To date, internal company emails identifying marketing employees’ involvement in the design of trials from high profile legal cases, have been the only worthwhile evidence to understand the existence and features of marketing trials.
Such trials, also referred to as seeding trials, are primarily intended to encourage the use of study drugs by doctors, and try to convert them into promoters of the drug to increase uptake.
Our paper, on the characterization of marketing trials, published today in Trials, therefore set out to determine if medical journals actually publish such trials, asking how often they occur and describes their important features.
From the 194 randomized drugs trials, published in 2011, in the top six medical journals a team of clinicians, journal editors and medical researchers independently judged whether a trial had components that were more marketing than science.
We then extracted the trial characteristics, comparing those that were marketing to those that were not and analyzed if they clustered together into certain groups, which could make future identification easier.
From our judgments, we found 1 in 5 trials were more likely to be designed for marketing purposes when compared to those that were not.
Marketing trials were more likely to have noteworthy manufacturer contributions to the authorship, data analysis and the reporting.
The numbers of centers recruiting participants were also much higher in marketing trials (median 171 versus 13 for those that definitely were not marketing trials). At times this didn’t seem logical: many of these trials were for common diseases and each center could have easily recruited more people, making the logistics and oversight much easier.
Marketing trials were more likely to focus on surrogate and composite outcomes (though not necessarily use as the primary outcome measure). These endpoints are often not important to patients, are less likely to influence practice and have beenextensively criticized for their overuse.
Marketing trials were also more likely to include generalizable language when it came to describing the intervention in everyday practice, potentially encouraging its use outside of the researched study population.
There were, however, some positive features of the marketing trials when compared to non-marketing trials: they tended to be better reported in terms of blinding, safety outcomes, and adverse events.
To find out more about the characteristics of the marketing trials we also wrote to the editors to ask their viewpoints.
Of those that responded, they reported marketing considerations played no part in their publications and that the novelty of the science was the major component of the decision to publish or not.
Editors were also not willing to share peer reviewers’ reports, which is disappointing, given they might provide more insight to the components we described in our paper.
We also emailed the corresponding authors of all 194 trials: 11 of which resulted in a delivery failure. Of the remaining 183 trials, only 55 authors responded and many of these only completed part of the questionnaire. Such a low response rate and the quantity of missing responses meant we did not analyze these results due to the inherent bias.
Why authors didn’t want to discuss their trials was unclear. If they did, it would manifestly make it easier to understand what is going on, particularly given the importance to practice of this topic.
Further investigations are therefore required – if not essential – to verify the extent of the problems and to genuinely elucidate what is going on, especially given the novelty of our approach and some of the limitations in our study, including the subjectiveness of our decision making at times.
If you wanted to identify marketing trials what should you therefore look out for? Our cluster analysis proved to be disappointing as there is no pattern that would consistently identify such studies.
However, marketing trials often include manufacturers as authors with input to many parts of the study, they use a lot of centers for recruitment, and at times, they suggest using the drug in populations that were not actually included in the study.
Based on what we found, marketing trials probably occur more often than I first thought. Perhaps, because clinicians who are involved in a company-sponsored clinical trial significantly increase their prescribing preferences for the sponsor’s drug, irrespective of international guidelines, which is deeply disturbing when you think about it.