FDA Matters Blog
Register to get regular updates from FDA Matters FDA Matters Home

Scientific Reductionism and the End of Medicine

“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.

Reductionism is an attempt or tendency to explain a complex set of facts, entities, phenomena or structures by another, simpler set of constituent parts. Historically, scientific reductionism has held that all biology can be explained in terms of chemical reactions. In turn, these chemical reactions can be explained at the atomic level by physics.

An example of scientific reductionism is the belief that a blueprint for understanding and curing all disease will result from mapping genomes (human, bacteria, etc.). In effect, the complexity of biology ultimately yields to a much simpler paradigm based on de-coding the meaning of each component in the human genome and then delivering medical therapy personalized to the individual’s genetic make-up. To oversimplify a little, biology then becomes a predictable “machine,” subject only to additional reductions that yield smaller pieces and even more insight.

In contrast, many scientists believe that the complexity found in biology is more than just the inability of scientists to simplify the tangle of life and disease. No matter how much we know about genomic causation and associations, we will never have a full picture of life nor unlock the secrets to all diseases. These scientists believe that life is more than the sum of its parts. To them, reductionism is not wrong; it just produces an incomplete vision of biology because it cannot account for systems effects. A new field, systems biology, is trying to develop ways to understand the complex, irreducible biological qualities of life.

FDA Matters views scientific reductionism as a source of actionable knowledge. But just as the book, “The End of History,” was more provocative than predictive, there is no “end of medicine” where human biology is reduced to the point of near-total knowledge and flawless cures.

Thus, personalized medicine will not defeat biological complexity. Further, the reductive process will incorporate knowledge from the human genome, but then take us past it into even more difficult and unpredictable challenges to understanding biology and curing disease.

Meantime, public policy is being shaped by the belief that biology and medicine will eventually yield answers that are concrete and totally reliable. But even when personalized medicine provides better targeted therapies, there will still be phase II and phase III clinical trials that inexplicably fail to show patient benefit. After approval, even the most well-documented and logical therapies may prove harmful and require modification or recall. FDA will need to constantly manage the expectations of Congress, the media and the general public to be sure that they understand that no amount of knowledge or evidence renders a medical therapy certain or riskless.

One prominent futurist has said: if we can just live long enough, progress in medicine will allow us to live forever. I say: not so.

The nature of medicine will be quite different 20 years from now, but unpredictability will still be common. As we develop vast amounts of new biological and medical information, old uncertainties about diseases and drug development will be resolved. New uncertainties will emerge.


One Response to “Scientific Reductionism and the End of Medicine”

  1. Ray Perkins says:

    Hi Steven,

    Biological research practices are at least reductionism-squared, perhaps reductionism-cubed.

    Genetics: If we were broad beans or onions, the practice of genetic reductionism in biology might work better. The one-gene, one-protein paradigm is almost OK for some things but not for others. For humans the relationship is something like 80:1 from an estimated 2 million distinct proteins but only 25,000 genes. Even throwing in a multiplier for mRNA still means that we miss 80-90% of bio-activity by only interrogating differences in genes. Genetics – missing post-translational modifications and diversity arising from protein degradation events – is, then, reductionism to the first power.

    Simple System Assays. The attraction of simple systems (e.g. a couple of proteins) is obvious and their use is a necessary starting point. Conditions can be varied in a controlled fashion and response variations noted and relatively readily rationalized. The risk is that simple-system behavior may or may not exist in more complex environments, and the overwhelming majority of simple-system test methods do NOT accommodate variations in complexity (indeed, many simple-system assays use only bits of proteins, etc., potentially amplifying the problem). Simple system assays are reductionism to the second power.

    New-Making Assays. The solution would seem to work with more complex systems. The question becomes, “How?” In many bodies of work, a response is anticipated – say an increase in some protein level – and tools are devised to detect only that protein. Popular choices for tools include custom antibodies and/or fusion to fixed surfaces. Such an approach can, obviously, only provide information on a single entity among thousands. Second, the detection tools are themselves biochemical actors, shifting native equilibria in the direction of the biochemistry induced by the tool. Reductionism to the third power.

    There are other examples of problematic reductionism in the ways we characterize biology – and, therefore, medicine and medications. At a deeper level, reductionism can only work if well defined principles underpin a discipline. Where are the biological principles that could, for example, extend low-probability genetic, simple system or news-making assays into a more general case? Are there equations that anticipate neural response in the shoulder to a minute change in pH in the foot? No. Biology appears to be wholly empirical, with consilience on both chemistry and physics.

    Too bad we’re not a bean or an onion. But then, an onion probably couldn’t reach the keyboard nor could a bean change a pipette tip.

Leave a Reply

You must be logged in to post a comment.

© 2009-2012 by HPS Group. All rights reserved. Permission is hereby granted to those wishing to quote or reprint from this site, providing it is properly attributed to FDA Matters: The Grossman FDA Report™.