The cardinal virtue of mathematics rests in its ability to describe objects, events, and relationships unambiguously. This is the meat of objectivity, the heart of truth. But what are we to make of it, when the clarity of the prosecutor’s argument dissolves before the gaze of a surprise witness? How do we account for it, when today’s truth becomes tomorrow’s illusion? And then becomes the truth again the day after?
If numbers are so penetrating, and statistics so reliable, why must trial outcomes be nitpicked at all? Where is there wiggle room left, through which doubt can squeeze, and the fortunes of war – or opinion – turn? And if mistakes are found in the original argument, then how can we be sure mistakes will not be found in the nitpick?
This is the common lot of humanity, that opinions change and dogmas are reversed, and it doesn’t matter, and never has, that the advocates of a dogma emblazon their credo on a flag, cut it in stone, or enact it into the laws of nations. Would anyone expect them to champion someone else’s convictions? Would anyone expect M. Cochrane to applaud the virtues of empirical method, in preference to the placebo control?
The controlled trial controls for bias, and controls for it very well. But that is not enough to justify confidence in the results it produces. To build confidence, in addition to eliminating bias we must guarantee that the trial itself is fair, for example, that the arms of the trial are reasonably well balanced, a trick very adequately achieved by randomization. That there are already many other systematic tweaks available to the researcher, to control for inaccurate measurement in controlled circumstances, tells us the placebo control is not the self-sufficient measure of medical efficacy that it claims to be. It even suggests, obviously I think, that changing the object to be tested demands new modifications adapted to the task.
Yet all the tweaks in the world do not release the controlled trial from the chains of its own limitations, and are not enough to certify its results with confidence: in addition, the controlled trial must prove its mettle to the satisfaction of those it serves. In relation to conventional medicine, in other words, the controlled trial long ago became a so-called “gold standard” in research for the simple and ample reason that its results made sense to the allopathic physicians who were its first, and have since then been its most faithful audience.
In short, the quantitative measurements it returned were consistent with the expectations of the conventional physician, expectations that were based on clinical (empirical) experience. In other words, quantitative research found – as any scientific enterprise must find – independent corroboration in evidence derived from an alternative source: in this instance, empiricism. It is therefore not surprising that, since the beginning, the controlled trial and the allopathic physician have been as closely allied as two peas in a pod.
But this is a remarkable and ironic fact, one that warrants repetition: controlled research gained its first and hitherto only ally in medical practice due to the implicit fact that its own “objective” findings were consistent with the “subjective,” empirically based expectations of conventional physicians! Had the controlled trial failed to produce those results, it is safe to assume that the allopathic community would long since have abandoned the practice of controlled research, and that the self-congratulatory boasting of its adherents would have been nipped in the proverbial bud…
…which raises the amusing question, why those who believe in controlled research accept the subjective testimony of the allopathic physician, but reject the testimony of the homeopath – and audiophile and cranio sacral therapist and herbalist … – on the basis that their opinions are, well, subjective. It appears from this that the disdain some in the research community exhibit toward empirical practice has more to do with whether the results it produces are in agreement with the results produced statistically, than with a principled preference for statistics over observation.
The typical rejoinder of the skeptic, that the allopathic physician’s observations have proven themselves by performing well under “objective” test conditions, represents so self-serving a line of reasoning as to be almost tawdry, or perverse: it assumes the veracity of its own method in order to certify the veracity of a witness who is testifying in its own behalf.
Surely, one would expect that such a transparently manipulative logical edifice would not fool anyone. Sadly, we know too well that it has in fact fooled nearly everyone.
As these considerations suggest, a major problem with specialization – and the placebo controlled trial is certainly a specialized research instrument – is its tendency to lose sight of defining contexts. As with biological inbreeding, lacking the moderating influence of context the self-absorbed intellect loses connection with the wealth of realities that flesh out the local observation. In this connection, a biological population flourishes because all individuals and all local groups share, through genetics, the experience of all other individuals and groups in the broader population: they have the inbuilt knowledge of distant environments that prepares them, when necessary, to adapt to those environments without prior direct experience.
By contrast, the inbred group “knows” (genetically) only the local milieu, and eventually its ability to adapt to distant or variable landscapes vanishes; as a biological entity it gradually but inescapably becomes unviable. In like fashion, the narrowed intellect, measuring one thing and knowing no other, thinks, tragically, that it has fully considered its subject, when in fact it has only fit within its grasp that which its own character permitted.
The more fanatical proponent of controlled research can barely imagine that elimination of bias is a necessary, but not in itself a sufficient condition for achieving a credible experiment. He would have scoffed when it was first suggested that randomization was needed to correct for mistakes in outcomes due to imperfections in the mechanism of the placebo control. In exactly the same way, in our own times he objects to the notion that the idealized format of the controlled trial must be tailored to the particular demands of the subject it seeks to measure; instead, he repeats a favorite nostrum, that if a medicine claims to have real effects in the real world, then it ought to be easy to demonstrate those effects in a controlled trial.
To be sure, this conundrum is not an original invention of the advocates of controlled research, or for that matter of statistical measures generally. Just for example, in the field of mental health practice, behaviorism has been heralded as being more objective than psychodynamic analysis or practice. In part, this is because treatment objectives designed around behaviorism are easy to define, simple to measure: is the patient employed 6 months after treatment commenced? Compare that to this: is the patient happier now than he was then?
One of the more disturbing consequences of this ideological shift was to encourage “stability in community functioning” as a widespread measure of success in mental health practice. Not without justification, it was claimed that the chronic psychiatric patient was made ill by the very system designed to help him: he became “institutionalized.” Then, with the new thrust to achieve “adaptive behaviors” for the psychiatric patient, programming succeeded in its move toward “de-institutionalization,” but with the consequence, as is well known by now, of an enormous growth in homelessness, former in-patients living in the streets of the community.
The fair-minded auditor, I would suggest, will look at these debates and conclude that the complexity and profundity of issues involved deserve a full and balanced review, a review, in short, that is poorly served by a droning reference to banners and buzz words. “Objectivity” is a valued commodity. Truth is more so. And “unbiased” experimental outcomes, that are deeply flawed and badly in need of nitpicking, are not nearly unalloyed enough to trade our future for.
Still, lest this little essay be misconstrued as biased against the placebo control, let me recommend to the homeopathic side, as clearly as possible, an attitude toward controlled research that assumes its complaints against us deserve also to be answered, in painstaking detail and on their own terms. The controlled trial is a legitimate instrument of scientific research, and if there are concerns how homeopathy has fared under its scrutiny, they should, simply, be addressed.
But this is not to say we should spend our days defending homeopathy against its perceived failure under experimental conditions. Indeed, in context of research science rather than medicine, the controlled trial itself has some questions to answer, and we are in good position to put them. After all, it is not sufficient to embrace the approving testimony of one class of empiricists – in this instance, the conventional physician – only to reject empiricism when its testimony is not so flattering.
For this is a debate, and assuming our side also has a right to make its case, then we can suggest that the clinical and popular success of homeopathy itself points to the failure of controlled research to accurately measure real world medical practices. In other words, while it is essential that the homeopathic community address serious questions from the research community, it is well past time to turn the tables, and require of the research scientist a serious and considered response to questions concerning the merits of his own methodology. This is especially true in the wake of investigations of the existing research record, which reveal significant problems with the quality and integrity of its work.
Simply put, the sweeping successes of homeopathic medicine, and its rapidly increasing popularity, cannot be ignored forever. Popularity is not a scientific proof, but it is evidence and must be accounted for. Further, we now have a clear record of recent successes for homeopathy under controlled conditions, and the reports of high level meta-analyses that support homeopathy’s claim to efficacy.
Therefore, since only a liar or a fool would claim his favorite child was perfect, the experimentalist must be asked, finally, to give an open and honest account of the weak points in his method. We know from history that the controlled trial is incapable of measuring even conventional medical practice, without introducing numerous technical modifications such as randomization or the crossover design. Therefore, it is easy to understand that, when trying to extend the reach of controlled research to fields such as homeopathy, audio performance, herbal medicine, and the like, it is the most natural seeming expectation that additional technical innovations will be required. The sincerity, or talent, of the research scientist who hedges in response to these expectations must surely be looked upon with suspicion.
In any case, it appears this challenge is already being met by at least some in the research community. This is reflected in the fact that we are already seeing a clear trend toward well-designed trials that present positive results for homeopathy. As that trend continues, it will become more imperative for the side of controlled research to explain itself, and to explain its persistence, for the past 200 years, in believing it had demonstrated facts that in fact turned out to be illusions.
Perhaps a dollop of humility will be the result, and perhaps that will help to spare others from becoming the next target of the statisticians. Perhaps it will even affect the parliamentarians in our midst, who might reconsider the wisdom of deciding how people should live their lives, on the basis of passing fashions in science. And perhaps the public at large will learn an important lesson, that the pronouncements of a self-appointed monarch – golden or otherwise – should not be taken at face value.
That having been said, may I wish everyone a happy and healthy new year —