Establishing a Framework for Improving the Quality of Clinical and Translational Research

Abstract

TO THEEDITOR: The discovery of serious problems with research conducted in a laboratory at Duke University 1 has had a profound effect on our approach to translational research. We believe that this experience provides a number of salient, broadly applicable lessons— not just for the field of cancer genomics, but for scientific endeavors and institutions across the spectrum of biomedical research. As part of our initial response, we have developed a set of principles for translational research. 2 This Translational Medicine Quality Framework (TMQF) Principles document reflects the work of a multidisciplinary faculty leadership group and a panel of external experts who represent clinical, laboratory, and genomic research, including quantitative sciences. These principles have been formulated to ensure that research efforts at Duke University are guided by clear and consistent standards for quality. In particular, this framework will enhance the degree to which both the public and the scientific community can be confident that basic and translational research resulting in human applications has been conducted and validated in a manner consistent with current knowledge of the issues, the necessary scope of expertise, an awareness of the potential pitfalls of translational genomic research, and the relevant regulatory requirements. Key points of the TMQF document include creating accountability for research quality and fostering a culture of open discussion and challenge of research beyond the individual laboratory and investigator; guaranteeing robust data provenance and reproducible data management; ensuring adequate quantitative collaboration as a standard in the research environment; and enacting heightened scrutiny of human research that involves conflicts of interest, whether individual or institutional. We are currently finalizing a document that will detail our specific plans for implementing improvements to research quality throughout the institution. As part of the corrective actions taken, we have conducted an intensive review of all publications that resulted from Anil Potti’s work at Duke University (including five that appeared inJournal of Clinical Oncology) 3-7 and for which Potti appears as an author or coauthor. After we excluded opinion pieces and reviews without original data, we were left with 32 articles that included more than 150 coauthors. In each case, we asked the coauthors to attest once again to the conditions of authorship and to collaborate to determine whether they could vouch for the reproducibility of the results. This effort led to the full retraction of 10 articles (nine published;