Maurie Markman, MD
One can only applaud recent efforts by clinical cancer investigators to query both patients and their providers regarding a wide range of potentially highly relevant topics. Important aims of such efforts include the establishment of a better understanding of the goals of patients, their families, and their oncologists in a particular setting and the subsequent development of strategies to strengthen communication and shared decision making essential to optimize individual patient outcomes.
In the best survey research, great care is taken to word questions and potential responses in a manner that makes it possible to optimally and objectively examine what the respondents actually mean by a particular answer to a specific query.
The investigators then attempt to analyze the responses and subsequently draw general conclusions based on their findings. Depending on the goals and outcomes of a particular research project, the results may lead to future investigative efforts and the authors may also suggest implications of their project that should inform current clinical practice.
However, just as in any research endeavor where the investigators are required to provide a subjective interpretation of objective responses or results, there is frequently room to question those interpretations. Such challenges may be most relevant in settings where the research is conducted by academic physicians, particularly individuals in a different specialty from the patient/physician population being studied, or so-called “non-clinical experts” who do not directly or routinely care for patients, and where these individuals are attempting to interpret the reasons for certain actions of clinically active community-based practitioners.
Study Finds Communication Gap
Consider, for example, the results of a most interesting analysis conducted as a component of a larger survey of patients and oncologists which examined communication between the 2 groups.1
The particular issue addressed in this paper highlighted how oncologists communicate their personal “belief” regarding a patient’s prognosis with that individual.
Based on the survey responses, the investigators concluded that oncologists very frequently fail to inform patients of their own opinion regarding prognosis, which is often—and unsurprisingly— considerably less optimistic than the patient’s own belief.
The investigators use these data to provide a rather striking, and one might even suggest somewhat condemnatory, conclusion that “patient-oncologist discordance about survival prognosis was common and patients rarely knew that their opinions differed from those of their oncologists.”1
One cannot rationally disagree with the investigators’ conclusion that inadequacy of communication between patients and their physicians regarding critically important management and prognosis-related issues must be addressed both in individual practices and in the community of oncologists as a whole.
However, it is reasonable and appropriate to suggest a meaningfully different alternative hypothesis for what has been observed in the survey results that may explain a limited or perhaps even a substantial part of the reported outcomes.
A Deeper Look at the Data
It is instructive in this discussion to remember the highly personal goal-setting decision by the late eminent evolutionary biologist Stephen J. Gould when he was diagnosed with peritoneal mesothelioma and which he subsequently described in a magnificent and now truly landmark essay, The Median Isn’t the Message
Dr Gould noted the highly disturbing and welldocumented median survival of far less than 1 year for this malignant condition. However, rather than focusing on the “imaginary patient” who is somehow prospectively known to subsequently experience median survival, Dr Gould elected to objectively examine the facts but then to see himself as being one of those patients whose clinical course would ultimately be described as being at the favorable tail of that overall highly distressing survival curve.
And what was wrong with this decision and conclusion? He ended up outliving his cancer for nearly 20 years.
Similarly, it is reasonable to hypothesize that oncologists may be completely aware of the statistically defined median and overall survival rates in a well-documented population-based survival curve in a particular clinical setting relevant to a given patient.