In most discussions of science and medicine, there is an implicit assumption that the human body is a machine—complex and biological, but still a machine. If we could only understand all the mechanisms, processes and parts of that machine, then we could prevent and cure disease. Yet, the further we travel into the biology of life, the more complexity we find and the less certainty and predictability.
“The human body as a machine” is a metaphor, not a fact. Once we accept this, FDA Matters believes we can become liberated from unrealistic expectations about medical discovery and FDA’s role as a gatekeeper for new products that benefit patients.
Medical progress is the direct result of basic and clinical research conducted according to rules of scientific evidence and proof that have been refined over several decades. Our knowledge grows and our methods of discovery improve through our nation’s support for medical research at NIH and regulatory science at FDA, as well as billions invested by industry.
Even still, we can’t consistently or logically explain why good scientific hypotheses tested in well-controlled trials often produce unimpressive clinical outcomes. Or why some medical products are clearly helpful for some patients in a clinical trial, but of no benefit to other patients who appear indistinguishable in baseline characteristics.
Researchers, medical products companies and FDA would love to know the answers to these questions. Until we do, we are powerless to explain, predict or prevent the persistent and expensive discovery failures that are frustrating patients (and investors) who are desperate for medical successes.
We imagine the failures come from the imperfection of our knowledge. This is certainly true, but not a complete answer.
For several hundred years, physicists assumed the world was a machine, a so-called clock-work universe. This was logical and worked remarkably well to explain and predict physical phenomena. Then quantum mechanics demonstrated unpredictable, random, counterintuitive outcomes in the physical world. Ultimately, quantum mechanics increased our knowledge of the physical world, but only after its uncertainty and inconstancy became part of the calculations.
Similarly, there is no clock-work biology just waiting for us to discover all the mechanisms, processes and parts. Some complex biological responses may prove to be “explainable” only through unpredictable, random, counterintuitive activity.*
This may be true of individual responses, as well. If each of us has a unique fingerprint and a unique personality, why should it be hard to imagine that we each have our own biology that can never be fully defined or predicted, even if we could identify all the unique machine-like (e.g. genetic) elements?
If the metaphor (man as machine) is incomplete–and some degree of biological response is inherently unpredictable, uncertain, unknowable, as well as individualized—then FDA’s current standards of evidence and proof are not optimum for stimulating medical discovery. Many products that would benefit patients may not reach the market.
FDA Matters believes the FDA understands this is a problem, at least at the senior management level. However, the agency currently lacks the culture or impetus to incorporate complexity, uncertainty and unpredictability into its approaches to approving medical products. Admittedly, it also operates in an environment where it is gently praised for approvals when they occur….and harshly vilified later if an unresolved or unexpected issue (i.e. uncertainty) is a source of clinical problems.
Patients, physicians, researchers and medical products companies want an environment where a higher level of patient benefit can be achieved, even at the cost of a slightly larger degree of uncertainty and unpredictability. FDA needs to accept and act on this challenge or face rising discord with those stakeholders.
Steven
* Note: there is some work specifically on how quantum mechanics may explain certain biological processes, such as photosynthesis, but the analogy here is only that random, unpredictable biological responses play a very large confounding role in biological discovery, even greater than commonly attributed. See also: “Biology faces a quantum leap into the incomprehensible” http://www.guardian.co.uk/science/blog/2010/nov/12/biology-quantum-leap
Earlier columns from FDA Matters that are related:
Post-Market Safety: Getting the Most Out of Inferences That Aren’t Proofs June 21st, 2011
In the FDA-regulated world, success is often defined as approval of a new product or indication based on two, well-controlled clinical trials. However, the scrutiny doesn’t end there. FDA’s mission includes determining whether already-approved drugs perform safely and effectively when used by large numbers of patients in routine medical practice.
To understand what happens under these “real world conditions,” FDA has expanded its post-market efforts, including development of a monitoring system (called Sentinel) that will be able to track drug usage and medical history information on tens of millions of patients. Although such information will be useful, it can only provide post-hoc inferences, not proof of causation. Even with this limitation, FDA Matters thinks developing the system is worthwhile and may provide multiple benefits. Read the rest of this entry
Scientific Reductionism and the End of Medicine December 27th, 2009
“For the last 400 years, science has advanced by reductionism … The idea is that you could understand the world, all of nature, by examining smaller and smaller pieces of it. When assembled, the small pieces would explain the whole.” (John Holland)
Have you ever heard someone accused of “reductionist thinking?” You probably will in 2010 because scientific reductionism is a critical, but rarely articulated, foundation of personalized medicine. Read the rest of this entry
A reader writes:
Steve—here was a dose of reality from my most recent participation in the annual BIO conference held in Washington two weeks ago: the number of startup and growth device and related firms that are now NOT starting in the US but going abroad as the conditions and environment are perceived to be much better, stable, and less regulatory burdened with delays and bureaucracy. This was not a handful – this was over 55 companies I heard from or directly about. In turn, in visits with over 30 different Country Representatives, the same question emerged over and over…” what is going on with your country Richard? Everyone from America is talking down about innovation, debt, economic recovery…” This is not good news at all about our future. Posted by Richard Seline
Hi Steven,
A little gem that I came across a few months ago has the wonderful title of “Can a Biologist Fix a Radio?” ( http://www.zephoria.org/thoughts/archives/2005/02/27/can_a_biologist_fix_a_radio.html). In it the author makes several cogent observations and arguments, one being a call for standardization in reporting biological/biochemical data. It almost seems absurd that such standardization does not exist, but the threads of the original argument is alive and well. Another point, in keeping with the title, is that our approach to biological research would be a positive hindrance in resolving the function of the various components and how they work together to make music. Part of the difficulty arises from a reductionist approach with two origins: 1. The design of experiments in accord with prevailing models (particularly those in vogue for drug development) and 2. Assays that either consider component activity in isolation OR assays that, themselves, alter performance. Not a good basis for tackling bio-messiness.
Regardless of whether we deal with relatively simple systems such as a single protein (e.g. structure determination) or “systems” biology, we all engage in some level of reductionism: The former in its isolation of the protein from all potential influences in its native environment ,and the latter in its piecemeal (sometimes news-making) assessment of activity. The challenge, indeed the necessity, is recognition of that reductionism and the degree to which it qualifies our results. This is a tall order, even assuming that we can divorce ourselves from model-driven experimental design. For example, methods to assess enzyme activity in simple systems do not translate into comparable assessment in intact cells. How, then do we compare the two sets of results? With current technology, such comparisons are difficult if not impossible. Only assays that function in both simple and complex systems permit collection of baseline data and deviations from baseline performance induced by other entities. We’ve begun to work with just such assays.
This is especially crucial in testing drugs, an area in which reductionism is spelled with several capital “R’s.” Most would-be drugs are screened in very simple systems, often including only drug/target or drug/target-fragment as components, and these simple-system assays seldom function in more complex systems.
Biology is not a system of on/off switches operating in isolation of their surroundings. Rather it’s necessary to imagine thousands (millions, billions?) of inter-molecular equilibria and intra-molecular equilibria. Those equilibria shift in response to both local and remote events. Thus one bio-actor may affect multiple events, and multiple events may affect a single actor.
A machine? Machines make sense – we’re talking biology here. It only has to work!