Science on the Potomac
Climate modelers funded by government paymasters have long been engaged in a process of burying key details in carefully maintained fog.
President Obama campaigned on a promise to bring sound science back to Washington. With health and climate, he’s instead leading the march toward antiscience bought and certified on the banks of the Potomac and codified in regulatory edicts that permeate every corner of daily life.
Science, the Obama administration tells us, will make medicine better and cheaper when “evidence based” standards instruct doctors how to prescribe and treat. Potomac science will inevitably be based on statistical studies of how, for example, the right drug did or didn’t keep the average body in a 2,000-patient study alive for five years longer. But the statistics of large crowds miss something: all the biochemical differences that distinguish and disassemble patients and diseases. Cutting-edge medicine works with molecules, not crowds.
Medicine’s first great revolutionary leap into the realm of the tiny was launched by the development of germ science in the late 19th century; in the late 20th medicine added systematic tools that now allow it to read and tame genes, proteins, lipids and other shards of troublesome human chemistry. Almost daily the discovery of some new biochemical fragment moves medical science another step away from the statistical mob and toward mechanistic medicine anchored in factors that vary widely from patient to patient.
The germ’s or gene’s or drug’s capacity to manipulate specific aspects of a patient’s chemistry is also a fact that science can establish with quite high certainty. The science gets progressively softer from there on up and is terminally squishy by the time it gets to Washington. Statistics are the last-resort scientific tools, lumping together what better science takes apart.
Peering at screens, not patients, Potomac statisticians can lose differences that might stand in the way of one-size-fits-all medicine. Or they can magnify differences by selecting just a few remote effects that tilt the scales in the desired direction--useful when trying to establish, say, that an expensive new drug is worse than the old, cheap generic. To bury a newer arthritis drug they might count the rare but easily flagged heart attacks in the arthritis-drug database while failing to consider how the drug might commonly relieve a bit more pain, thus permitting a bit more exercise that leads to a smaller risk of heart attack.
Climate modelers funded by and accountable to government paymasters have long been engaged in a very similar process of burying key details in carefully maintained fog. They set out to model something far too complicated for their computers and their own (modest) scientific talents. They discovered to their dismay that however hard they labored to account for everything that might matter, they couldn’t even predict the past, like the medieval warming period or the different temperature trajectories of first and second halves of the 20th century. Still less could they predict the future. The predictions of the early models diverged widely as they whirled off into the future. None of the 1997 models predicted the temperature plateau that materialized as the ink was drying on the Kyoto Protocol.
So to make things come out right, the modelers began losing, smoothing and fudging the raw temperature data, and inserting dials into their models that could be twiddled at will. The dials adjust how the models treat all the minutiae that certainly shape climate, but in ways too complex to be anchored in any hard data at all--water vapor, dust, aerosols, plankton, soil bacteria, the rate of mixing of the top few inches of the ocean and so on. Dimly aware of what better scientists would say about all this, the modelers strained to conceal the details of their work.
Potomac medicine will be scientific when its databases include every last piece of patient-specific biochemical data that might affect the progress of a disease or its antidotes. But Potomac science can’t possibly keep up with the torrents of information now flowing out of human bodies. Smart doctors and patients know it, and they’ll fight to the death to dodge or dismantle Washington’s edicts.
Potomac climate medicine will be scientific when every last data point and every last line of code are posted on the Web. When the data and code come out into the open, as they inevitably will, better scientists will quickly demonstrate how very modest changes in the smoothing, fudging and twiddling can jerk the predictions up or down, or turn them upside down.
This piece originally appeared in Forbes
This piece originally appeared in Forbes