Imprimir Republish

Good Practices

More transparent conflicts of interest

Boas Práticas aDaniel BuenoA group of specialists in medical informatics is proposing the creation of a global public registry of conflict of interest disclosures by researchers in the field of biomedical science.  These statements that are required, for example, when investigators submit papers to scientific journals or projects to funding agencies, identify the author’s associations with parties that are potentially interested in the research study, such as pharmaceutical companies or lobbying groups that could in theory generate bias in the results.  The idea for the  global registry was suggested in a review article that appeared in the inaugural issue of the journal Research Integrity and Peer Review, signed by Adam Dunn, associate editor of the new publication and researcher at the Australian Institute of Health Innovation, along with colleagues at the Harvard School of Medicine.

The proposal is based on five elements.  The first is to establish a commitment among funding agencies, institutions and journal publishers to enforce existing rules regarding conflicts of interest that are not always followed.  Dunn cites scientific literature that estimates that between 43% and 69% of articles and clinical trial reports contain gaps in information regarding conflicts of interest.  The article mentions the example of the initiative that in 2000 mandated the pre-registration of all clinical trials conducted in the United States.  It only became effective a few years later, however, when funding agencies and public institutions began to oversee compliance with the measure.

The second element refers to the definition of rules for providing and updating disclosures, in addition to creation of a team of auditors that can verify the authenticity of the data.  The third element concerns the possibility of linking the records of conflicts of interest disclosures available in databases maintained by science journals, universities and funding agencies.  One idea involves adopting an efficient system for identifying researchers, which could entail use of the Open Researcher and Contributor ID (ORCID) (see Pesquisa FAPESP Issue nº 238).  This resource would allow easy identification of all conflicts of interest involving each author.  The fourth element refers to creation of a consistent taxonomy for describing the various types of conflicts of interest to permit a more precise understanding of their implications.  The fifth involves the development of electronic tools that would automatically populate the disclosure forms based on information previously provided.

Dunn and his colleagues admit that creating a public registry may not be enough to prevent the publication of biased research, but it will go far to help better understand the effects of the problem and provide more trust and transparency to the work of scientists in the field of biomedical research.