Evidence in oncology. The Janeway lecture 2000

The purpose of this address is to evaluate the types of evidence that have fashioned clinical decision making in oncology over the past century and to determine the appropriate weighting of evidence at the start of a new century in the shaping of patient care, clinical research, and health policy.

Three separate analytic approaches were used: (1) a review of publications that altered medical practice, (2) a consideration of the misperception that little progress has been made by looking at the relationship between 5-year survival rates and national mortality rates, (3) an evaluation of the quality of data from clinical trials.

Published reports suggest that res ipsa loquitur was the dominant mindset of researchers in the first half century and continuing into the second half century. However, recognition of the scarcity of dramatic improvements in outcome and the possibility of incremental improvements led to the mounting of prospective randomized comparative trials that could identify such incremental improvements. Findings from these trials have profoundly altered patient care in the past quarter century. Data suggest that there is a sequence of events-increased survival rates in patients at research institutions followed by significant increases in survival rates nationally-followed by a reduction in annual mortality rates that do reflect improvements in treatment. Phase III comparative clinical trials yield the highest quality data in oncology. Meta-analysis of such data may be useful, but the most compelling data that alter medical practice come either from comparative clinical trials showing such significant differences in results as to necessitate their early termination on ethical grounds or from replicated phase III trials.

The nature of evidence influencing clinical practice and research has changed greatly in the past century. Although res ipsa loquitur still pervades the thinking of some oncologists, recognition of the incremental nature of progress is now the more dominant force in contemporary oncologic research. In addition, as molecular targets become the focus of therapeutic efforts, it will be especially important to proceed as quickly as possible to phase III comparative trials rather than to phase II studies.