Is double blind drug testing really a ‘gold’ standard?

Most adults are aware of the placebo effect, where the belief in some medical benefit can enhance that effect. But few know that there are twists to the placebo effect that have a significant impact on drug testing. It is the existing ‘gold standard’ drug trial, as characterised by the following main attributes, which is impacted:

  1. One set of patients receives no treatment.
  2. A second set of patients receives the drug being tested
  3. A third set of patients receives a sugar pill (placebo) that looks like the drug
  4. The trial is double blind – neither the patients nor the administers of pills know who is receiving the drug or the placebo
  5. The trial is randomised – the patients are randomly assigned into these groups

The key to the double blind aspect is that patients should not be able to differentiate between the placebo and the drug. Recent research with anti-depressant drugs showed that this is not always the case. It appears that many patients were able to work out if they were receiving the drug, because they could detect the side-effects of the drug. They could sense that they were taking an active substance as opposed to a placebo pill. Research was carried out to try to cancel this detection process by using an alternative drug instead of an inert placebo tablet. The drug would be one that generates similar side effects to the drug under test, but not any anti-depressant effect. The net effect was that the drug was no longer demonstrably more effective than placebo. Minimising the flaw of drug detection (that had negated the vital double blind aspect of drug testing), showed the drug to be much less effective as an anti-depressant.

placebo effect

To explain this, it appears that by being able to detect that a drug rather than placebo was being administered, the patient’s expectations of healing would be elevated. Conversely, a patient that knew he was taking a placebo would have a reduced expectation, and hence lowered placebo effect. There is a profound conclusion to be drawn from this: that drugs shown in scientific studies to outperform placebo may actually have been operating by virtue of the placebo effect in addition, or in place of any biochemical effect. The existing gold standard drug trial is not as sound as it is made out to be.

There is another twist to reveal, courtesy of ‘New Scientist’ magazine. Researchers discovered that the effect of the anxiety reducing drug Diazepam virtually vanished when it was administered unknowingly. Likewise, it was discovered that morphine could be provided at fairly high levels to patients in pain with no effect on that pain if administered without the patient’s knowledge. This too was a telling discovery. Morphine is an extremely well established and effective treatment for pain, effective in most people. Yet, it seems, a part of this very efficacy is due to the placebo effect. This links in well with the placebo detection problem – anti-depressant drugs being barely any more effective than placebo when the placebo used was indistinguishable from the drug.

But it is much more profound than this suggests, as Professor Irving Kirsch inadvertently discovered when he embarked on an exploration of the placebo effect. This was lucidly and succinctly detailed in his aptly titled book “The Emperor’s New Drugs”. Via meta-analyses of anti-depressant drug trials, he discovered that the drugs under test uniformly performed 25% better than placebo. This implied that the advantage was caused by the chemical effect of the drug on the taker.

But when he dug deeper, he found that this 25% anti-depressant edge was in fact a super-placebo effect. When an anti-depressant drug is taken, via trial or in daily life, the side effects are the signal to the taker that the drug is doing something. The stronger the side effects, the stronger the anti-depressant effect on the taker. Just as larger placebo pills have a larger effect, so it is with anti-depressant drugs. We have evolved a belief that the stronger the medicine, the more effective it will be. When an analysis of Prozac, the most famous of anti-depressant drugs was performed, the correlation between the degree of side effects and the efficacy of anti-depressant action was almost perfect (it had a 0.96 correlation value).

It appears that there may be minimal or no anti-depressant effect attributable to chemical action by anti-depressant drugs. And this is a staggering conclusion. Step right back, and picture the reality of the situation. A drug such as Prozac, taken by millions, earning countless billions for the manufacturer, may simply be a placebo tablet. The effect felt is mostly a placebo one, which is actually enhanced by its side effects – the placebo effect is enhanced or caused by the damaging, unhealthy side effect actions on the body. This is a paradoxical, but deeply sad conclusion, and illustrates the degree to which our health can be and often is subverted by the pharmaceutical industry. A subversion assisted by the regulatory bodies, who even help to cover up the existence of negative drug studies.

As an aside, but one that is important to note, this does not imply that all drug trials are subject to such profound super-placebo effects. It appears that the placebo effect is most pronounced with pain and depression suppression. But the net effect of these findings is that the gold standard drug trial is seriously flawed. And that there are likely to be drugs on the market that passed drug testing which would have failed such testing if the testing flaws were limited or removed.