Abstract
Several philosophical theories of scientific evidence (or confirmation) have been proposed. I argue, however, that none satisfactorily explain the way that evidence is gathered and used to confirm the variety of hypotheses that are commonly formulated in clinical medical science. I consider five philosophical theories of scientific evidence, and Inference to the Best Explanation to the extent that it can be considered a theory of evidence or theory choice. I argue that none adequately explain confirmation in clinical medical science, and I offer a new account of evidence that I argue does constitute a satisfactory explanation, and which I am calling the weight of evidence account. I divide hypotheses that are found in clinical medical science into three categories: therapeutic, etiologic, and diagnostic. Therapeutic hypotheses are those that are concerned with treatments or other medical interventions, etiologic hypotheses are those concerned with the causes of disease or other adverse medical conditions, and diagnostic hypotheses are those considered by clinicians when making a diagnosis. I illustrate the methods of gathering and using evidence in the confirmation process by using examples of each type of hypothesis drawn from the clinical medical scientific literature. I also argue that the weight of evidence account supplies a satisfactory explanation and rationale for the ìhierarchical pyramidî of evidenceñbased medicine, with randomized clinical trials and their derivatives, meta-analyses and systematic reviews of randomized clinical trials, at the top of the pyramid, and case reports, case series, expert opinion and the like at the bottom. Cohort, case-control, cross-sectional, and nonrandomized clinical trials fall into the middle. I illustrate the development of various ìlevelsî of evidence by considering the evolution of less invasive surgical treatments for early breast cancer, and argue that the weight of evidence account satisfactorily explains the notion of levels of evidence and other efforts to rank evidence. In addition, I provide a defense of randomization as a method to maximize accuracy in the conduct of clinical trials. I also consider ethical issues surrounding experimentation with medical therapies in human subjects, and illustrate and discuss these issues as they arose in studies of respiratory therapies in neonates and treatment for cancers of the anus and testis in adults. I argue that in many cases sufficient evidence can be accrued to warrant generally accepted new therapies without the need for evidence derived from randomized clinical trials.