Form follows function: pragmatic controlled trials (PCTs) have to answer different questions and require different designs than randomized controlled trials (RCTs).
Aim Rising concern for demonstrated real world comparative effectiveness has heightened interest in “pragmatic trials” design. Pragmatic trials investigate whether the efficacy, presumed or found in explanatory trials under ideal conditions, can also be detected under real world conditions, i.e. effectiveness. It is also recognized that ‘real world’ effects which are usually addressed in public health research gain growing interest in confirming the ‘road capability’ of results obtained under ideal study conditions. This paper demonstrates that studies under ideal or real world conditions use different methods, generate different information and cannot replace each other. Subjects and methods The PCT design meets four requirements of public health and of effectiveness research. It includes all individuals who presented with the selected condition. It classifies the included individuals according to baseline risks. It enables plausibility controls. Finally, it compares the outcomes resulting from specified and notspecified interventions or treatments. Proposal We propose a pragmatic controlled trial (PCT) design in which patient preference and other co-factors crucial in determining the actual effectiveness of interventional options will not be neutralized by concealed randomization and blinding. This design is applicable to record the selected interventions and generated outcomes in day-to-day health care and is capable of incorporating preference and other participative factors into assessment of effectiveness. Conclusions The PCT design is useful for public health research, e.g. the effectiveness of interventions to change smoking habits or to prevent death frombreast cancer, as well as for comparative effectiveness research where it will supplement the traditional randomized controlled trial (RCT).
SiteringJournal of Public Health 21(2013) nr. 3 s. 307-313
Følgende lisensfil er knyttet til denne innførselen: