Seite 2: Biased Observers

Which role do the experts play who are carrying out the studies?

Michael P. Hengartner: A classical finding from psychology is that there is an observer effect in experimental studies. That is, if the study participants or the researchers know to which experimental condition they belong, then the result of the study will be different. That is precisely the reason why neither the patients nor the experimenters should know in such studies who gets placebo and who the active medication. This is called double-blinding.

But this often does not work: Both the patients and the experimenters usually have years of experience with the effects of psychopharmacological drugs. Thus, if a patient does not report some of the common side effects, such as a dry mouth, then both suspect that a placebo is used in this case, even if nobody knows this officially.

This can bias the study’s findings. After all, the experimenters ask the patients questions and assess the severity of the symptoms - whether they are mild, moderate, or severe -, from which then a general depression score is calculated. There are studies which have shown that about 90 percent of the experimenters know after a while who is getting placebo and who the active drug. This severely undermines the trustworthiness and validity of the study.

In some studies the patients were allowed to fill out questionnaires to assess their symptoms themselves. In such cases, the effects became smaller or disappeared completely. Clinicians and patients thus have different opinions about the drug effects. Besides that, the questionnaires commonly used often fall short of measuring the problems in the patients’ real life. Can they work? How are their relations with other people? These are the things that really matter to the patients, independent from the severity of their symptoms.

Limited Representativeness

And in which respect can the studies said to be representative?

Michael P. Hengartner: That’s another important point. In addition to people who are strongly responding to placebo, those who have other problems like anxiety or substance abuse disorders, like alcohol or drug addiction, are excluded from the trials. We call this comorbidity. Usually patients suffering from psychotic symptoms or suicidal thoughts are excluded, too.

Yet, every clinician knows that patients with severe depression often have a variety of problems. The studies would thus have to show which effect the antidepressant drugs have in this typical group of patients, but they are investigating a carefully selected subgroup instead. Therefore the results of these placebo-controlled trials cannot be generalized.

You also write that antidepressant drugs can increase the suicide risk and damage health. Could you explain that?

Michael P. Hengartner: This discussion is more than twenty years old. It has actually been found right from the early 1990s that the then new selective serotonin re-uptake inhibitors (SSRIs) can cause extreme psychomotor agitation and suicidal behaviors, in healthy people as well as in those with a mental disorder. That is, people can get uncontrollable aggressive impulses, suicidal ideation, or take steps to harm themselves.

One problem was that the efficacy trials included too few subjects for these effects to become statistically significant. One must know that the scientific community takes the 5%-threshold of significance more important than justified by the statistical theory. Thus only later meta-analyses with the data of thousands of people confirmed that antidepressant drugs can lead to an increase of suicidality. David Healy, whom I already mentioned, pointed that out very early and got many enemies because of that.

Moreover, when suicide attempts occurred in the previously mentioned washout-phase, then they were attributed to the placebo group. That made it less likely that such problems were noticed when they were caused by the investigational drug. It is also difficult to measure them, leaving aside lethal suicide attempts. In some studies suicidal ideation or self-harm were simply recorded as "emotional lability" or "worsening of depression" to conceal that problem.

Health Risks

And what do you mean by possible damage to one’s health?

Michael P. Hengartner: This is related to the long-term use of psychopharmacological drugs, thus for a period of many years. One must take into account that these are psychoactive substances. That is, these drugs do have effects in the brain, even though, in my opinion, they do not have specific anti-depressive properties.

It is also known that serotonin has various effects in the whole body and that it influences the immune system and metabolism for instance. This can eventually lead to a dysfunction of basic bodily processes. And of course to neurobiological disorders owing to enduring changes in neurochemistry.

I don’t want to condemn psychopharmacological drugs. One has to deal with them more consciously, though, and to consider how the substances work in the whole body. This has to be included in the risk-benefit analysis. By now several studies have demonstrated that the long-term use of psychopharmacological drugs increases certain health risks.

What you are summarizing in your article is the result of many years of research. These are scientific publications that have been available in libraries for many years. Why is so little of that knowledge disseminated to the public?

Michael P. Hengartner: One has to realize that mainstream psychiatry has made itself very much dependent on biomedical concepts and the prescription of psychopharmacological drugs. These are the disease models investigated most frequently and the treatments prescribed most often. This message is also spread successfully in the media and marketed aggressively. The hopes in Biological Psychiatry have been and still are very high. Opposing views, by contrast, have been ignored for a very long time.

This is also due to the systematic biases described earlier. We now know that some studies found a meaningless difference of one to two points on a depression scale of 52 points between the placebo and the antidepressant group. However, the actual publication eventually reported a difference of more than ten points, because only a selected subset of the patients has been included in the analysis. Physicians and researchers then read these biased reports and got a wrong impression.

It goes without saying that those working in the clinics or hospitals have a great wish to offer patients something that really helps. The doctors prescribing the drugs do not want to give up the belief in their effectivity.

Increasing Prescriptions

But why are still so many antidepressant drugs prescribed? You even say that the prescriptions are still increasing.

Michael P. Hengartner: There is probably more than a single cause. I think that the helplessness in clinical practice also plays a role. What didn’t clinicians already try with severely suffering patients? They want to help people, after all, such that they feel better.

Additionally, resources and alternative possibilities are scarce. Psychotherapy is difficult, for example, in the case of an acute and severe depression with psychotic symptoms. The patients might not participate or stop the treatment quickly under such circumstances.

Another reason is, as I already said, that the effectivity is of the medications is overestimated. Because this opinion is so widespread, critical studies have more difficulties to get through the peer-review process. The referees, who should actually be neutral and independent, consider the critical view as wrong and thwart their publication. One can also be marginalized quickly as an irrational polemic or even sectarian.

How about the official treatment guidelines? Are they clear?

Michael P. Hengartner: These express another problem, that there are different opinions at different places about the best treatment. For example, the guideline of the American Psychiatric Association instructs doctors to prescribe antidepressant drugs already in cases of mild depression. By contrast, the respective British committee strictly advises against this and states that one should rather wait and carefully observe in such cases.

It is interesting that both committees, the American and the British psychiatrists, more or less refer to the same studies - but draw completely different conclusions from them. Thus, what should a physician do? Also consider that he or she participates in continuing medical education that is mainly funded by the pharmaceutical industry, where somebody comes to praise medications.