Double blind trial

The double-blind test is an elaborate instrument for estimating the size of the specific effect of a drug. It represents the core of evidence-based medicine. The need for such an instrument arises from the fact that people might become healthy independently of any therapy (spontaneous remission) and by non-specific effects like expectations, environmental stimuli and other influences (placebo effect).

In the double-blind trial, patients receive either the drug to be tested (verum) or a sham drug with the same appearance (placebo). They are randomly assigned to the check or control group (randomization). Both groups should not differ too much in their characteristics (severity of the disease, age and sex of patients, etc.). If patients do not know whether they are receiving a verum or a placebo, the trial is called blinded. If the therapists and all other persons involved do not know it either, in order not to influence the result by unconscious signals, one talks of a double-blind setting. This basic scheme exists in various modifications. If the therapeutic effect of the substance to be tested is stronger than in the 'placebo arm', an effective therapy is assumed. The double-blind test has some fundamental shortcomings that significantly limit its significance for therapies.

• Randomization is based on certain criteria or markers (surrogate parameters) that are thought of representing the disease. For other events, the double-blind test is more or less blind. The best known example was Sildenfali (Viagra©). In a double-blind test on hypertension, the substance proved hardly better than the placebo. Only a burglary at the company headquarters made the company aware that they had a potent agent.
• Patients in double-blind trials are generally highly selected and have clear diagnoses and symptoms. In general, they do not correspond too much to the general population to be treated later.
• Patients in double-blind trials usually receive a monotherapy. As a consequence, it is always doubtful in how far emerging results can be transferred to multimorbid patients.
• Since double-blind tests are very time-consuming, they are usually only carried out for 3 to 6 months. Longer tests are rather an exception. However, longer observation times can reverse the results. Therapies can be helpful short term, but harmful long term.
• With a double-blind test, it is not possible to distinguish whether the robustness or resilience of the patient is increased or not. Quite often the regulatory capacity is restricted by drugs, which might have a damaging effect under changing circumstances. For example, beta-blockers have a wide therapeutic spectrum, but they lead to a reduced adaptability towards heat. This might lead to higher incidences of heat stroke and death during a heat wave. That is, a therapy with a significantly lower effect under constant conditions might be much more effective under changing conditions.
• The double-blind tests make statements about collectives, not about individuals. The question to what extent a result is relevant cannot be statistically assessed. In a double-blind study on mattress hardness, it was found that a medium-soft mattress provided the best result for the majority of people. But a not insignificant number of people slept better on a hard or a soft mattress. What is the cosequence of that knowledge? Should everybody sleep on middle hard matresses?
• In statistics, there is a fundamental uncertainty as to what helps whom and under what circumstances. This was somewhat sarcastically summarized as follows: "Large numbers provide a statistically accurate result, of which one does not know to whom it concerns. Small numbers provide a statistically unsuitable result, but one knows better whom it concerns. Hard to decide which kind of ignorance is more useless."

The inadequacy of the double-blind test can be observed even in such a simple and standardized disease as 'Otitis media'. In the first major studies on the therapeutic effect of antibiotics in this disease, it was found that they were no more effective than a placebo, but increased the number of recurrences. First of all, this result showed that the effect of an antibiotic therapy in otitis media was significantly overestimated. But a uselessness of antibiotics did not necessarily make too much sense physiologically. After certain criteria for otitis media were established (redness and protrusion of the eardrum, pain and fever), in later studies antibiotic therapy was somewhat more effective than a placebo, but still associated with more complications. Moreover,these studies did not consider the question of resilience and/or long-term impacts of the therapy, on the intestinal flora, for example.

Double-blind studies thus provide clues that could and should influence the therapeutic action, but under the above-mentioned reservations. However, they cannot be regarded as a gold standard and therapeutic imperative.The overall situation of the patient must be assessed and considered and taken as the basis for any therapeutic action.