These cases illustrate how the textbook RCT design has been altered to meet constraints imposed either by the researchers' objectives, or the actions of individual subjects, or both, in a particular social/medical/ historical context. Moreover, the adjustment from the "ideal" was understood to be a compromise, not a theoretically superior restructuring of the method. In practice, rather than in theory, the RCT may be a more flexible method than the mathematical model might suggest. Is it possible to formulate an RCT theory in which trial designers may make "accommodations" and "allowances" or to adjust material-mathematical model in order to address a socio-medical question?

     Each clinical trial involves several actors with specific goals: the sponsoring organization, the medical researchers and clinicians, and the patient/subjects. If these goals are very nearly similar, as for example in the streptomycin trials, the result of the RCT will likely receive a common interpretation and therefore the textbook model will be seen as successful. However, if the actors have different goals, if there are different formulations of the social-medical problem the RCT is to address, the tension between these is likely to structure the design and execution of the trial. Whether or not the trial in practice deals effectively with this tension will determine whether the results are interpreted as "solving" the problem and whether the model will be seen to "work".

     To consider the problem from a different aspect: a trial which precisely measures what it is designed to measure, in adherence with statistical theory, will have internal validity. However, unless its results can then be interpreted to apply in some sense to the population at large, it will not have external validity. If trial participants have different views as to what will constitute external validity, what will solve their particular socio-medical problem, these views are likely to force adjustments in the internal trial design.

     In fact, the randomized clinical trial, in practice, is more than a mathematical exercise, just as the methodology was only one part of the FDA's regulatory strategy in 1970. The technology incorporates not only the statistical assumptions and practice norms, but the goals and concepts of organizational, clinical, and social participants as well. Whether or not the material-mathematical and the social-medical problems inherent in a clinical trial are congruent, they are interdependent. The RCT results and their application in medical practice can only be understood in the context of both.

back