“What gets us into trouble is not what we don't know. It's what we know for sure that just ain't so.” Mark Twain
I am sometimes asked what my biggest surprise has been as director of the National Center for Advancing Translational Sciences (NCATS), the component of the US NIH focused on the science of translation. Without hesitation, I answer “apperception” — meaning ‘to comprehend a new idea by assimilation with the sum of one's previous knowledge and experience’. Translational science is a new idea1 and turns out to be challenging to assimilate with what is believed to be true by many stakeholders based on their previous experience. This difficulty with apperception is surprisingly resistant to data and substantially impedes progress in translational science that would otherwise be accelerating the development of new treatments for disease — a goal all stakeholders share.
I offer here several of the most prevalent and robust misconceptions I have experienced in speaking to diverse audiences about translation, along with some ideas on causation and amelioration. I present these as one trained in, and unflinchingly supportive of, basic scientific investigation. And with apologies, I present each misconception in its purest form without subtlety or qualifications, for the purposes of exposition.
Common misconceptions about translation
Translation does not exist. This is a surprisingly widely and firmly held belief, with multiple variants, including the ‘basic discovery ≡ intervention ≡ health improvement’ false equivalency. This seems to derive in part from well-placed respect for the wonders of fundamental science, along with widespread lack of knowledge of how interventions (drugs, devices, behavioural interventions and medical procedures) are developed, separation of the stages of translation into dozens of ‘silos’ and a paucity of individuals who have first-hand experience in more than a few of these silos.
The result is that a fundamental discovery is commonly believed to be tantamount to a marketed intervention. Furthermore, the widely used term ‘technology transfer’ encourages this notion, as it implies that a basic technology can be transferred to a commercial organization and immediately marketed without substantial further investment or risk of failure. To those with this simplistic world view, the concept of translation as a discrete stage between fundamental discovery and health improvement is confusing at best and a fiction at worst.
Translation is a ‘thermodynamically favoured’ process. The concept here is that translation does exist, but it occurs naturally and automatically. This is also known as the ‘snowpack in the mountains’ model, in which fundamental science is the metaphorical winter snow that, in the spring, naturally melts and runs downhill and provides the needed product, an intervention that improves human health.
One source of this misconception appears to be the common statements at the end of scientific and lay publications about fundamental discoveries that note the potential for the discovery to lead to a drug or other therapeutic intervention in a few years. This must be a misconception given that the remarkable success of basic science in delivering fundamental insights in the past several decades has not led to the number of drugs or other interventions being approved increasing to a corresponding extent.
Translation is straightforward and does not qualify as science. Those who do not ascribe to the first two misconceptions may still hold firmly that translation is simple, rote (‘crank-turning’), well understood, and science- and creativity-free. For example, I was once introduced by a well-meaning distinguished professor as one whose work “entails all the creativity of running a Toyota factory”. In this misconception, the failure so commonly observed in the process of developing and deploying an intervention to improve health is solely the result of the poor scientific quality of translational research and its practitioners, particularly (although not exclusively) in the biopharmaceutical industry.
The principal source of this misconception appears to be the aforementioned siloing of training and experience in the research ecosystem. Consistent with this hypothesis is the common experience of those, including this writer, who rapidly and with some embarrassment reverse their position when they move into translational research, particularly in industry, with its technologies, disciplines, focus and research ethos that are so distinct from that in the academic institutions they came from.
This misconception also stems from the fact that translation unavoidably deals with humans rather than a reductionist or model system. So much of science, and so much progress in it, has been produced through the study of biomolecules, cells and inbred model organisms such as yeast, flies and mice that many scientists sometimes lose sight of their function as admittedly simple ‘models’ of the intact, outbred human. Translation is additionally difficult because its focus is not humans who are healthy, but those with a disease, with variability in the disease and the therapeutic response added to variability in humans resulting in exponentially greater complexity. But because of the complexities of human experimentation that do not exist when dealing with model organisms, the science is often unavoidably less rigorous. This has led to more academic prestige accruing to those who rigorously study a simple system, rather than somewhat less rigorously study the human system that is the ultimate application for much of science and the rationale for most of its support by the public. A parallel development to the ascendancy of this misconception is the relative devaluation of clinical investigation, clinical observations and patient reports in the academic science field.
This said, translational researchers have contributed to this misconception by the words and schematics we use to describe what we do. For example, drug development is universally referred to as a ‘pipeline’ — a term that connotes a passive process in which the starting material exits the conduit unchanged with 100% success and progress is dependent only on the diameter of the pipe, which is clearly an inaccurate description of the long, multifaceted and failure-prone process of drug development. Visually, the translational process is commonly illustrated as a succession of chevrons of equal length and width, a similarly inaccurate and misleading schema. An alternative has recently been published2, which, if widely adopted, will help address this misconception.
Translation is a unidirectional process. The widespread misconception that the translational process proceeds from basic research to clinical medical practice to impact on public health has only been possible in the past 60 years, as reductionist science focused on particular genes, proteins, pharmacological modulators and cells has become feasible. For the thousands of years of translation before this, the process necessarily started with clinical observations in populations or individuals, which led to the study of therapeutic interventions (such as opium for pain or willow bark for fever) or fundamental insights (for example, the identification of the causative organism in the 1849 London cholera epidemic).
Once an investigational therapy gets into humans for the first time, regulatory approval and marketing are all but assured. This misconception primarily benefits stakeholders with phase I assets who wish to sell them or raise capital to support the very expensive and risky remaining stages of clinical testing and regulatory approval. In fact, data consistently indicate that 80–90% of interventions that progress to testing in humans for the first time are never approved3.
Regulatory approval is the end of the translational process. One of the primary problems with a unified vision for translation is that each participant has a specific view of the outcome of translation, on two axes: technical specialty and stage of translation. This is why NCATS’ definition of translation is intentionally holistic, with a clear end of the process being improvement in ‘the health of individuals and the public’. In other words, if translation is concerned with transforming the potential of discoveries into the reality of impacts on health, the translational process must continue until it reaches a point at which such impacts (or lack of them) happen, which is in the community, not in an animal model or a clinical trial. Our job as translational scientists is not complete until the intervention has reached, and been shown to be useful in, improving the health of individuals. This then reaches the domain of implementation science, community engagement research and population health — the challenges of which are as large, if not greater, than the more familiar earlier stages of translation.
My purpose in pointing out these misconceptions is not to diminish the intent or accomplishments of many dedicated and skilled scientists and physicians. I myself once believed unquestioningly in all of them. Rather, it is to help further understanding among researchers of different backgrounds, advance the ongoing discussion on the reasons for and solutions to the translational gap between basic discovery and human health advances, and to help avoid unproductive or even counterproductive scientific and policy decisions. Only by advancing our common understanding of the complexity of translation, translational research and translational science1,2 will translational gaps be narrowed and ultimately eliminated.
Competing Financial Interests
The author declares no competing interests.