AS A YOUNG CLINICIAN, I was interested in making a difference; it did not matter how much of a difference, as long as I could claim some patient benefit. And I really didn’t care what benefit: better survival, less local recurrence, shorter hospital stays, fewer narcotics—the specifics did not matter. With retrospective observational studies, it was easy to claim a benefit when I was demonstrating a potential or probable association. I intervened, and the patient improved! If I could prove that claim and make it look scientific with a randomized controlled trial, was I then done?
Sir Murray F. Brennan, MD
Early on, I thought I was done. My greatest challenge was to convince others to do as I had done. They had to believe my randomized controlled trial was applicable to the patients they saw, which was no easy task. Just because “the Chairman says it was so” did not mean it was so. Even with faculty compliance, it was difficult. I had to convince my skeptical colleagues that the difference I showed in a randomized trial was sufficient for them to change their entrenched practices.
As I aged, I also matured, but I still wanted to make a difference. Should we discuss the fine line between maturity and senility? A case for another day, if I can remember. Now, I still want to make a difference, even claim a benefit, but I am more interested in and aware of the consequences of my intervention. What does this presumed benefit cost?
If we have a randomized controlled trial that improves the outcome of any variable by 20%, that might seem to be enough, but if we’re studying a population that already has a 90% survival, 20% of the remaining 10% who might benefit is 2%. Am I justified to treat 100 patients, of whom 90 cannot benefit from my intervention because they were never going to have a recurrence or were already doing well on the current regimen, when 8 will not benefit regardless of my intervention?
The literature is replete with such studies. Should it bother me that 2,000 patients had to enter such a trial to see 40 benefit and statistical superiority claimed? Is it enough that the U.S. Food and Drug Administration approves the treatment and Medicare will pay for it?
Benefit, Cost, and Value
WE COULD JUSTIFY treating 100 patients for 2 to benefit, where benefit means alive vs dead, if there were no consequences to those who were to be alive regardless of the intervention. But no intervention—even no treatment (ie, placebo)—is ever without cost, if only emotional or financial. Do we ask if a new treatment regimen that will benefit 2 patients in 100 is accompanied by a complication rate that exceeds 2%? That depends on what you mean by complications, I hear you say. Serious complications are what matter. Grade 3 or more is your interest.
“It is encouraging to see physicians ask patients what they are willing to pay in time, money, emotion, or side effects for the presumed benefit. After all, the patient is the one with the disease.”— Sir Murray F. Brennan, MD
Tweet this quote
I recently developed cyclist ulnar palsy, and it is very annoying as I try to type—could I still perform surgery? Should I? Peripheral neuropathy may not trouble me much if I am a singer in a band, but what if I am the drummer, the guitarist, or the cellist (or the surgeon)? Too late now for me to be concerned, most trials where paresthesia is a potential side effect have excluded those over age 75, and there has been no benefit for those over 65.
At what operative mortality am I justified to remove a benign but potentially precancerous lesion? If there is a 2% benefit and a 1% complication, is that enough? What if a 5% benefit is associated with a major morbidity—how much benefit is enough? Should all intraductal papillary mucinous neoplasms be operated upon? They are all potentially precancerous, but the risk of malignancy in some is less than the operative mortality.
How do I approach this problem? How do I define benefit, or more importantly, value? If value is benefit divided by cost, how do I define benefit or cost? Cost is a widely disparate variable. Can we define both benefit and cost as being simply financial? Can we ever use the same units of measurement for both cost and benefit? Dividing days gained with the intervention by days lost because of the intervention is attractive but limited. Should I use days of “good life” gained divided by days of “good life” lost? What then is the acceptable ratio? Maybe 10:1 sounds good, but 3:2? You decide. When is what I do valuable?
IT IS TIME to acknowledge we may have the approach backward. If we have 90% survival and need a 4,000-patient trial to get to 92%, should we not be looking for less treatment that gives the same benefit with fewer side effects? If the rate of heart failure from doxorubicin and trastuzumab is 2% to 4%, and cardiac function impairment is 10% to 20%, should we not look at the converse, ie, no improvement of the 90% but no loss of effect and fewer complications or side effects for the 90% who cannot benefit further?
How do I answer my own confabulation? Should my earlier preoccupation of benefit first and financial cost second change to nonfinancial cost first and then benefit?
For me, it is time to accept that I may no longer be able to make an improvement, but I can maintain the same success at less cost. We have done that in operative mortality for major procedures.
In 3 decades, we have improved operative mortality for complex operations from > 20% to < 2%. We may not have improved cancer outcomes, but we have maintained “success” at fewer lives lost. With minimal-access techniques, we have further minimized perioperative morbidity.
Perhaps we should ask the patient, what does he or she want? It is encouraging to see physicians ask patients what they are willing to pay in time, money, emotion, or side effects for the presumed benefit. After all, the patient is the one with the disease. Isn’t that the value I was seeking? ■
Dr. Brennan was Chairman of Surgery at Memorial Sloan Kettering Cancer Center for 21 years and is now Senior Vice President of International Programs and Director of the Bobst International Center, Memorial Sloan Kettering Cancer Center, New York.
Disclaimer: This commentary represents the views of the author and may not necessarily reflect the views of ASCO or The ASCO Post.
DISCLOSURE: Dr. Brennan reported no conflicts of interest.