Words, Truth, and Joy in Scientific Research
It’s my pleasure to introduce today’s guest post by Timothée Olivier. Dr. Olivier is a medical oncologist based in Geneva University Hospital (Switzerland). He joined me and my team (@vkprasadlab) at the University of California San Francisco (UCSF) for a one-year research program starting in November 2021. I must say, I deeply enjoyed his company, working with him on these many projects, and this essay of his. These are his reflections of the year.
Vinay Prasad MD MPH
Words, Truth, and Joy in Scientific Research
Words are more than mere information: they carry emotion, hope, good or bad news. Any provider interacting with patients with cancer and their relatives has to learn, sometime the hard way, when to use some words, and when to avoid others. In other situations, mastering the art of silence – “no words” – may be highly valuable.
Words can mislead people, and this theme is central in many of our research projects. In one project, we described how a “Physician’s or investigator choice” treatment was often an illusory choice (1): it was restricted to limited options, potentially preventing the most efficient treatments. As a potential consequence, a new drug may be tested against a straw man comparator. Consequences? Prescribing doctors may wrongly infer from the report that the choice was unfettered, and they may draw conclusion about the new drug based on this false assumption. Ultimately, this can affect patients’ care.
Truth is an audacious word; one can also use “fairness” . A central theme of our research is trying to get closer to the truth when it comes to inform patients about anti-cancer treatment, it’s efficacy, and toxicities.
An important trial in melanoma was the CheckMate-067 trial. (2) I remember when I saw the 5-year updated results, and the fact that quality-of-life was the same in the 3 arms of treatments. I knew one arm – the combination treatment – was much more toxic than others. I remember thinking: “Really? How is this possible? Same quality-of-life with such differences in potentially long-lasting toxicities?”. This was simply not congruent with my practice.
I decided to take a deep dive in the trial, and findings were sobering: the proportion of people not filling-out their quality-of-life questionnaires were much higher in the combination arm, and this was not taking into account in the analysis. In other words, quality-of-life looked the same, probably because people who were suffering the most were the least likely to send back their quality-of-life questionnaire! (3)
I learned important lessons during this project. First, when a report doesn’t fit with your experience: follow your instinct, explore, investigate. Second, behind many research projects are human stories. To me, this was the story of a 46-year-old woman treated with immunotherapy for metastatic uveal melanoma and who developed bilateral hearing loss and severe gait imbalance as a result of immune-related toxicity.(4) And this was the story of many other patients I took care of.
Our work is all about fairness: knowing potential pitfalls in quality-of-life analyses helps me, as a physician, better inform patients and my own practice.
No one can deny cancer is a terrible diagnosis, and a too often life-threatening disease. In a sense, it may appear inappropriate or tabou to talk about joy in the research-field of oncology. However, joy is present in many discussions between physicians and researchers, and is a highly powerful motivation in the research process. It has to be acknowledged… and praised!
Sometimes, it feels like we are playing a game, trying to solve a problem, learning to understand and appraise the surrounding reality. In other instances, there is joy when a new idea pop-up, when your colleagues or mentor reacts (positively!) to a new project proposal, when a work is finished, and when an article is accepted!
Finally, there is a deep joy in seeking for more fairness for patients, in aiming better for people. Their is joy in our purpose.
A project that gave me joy? Surprisingly, my first thought goes to one of the most complicated works I led. We looked closely at dose reduction rules and other rules differences between treatments in randomized trials. At the beginning, it was very hard, fastidious, and time consuming, like when you begin an enormous jigsaw puzzle. However, when later came the data analysis phase, our findings were astonishing... in 55% of head-to-head registration trials, those rules were biased in favor of the new drug! (5) The implications for this finding were massive: in those cases, it is difficult to tease apart between whether the new drug was truly better than the old one, or just looked better due to favorable rules.
Science as a living process.
Here is what I learned, more than anything, during this research-year: the beauty of the scientific process.
The scientific process is a safeguard: in the pursuit of joy, you can risk exploring only positive findings, or be pleased by fancy or popular explanations and fool yourself… and others!
In the pursuit of nice or clever words, you can distort your findings, again to please others interest. Your priors’ beliefs and expectations may mislead you. Rigor in the scientific process helps you to avoid tricks and traps. The scientific process teaches you to accept what you see, to question your data, to always push one step more.
What I learned? Science is a vivid dynamic process between the seek of Truth, the seek of Joy, which ultimately manifests through Words. And I’m grateful for doing scientific research.
Dr. Olivier is a medical oncologist based in Geneva University Hospital (Switzerland). You can follow him on twitter @Timothee_MD
1. Olivier T, Haslam A, Prasad V. Reporting of Physicians’ or Investigators’ Choice of Treatment in Oncology Randomized Clinical Trials. JAMA Netw Open. 2022;5(1):e2144770. doi:10.1001/jamanetworkopen.2021.44770
2. Larkin J, Chiarion-Sileni V, Gonzalez R, et al. Five-Year Survival with Combined Nivolumab and Ipilimumab in Advanced Melanoma. N Engl J Med. 2019;381(16):1535-1546. doi:10.1056/NEJMoa1910836
3. Olivier T, Haslam A, Prasad V. Informative censoring due to missing data in quality of life was inadequately assessed in most oncology randomized controlled trials. J Clin Epidemiol. 2021;139:80-86. doi:10.1016/j.jclinepi.2021.07.013
4. Köessler T, Olivier T, Fertani S, Marinari E, Dutoit V, Dietrich PY. Ipilimumab-related hypophysitis may precede severe CNS immune attack. Ann Oncol. 2016;27(10):1975-1976. doi:10.1093/annonc/mdw255
5. Olivier T, Haslam A, Prasad V. Dose modification rules and availability of growth factor support: A cross-sectional study of head-to-head cancer trials used for US FDA approval from 2009 to 2021. Eur J Cancer. 2022;172:349-356. doi:10.1016/j.ejca.2022.06.023