As an undergraduate in the 1960s, I was encouraged to read The Two Cultures and the Scientific Revolution by Charles Percy Snow (“CP” to friends), a British physical chemist and novelist (1905-1980). The book was based on a lecture given at Cambridge in 1959. For context, this was two years after Sputnik and panic over the relative strengths of socialist technocracies vs. traditional democratic debating societies. While the circumstances in the United Kingdom at the time properly disturbed Snow, the title gained more importance than the book. We do like things that come in twos—a binary choice is always convenient for making an argument. I will steal that notion here.
The fact that “humanists” and “scientists/engineers” had little to say to each other disturbed Snow. The presumed clash of cultures was an active topic for debate in undergraduate seminars. I’ve not forgotten it. This was a time when entire college freshman classes were expected to read a single book and debate its implications after arriving on campus, anxious about leaving home. The “two cultures” debate became entwined with the Cold War and the marvelous freeing of government largess to encourage we science majors to keep going. English and history majors are still complaining.
I am retiring from academic life this summer after 47 years. I use this as an excuse to briefly review a more recent cultural debate within the sciences alone. I am reducing the complexity of academic and industrial science to sharply drawn choices, each of which has taken 50 percent of my time. Reality is more nuanced.
Creativity
In academic science and engineering, creativity is the dominant research theme. It is what we expect of Ph.D. students and faculty. Something new is our goal, and getting it published is how we reflect success. Early in my career I worried about not being able to achieve this. Much to my delight, I found it relatively easy to do new things in chemistry. The harder objective was to discover something important, something that really mattered.
Five decades later, it is both easier to be novel and harder to have it matter. We have far more channels to disseminate what is new. Measurement tools enable data that CP likely could not have imagined 60 years ago. The number of scientists and engineers exploded with the cold war, and pressure for both funding and notoriety has expanded unsustainably ever since. IT tools do now enable us to find what others have achieved and they can find our gold nuggets.
On the other hand, publications are a very poor source to reveal what matters because most of what matters is not ever published. Instead, it is purchased. The criticism that academic papers are often not reproducible is fair only relative to the puffed-up claims for it. In this culture, weak validation is expected. Through an iterative process, the good stuff rises. The top academic labs slow down a bit and get closer to the truth when they have the resources to do so. This is rare. Validation takes a larger team and a longer time.
Consistency
In business, we value consistency, because the customers expect it and the regulators demand that we prove it. This seems obvious, but bad behavior has enforced the evolution of GxP and ISO quality assurance formalities as technologies evolve.
While there were earlier grumblings, the 1970s really set these formalities in motion. We are not looking back. Consistency encourages bureaucracy and discourages creativity. We can’t properly change methods in the middle of a funded multiyear study. Adaptive clinical trials suggest that contrary possibilities are not impossible.
Over decades, commercial firms at scale have largely abandoned discovery research, deferring to academics and small boutique firms with fewer bureaucratic overheads, and often not restricted to reporting earnings every 90 days. Losses for young firms suggests investing, belief in the cause. Beyond wishful thinking, the cause must be supported by data and the data must be trustworthy. The Theranos tragedy supports this comment. Thus, we have the need for validation and we place much of that requirement in contract research/manufacturing firms subject to audit by both clients and regulators. It’s working. Publications in this environment are not a priority.
Conclusion
With R&D, there is never a clear boundary where R ends, and D begins. Decades ago, pharma might internally “throw the project over the wall to development,” implying that the data “now really count” and must be subject to the QA formalities. Someone is now watching, perhaps from an “agency.”
What goes on in academic labs is not as bad as the recent claims often made for it. A lot of what is published has the purpose of educating students and demonstrating possibilities far from statistical validation. Hospitals and bioharma may favor “Six Sigma” while students favor graduation and faculty favor more grants.
Recently, I’ve been asked to review several manuscripts on well-validated bioanalytical methods for drugs dosed in mice. Given that these oncology drugs were already approved for some human indications and in clinical trials for others, I failed to see the data having much value. On the other hand, some newly minted Ph.D.s will now have a good feel for what we expect in commercial science. That is priceless. Most will rarely publish again.
In our May issue, Randy Willis referred to the Twitter feed @justsaysinmice that reminds us that potential cures for human ailments are often decades away from early discovery in cell culture or animal models. Those crafting press releases tend to ignore that reality.
Your assignment for this summer is to consider your role on the creativity-consistency axis. If you are on the right side, you will deliver trustworthy results on time, but may get cynical about what’s new. If you are on the left, you are likely to rush to conclusions and to print while finding rigor to be boring.
There is a lot of gray space to nudge into for a more balanced and satisfying career. You can stay up to date on the science and be consistent. As you crawl into your sleeping bag at camp this summer, think how many drugs we would have if pharma was held to the development standards for the Boeing 737. None. Many more people die in a week from acetaminophen and in a day from opioids. Just saying. Get good data for creative projects. It matters.
Peter T. Kissinger (who can be reached at kissinger@ddn-news.com) is professor of chemistry at Purdue University, chairman emeritus of BASi and a director of Phlebotics and Prosolia.
