Answer by Tirumalai Kamala:
First, I don’t think that “creating networks of patients who willingly provide anonymised data for the sake of research” is the primary problem. A recent high-profile case well illustrates the point that structural and cultural constructs of the prevailing biomedical research enterprise are the true obstacles. The case?, a California start-up, with the ambitious goal of sequencing the human microbiome. It required participation by citizen scientists who would receive home-test kits to swab their noses, mouths, etc and send to the company for sequencing. Company would sequence and send back the data. All without the oversight of the traditional IRB (Institutional Review Board). Upshot? Bioethicists and other stakeholders of the current biomedical research enterprise went up in arms. Result? Company retreated, and got approval from an independent IRB ( ). Clearly, recent exponential technological change appears to have brought us to a crucial inflection point that could fundamentally change the business of science. How did we get here? Maybe examining some of the structural and cultural constructs of the prevailing biomedical research enterprise will help better identify the true constraints and reveal efficacious solutions.
Let me give a little personal perspective on this matter as a minuscule cog in a giant pharma wheel. About two years back, I was assigned to be part of a small multi-country team instructed to put together a blueprint for a translational research approach within the company’s R and D structure to help accelerate vaccine development.
It’s a familiar refrain. Drugs and vaccines take ever longer to get to market, fewer of them do so, and costs for each are exponentially greater than a few decades earlier. The key here is the ballooning in late stage failures, failures that typically happen in Phase III, after 10 years or more of R and D, and millions of dollars of investment. There is an urgent and severe need for these inevitable failures to occur much earlier, preferably within the first two to three years of a project green light.
Moving on from personal perspective, let’s examine scared cows, i.e. key structural constraints that stand in the way of effectively harnessing technology for medical research, and in creating comprehensive human bio repositories.
1. Prevailing publish-or-perish practice of biomedical science. The prevailing paradigm of publish-or-perish makes it an inherently competitive, winner-take-all, zero-sum game. Yet, science is not a game, especially for the millions dying painfully and senselessly from preventable diseases that currently lack effective treatment. As we routinely acknowledge, most of the low-hanging fruit in terms of human disease has already been harvested, back in the era wheneither ruled the roost wholesale or still had a say. What remains is either currently intractable without large-scale co-operation or unfashionable, in that the preponderance of suffering is by those in the Third World. Needless to say, publish-or-perish presents a formidable structural and cultural obstacle for meaningful and effective co-operation in tackling those human diseases that today lack effective control or treatment.
2. a) Publish-or-perish rewards selfishness. In the US, for example, the scientific landscape underwent a sea change following the Great Recession. Especially for biomedical research, the NIH gravy train came to an abrupt and sudden stop. With it, competitiveness only intensified, as job security evaporated for multitude of scientific staff, universities severely curtailed tenure, and competition for fewer grants increased. As this report () suggests, this has increased the existential angst among researchers’ that they get proper credit for their research product. Ever more so, our grants, careers, and tenures depend on that.
b) Publish-or-perish rewards novelty not replication.To my knowledge, no one has done more pioneering work to spotlight this pernicious and corrosive aspect of the prevailing biomedical research enterprise than, starting with his modestly titled 2005 shot across the bow, “Why Most Published Research Findings Are False“ ( ). If prevailing biomedical research structure constantly goads us to pursue novelty, any surprise we routinely end up with catastrophic late-stage therapeutic failures? After all, we are not rewarded for replication!
c) What are possible reforms to change this paradigm structurally and functionally? Would a change in the reward structure incentivize a move away from competition to co-operation? That certainly seems to be the thinking behind the US-based NSF’s recent directive that, rather than assess a scientist’s impact by their peer-reviewed publications alone, non-traditional research products such as data sets, software, videos could start to become part of the “actionable, accessible and auditable metrics‘ used to assess their productivity (). However, this is merely a small step in the right direction. One government agency’s policy directive alone will not suffice to trigger a structural and cultural change.
d) Creating bio repositories takes time, effort and administrative resources. We scientists are currently not funded and rewarded for such efforts. Funding agencies like the NIH primarily assess research publication metrics. Will they and other funding agencies follow NSF’s lead and change their assessment metrics to allay our existential need for appropriate scientific credit?
3. Dominance of inappropriate animal models. Let’s re-visit‘s famous directive for governmental funding for basic science, Science the Endless Frontier ( ). All well and good in a post-World War II US, flush with money and resources to make that happen. Such explosion in scientific investment propelled a commensurate expansion in scientific personnel. Science became a stable, well-paying career for many for whom it wasn’t a vocation.
In hindsight, so easy to see that swelling ranks of careerist scientists combined with the publish-or-perish edict mandated an expedient animal model for biomedical research. Enter the mouse. Short life-span, relatively hardy, easy to maintain and feed, and courtesy fanciers like Abbie Lathrop (), easily available inbred lines. Match made in careerist heaven. Fatal flaw? Mouse does not simulate human physiology and disease. Yet, here we are stuck with it as the necessary de facto proof-of-principle, in both basic and applied biomedical research. In fact, the entire product pipeline of any number of companies that support biomedical research consists of products for the mouse model and little else. In both basic and applied biomedical research, the mouse model is the first gatekeeper. As a result, resources that over the decades could have been better used to develop human bio repositories, and reagents and tools for human biology research went instead to building the current mouse model biomedical research infrastructure. Inevitable result? We understand a lot of basic biology…of the mouse. Of the human? Not as much.
How do we break free of this fruitless stranglehold? I think the inflection point will come through a combination of money constraint and technological development. In that light, the post-Great Recession era of tight science budgets may well be a boon. Animal colonies are notoriously immensely expensive. Shrinking budgets that preclude them could still support in vitro and in silico human studies. As more biomedical research labs get pushed in that direction, the underlying support infrastructure that provides reagents and tools will hopefully expand to fill these labs’ changing need. Slowly, the stream could expand to become the norm.
4. Industry compensation structures reward progression not truth-seeking. Why do we have so many late stage failures in the first place? Michael Ringel et al recently wrote about this at length, exposing the current reward structures that drive this behavior and the changes necessary for reversal (Ringel, Michael, et al. “Does size matter in R&D productivity? If not, what does?.” Nature Reviews Drug Discovery (2013)).
5. Potentially Fruitful Niche for Individual Private Enterprise. The structural and cultural constraints in current biomedical research create room for the wealthy private individual(s) to play patron and fund their vanity science projects. As I recently learned while researching another A2A, one of the most exciting breakthroughs in more than 30 years of HIV vaccine research () was privately funded. “This remarkable and totally unexpected breakthrough was obtained by an investigator-driven research that was not funded by the usual governmental and large scale organizations that support most of the ongoing HIV vaccine research world-wide. It was sponsored by a private benefactor who funded the project to the tune of 13 million Euros. This illustrates once again that success in basic vaccine research is unpredictable and that “risky” projects based on unorthodox thinking may deserve as much funding as the “safe” projects that are often preferred because they abide by current fashionable paradigms” ( ).
6. Incentivizing human biomarker research. Late stage failures negatively impact every daisy in the biomedical research chain, from the funder, basic researcher to clinician. As more of these accumulated in the last two decades, a push to accurately identify biomarkers, i.e. surrogate endpoints, is taking root. (). Some interesting case studies here ( ). Hopefully, prevailing economic constraints will continue to incentivize such pragmatic approaches.
Thank you for the A2A, Anonymous.