The impact of patient treatment preferences on the interpretation of randomised controlled trials.

Reliable information about aggregate main treatment effects in cancer research comes from randomised controlled trials (RCTs). The possibility of important interactions, such as between treatment preferences and their effects, is necessarily subordinated in the quest for evidence about main treatment effects. If patient preferences can influence the effectiveness of treatments, for which there is some indirect evidence, then those estimates of the treatment's main organic effects from unblind RCTs might be wrong. RCTs clearly disallow patient choice and it is, therefore, important to know the extent of any preference effects in order to interpret the RCT evidence. It may be important to know whether they exist, and where and by how much they affect outcome. It is argued that measuring these effects reliably is methodologically difficult, and will require massive trials each directed at measuring one particular preference effect. Such effects have a slightly fanciful image, particularly in cancer treatment, and may be transient. Given the current uncertainties about their true nature and plausible biological mechanisms, the accumulated evidence is unlikely to provide sufficient justification for investing in such trials, given other current priorities.