Are Impact Factors corrupting truth and utility in biomedical research?

What motivates governments to spend taxpayer money on iomedical research? A pragmatic answer would be to improve uman/animal health and to support biotech/pharma. Clearly, he research output from such expenditures should be both true repeatable, robust, reliable) and useful (helpful, applicable, growng knowledge). Most biomedical research funding is competitively warded as twoto five-year grants or fellowships, with success ates often ≈20% or less, depending upon the grant mechanism. he efforts of ≈80% of applicants (mostly taxpayer-funded scienists) are thus arguably wasted, and therefore expensive in terms f funds, human resources and opportunity costs [1]. A key perforance indicator (KPI) or criterion for award of research funding is ournal publication output, whose quality is often judged by the mpact Factor (IF) of the journals within which the applicant(s) ave published. A timely paper in a high IF journal (e.g., Nature, ell, Science), can mean the potential difference between continued unding or searching for a new position. Over-reliance on IFs as KPIs as been deemed inappropriate for a range of reasons including: (i) he processes used by the Thompson Corporation to calculate IFs re neither transparent nor openly available (although it should e stated that this represents intellectual property of a for-profit ompany, not something usually made public), and was actually riginally devised as a metric to help librarians determine which ournals to purchase; (ii) very bad publications can end up in high F journals [2], whereas publications with profound benefits for ealth or science may not (inter alia Gardasil [3] and DNA sequencng [4]); (iii) journal IFs are an average with enormous and growing ariances, with the most cited papers increasingly coming from ournals with lower IFs [5]; (iv) IFs can be “gamed” by editorial olicy [6]; (v) IFs depend on the field, such that bigger fields have ournals with higher IFs; (vi) the IF is more a measure of popularity i.e., how many researchers cite papers in the journal), rather than measure of truth, utility, or value to the community [7–9]. Such considerations resulted in the “San Francisco Declaration n Research Assessment,” which promoted 18 recommendations nd concluded, “Do not use journal-based metrics, such as Journal mpact Factors, as a surrogate measure of the quality of individual esearch articles, to assess an individual scientist’s contributions, r in hiring, promotion, or funding decisions” [7]. The Australian edical Research Council similarly stated, “Journal IF is not a sound asis upon which to judge the impact of individual papers,” with unding applications no longer including IFs [9]. However, it has ecome an article of faith that the best science ends up in the best ournals, and the best journals have high IFs. This subjective corelation is likely true in the broadest sense, but is often treated as bjective and near absolute, with remarkably little effort expended