It has been nearly a year since the National Institutes of Health (NIH) made public its Roadmap, an attempt to define the institutes' priorities to accelerate biomedical progress. One of the main themes of the Roadmap is the re-engineering of the clinical research enterprise, placing a special emphasis on the need to promote translational research. Although the importance of translational research is now generally recognized, there is widespread disagreement on what the term actually means.
Several NIH institutes have embraced a precise definition of translational research: “the process of applying ideas, insights and discoveries generated through basic scientific inquiry to the treatment or prevention of human disease.” But is this definition what the biomedical community has in mind when preparing a grant proposal or writing a paper? Earlier this year, we conducted two separate surveys to answer this question and to understand what researchers considered exemplary cases of translational research.
Approximately 44 percent of the more than 3,000 Nature Medicine readers who completed our survey subscribe to a definition similar to that of the NIH. But a significant 30 percent said that only those studies with direct clinical application should be regarded as translational research. Even though the NIH has been criticized for being overly pragmatic in their support of translational research, a large fraction of scientists seem to overlook the fact that the NIH definition refers to the process of translating basic insights into therapies and not to the clinical applications per se. In other words, taking the NIH definition in a broad sense, studies on animal models of disease are translational, provided they are relevant to the human condition and allow us to make specific predictions about diseases in patients.
Our surveys also revealed the fractions of people who hold radically divergent views on the meaning of translational research. At one end of the spectrum, 10 percent of respondents had never come across the term. At the other, less than 1 percent of scientists held a more elaborate view of translational research, incorporating an aspect that is missing from the NIH definition and that we regard as particularly important—the bidirectional interplay between the laboratory and the clinic. For this small fraction of researchers, translational research goes full circle from bench to bedside and back.
Indeed, the successful translation of basic insights into medical practice does not depend only on the contribution of basic science to the clinic, as is commonly assumed. It also depends on the ability of researchers to use the results of well-designed clinical trials to refine the usefulness of the experimental models in which advances are originally made. If the success of translational research depends on this bidirectional communication, it is worrisome that more than 99 percent of the scientists we surveyed might not be conscious of its importance.
Perhaps this lack of reciprocal communication helps explain the results of a study by Contopoulos-Ioannidis et al., published last year in the American Journal of Medicine. The authors quantified the “translation rate” of “highly promising” basic research into clinical application. Of 101 studies that, according to the authors of the original papers, had clear therapeutic or diagnostic potential and were published between 1979 and 1983 in high-profile journals, 27 led to a randomized clinical trial and only five gave rise to a licensed clinical application.
There are many possible explanations for the relatively low translation rate. But one idea, advanced by one of the authors of the analysis, is that studies in model systems often fail to mimic the complexity of diseases and the difficulties in dealing with heterogeneous groups of patients. Alas, this complexity has led many researchers to advocate the use of a different yardstick for the evaluation of translational research. A majority of the scientists we surveyed feel that, because research in humans is fraught with difficulties, the criteria for its evaluation should be more sympathetic than those used to evaluate basic research. Accordingly, they argue that editors and referees should be more lenient with papers reporting human data, lest journals prevent the advance of the translational process.
By contrast, a small fraction of researchers do not see the need for a double standard, arguing that the best translational research needs to be as good as the best basic science. One corollary of this assertion is that high-quality translational studies are remarkably limited in number. Still, these studies find homes that are as highly visible as those of the top basic studies. According to our surveys, the best examples of translational research are published in journals such as Science, Nature, The New England Journal of Medicine and Nature Medicine.
We are fully committed to continue being a vehicle for the publication of studies that apply basic scientific discoveries to the treatment and prevention of human disease, and welcome submissions from researchers engaged in translational research. At the same time, we subscribe fully to the idea that the drive for translational research should not take place at the expense of excellence. We are therefore firmly committed to maintaining the high standards that have characterized Nature Medicine over the past decade. If translational research is the new face of biomedicine, quality is the one thing that must not be lost in translation.
About this article
Cite this article
Lost in clinical translation. Nat Med 10, 879 (2004). https://doi.org/10.1038/nm0904-879
This article is cited by
Journal of Behavioral Medicine (2019)