부산대학교 도서관

However, although the individual molecule has no pressure, the individual patient does have an explicit treatment effect. This extreme example may be seen as simply the knocking down of a straw man; but in evidence-based medical practice there seems to be an all-pervasive, unquestioned assumption that individual probability is equal to the frequentist population probability - the average treatment effect. Using frequentist methods, all the variation in individual (between-patient) response is assigned to the random-error term in the analysis - and ever more precise estimates of average treatment effect are sought by designing ever larger trials. An eloquent paper by Gill et al. entitled " I Why clinicians are natural bayesians i " [4] showed how this process is used for diagnosis of patients, but the same reasoning can also be applied to patient treatment. [Extracted from the article]