Disruptive science is leaping forward, not limping along

A recent study and accompanying news story in the preeminent journal Nature provocatively concludes that disruptive innovation in science has dramatically and mysteriously declined 90% since 1945.

The study has prompted a wave of news coverage and tweets decrying the apparent languishing of modern science. We feel that the authors make interesting observations on publishing trends, but their conclusions seem to be quite disconnected from the valuable and transformative innovations that benefit humanity.

The authors reached their troubling conclusion via an enhanced form of citation analysis that tracks the extent to which previous studies are mentioned in new papers. The method is intended to separate truly novel ideas from incremental advances. Scrutinizing nearly 50 million papers and patents from 1945 through 2010, the authors report a trend away from what they consider real breakthroughs in favor of more routine progress.

If this assessment is fair, it should worry scientists and non-scientists alike. Scientific innovation is, after all, the critical engine that makes it possible to tackle the tremendous challenges humans must overcome to survive, such as pandemics, global food insecurity, and climate change. If innovation is steadily declining, it would be cause for serious global concern.

To be sure, the authors’ new method for analyzing scientific work is helpful, insofar as it points to certain systemic flaws in scientific funding and publishing. Grants, for example, often err on the side of safe bets, resulting in published research that only marginally advances existing knowledge. We applaud its conclusions to encourage research quality, sabbaticals, riskier and longer awards for individual careers over specific projects, and the gift of time for researchers to step outside the fray and read widely.

The negative overall trend may partially stem from an increasing denominator of “non-disruptive papers,” given that more than 3,000 papers are now published per day — a staggering increase since the 1940s — driving down the “disruptiveness index.” Yet total innovation and impact may still be increasing. Furthermore, the paper’s analysis also stops at 2010, missing the last crucial decade of rapid progress.

Here’s the big question: Does this analysis really represent a decline in radical innovation that improves fundamental understandings and technologies?

The example of the “most disruptive” innovation in the last half century the authors cite is a 1983 patent on a better way to get DNA into cells. But moving DNA into cells was not at all surprising in 1983 with famous precedents in 1944 and 1978. It was soon largely replaced by electroporation, cationic lipids, viral capsids, and microinjection. Meanwhile, the “least disruptive” innovation in the Nature analysis is David Baltimore’s 1970 paper on reverse transcriptase, an enzyme that catalyzes the formation of DNA from an RNA template. At the time, going from RNA to DNA was conceptually very surprising and the paper, as a bonus, provided a way to study and commercialize RNA that has not been replaced. That low-disruption paper earned its author a Nobel Prize, while the supposedly bolder example did not.

Consider the incremental advances of today, like the mRNA-based Covid-19 vaccines. The authors of the Nature paper view these vaccines not as a genuine breakthrough but as the result of decades of stepwise research, from the discovery of mRNA itself to viral genome mapping and to the development of the lipid nanoparticles that stabilize the fragile molecule for delivery into the human body. But given that these vaccines saved an estimated 20 million lives in a year, we believe it’s fair to call their arrival in record time truly groundbreaking. These vaccines literally represent a break from the vaccines of the past, a new paradigm upon which vaccine developers and scientists will continue to build.

In the near future, what if new categories of cancer therapy (building on the development of CAR-T immunotherapy), break-even nuclear fusion reactions, exponential improvements in reading and editing genomes, machine learning solving the six-decade old protein design challenges, the colonization of planets, carbon sequestration, nitrogen fixation, and aging reversal therapies emerge? One could imagine that none would nudge the disruptiveness index upward because long-standing, profound problems are defined by multiple mature disciplines. That said, these achievements would rank among the most amazing achievements of humanity!

Gunther Stent declared “the end of progress” in 1969, decades before the life science revolution got underway. Similar sentiments about the end of physics breakthroughs immediately preceded quantum and relativity revolutions and many decades of breathtaking applications. Perhaps it’s time to embrace the spirit of anti-complacency and pro-innovation without hallucinating a depressing decline in the face of remarkable progress.

A more relevant metric to gauge the success of the scientific enterprise is the extent to which advances — whether incremental or totally novel — translate into real human impact. Imagine a world where everyone lives longer, healthier, and happier lives, in harmony with the planet and with each other. Now that would be truly radical.

Juergen Eckhardt is a senior vice president and head of Leaps by Bayer, the impact investment unit of Bayer AG. George Church is a professor of genetics and of health sciences and technology at Harvard Medical School and the Massachusetts Institute of Technology.

First Opinion newsletter: If you enjoy reading opinion and perspective essays, get a roundup of each week’s First Opinions delivered to your inbox every Sunday. Sign up here.