 platelets/mcL. His chimerism converted to fully donor. He left the hospital feeling better than he had at any time since before his transplantation.
If I have learned anything from this, it is that the more experienced I become, the less I know.
The biological complexity of the diseases we treat and the physiologic responses we induce with our treatments are staggering.
Moving forward, my hope is that we can leverage the power of “big data” to develop personalized prediction models for outcomes that matter to our patients. We also must make this information as interpretable and digestible as possible, so our patients can spend their time integrating their own values and preferences with the decisions they are making.
No matter how sophisticated our prediction models might become, we will need to retain a large dose of humility because decisions made in an environment of significant uncertainty — which arguably is much of what we do in the practice of hematology/oncology — can lead to unpredictable outcomes on the other side.
For more information:
William Wood, MD, is an assistant professor of medicine in the division of hematology/oncology at the University of North Carolina in Chapel Hill. He also is a HemOnc Today Editorial Board member. He can be reached at UNC Health Care System, Division of Hematology and Oncology, 101 Manning Drive, Chapel Hill, NC 27514; email: william_wood@med.unc.edu. You also may follow him on Twitter (@WoodBD).
Disclosure: Wood is a medical consultant for Best Doctors.