SUMMARY

Nowadays, two simultaneous phenomena are taking place in the field of Medicine. On one side, the idea of basing the treatments and other issues involving the patient management on clinical research findings, especially clinical trials, has become stronger. On the other side, a suspicion is growing that a lot of these research findings are nothing but false. The most noticeable manifestation of this phenomenon would be the high rate of non-reproducibility through subsequent researches of most of the published studies.

A research finding that is “true” (and not a false positive) is dependent on three factors involving probabilities (the probability previous to the carrying out of the study that such result was true, the statistical power of the study, and the chosen level of statistical significance) and on a systematical factor (presence of biases). A biased study would be that in which, independently from the intentionality, one or more factors regarding the design, the sampling and data collection, the analysis, the presentation of results or the study’s publication itself, tend to bring about erroneous conclusions regarding the research subject. In this article, the biases related to the financing of the clinical research through institutions or enterprises whose interests can be affected by the results of these studies are reviewed.

The financial bias in clinical research has been shown in many studies that, basically, report that the trials sponsored by the pharmaceutical industry have, as a whole, a 3-5 times higher probability of presenting favourable results to their sponsors, than those without sponsorization. This phenomenon is produced in cases such as presentations within medical congresses, economical assessments, it is independent from the quality of the trial, it appears in every journal, even the most prestigious ones, it has place within the medical, surgical and other devices treatments, and every clinical field or specialty.

The mechanisms that contribute to the presence of this bias seem to be: 1) the disregard to the principle of equipoise, one of the core principles of ethics regarding human experimentation; 2) the presence of control clauses within the research contracts, inflicted by the industry and accepted by healthcare or scholar institutions and private research enterprises; 3) the use of non-adequate comparative alternatives, or incorrectly dosed ones; 4) the patient selection and the inclusion/exclusion criteria; 5) the manipulations of the ex-post data, including changes in the results measures, and the excessively creative analysis; 6) the omission of the negative results and the exaggeration of those positive, or the confusion regarding the report of the results; and 7) the publication bias, that fosters the diffusion of favourable trials and limits or hinders that of non-desired results.

Besides, the sponsored studies are disproportionately focused on the technologies and treatments that have a commercial value in spite of those having less commercial value, on “when” instead of “how” to intervene or even “if one should intervene or not”, and on the outcomes of interventions more than on their risks, in such a way that there is only “evidence” when somebody wants -and finances- something to be “evident”. In the last decade, several initiatives aimed at limiting the impact of the financial bias have been developed: mandatory trial recording, changes in the “uniformity requirements” to avoid the “caught” works, guides to handle the relations with the industry, governmental rules, good practice controls in the trials, the publication of every trial and some other one.

In any case, at the time of deciding whether the results of a research will be used or not to treat patients, it should be taken into account that the trust in the results of a trial must be greater: 1) as long as the sample size is bigger, 2) the effect range is greater, 3) the number of valued associations lessens, 4) the stiffness of the study (regarding its design, operative descriptions, result measures and analysis) is increased, 4) the involved financial interests decrease, 5) the comparator used and the study patients look more alike the treatments and the real patients.