Are antidepressants just placebo?

Published: 02-09-2020

Maybe not “just”, because the effectiveness of antidepressants is not low, but – as Irving Kirsch proves – their effectiveness is only due to the strong placebo effect. He came to such conclusions thanks to a detailed analysis of studies on antidepressants, including those that have not been published.

Who is Irving Kirsch

Irving Kirsch is a clinical psychologist based at the University of Connecticut. Due to the interest in the placebo effect, he decided to investigate, together with Guy Sapirstein, the strength of this effect in treating depression. Before starting his research, he was a supporter of antidepressants. He knew that part of their therapeutic effect was due to the placebo effect (the belief that the drug would help which actually helps), but he believed that the active ingredient in the drug was also important. While conducting psychotherapy, he often advised his depressed patients to see a psychiatrist who would prescribe medication for them. However, his analysis of the research led him to conclude that the overall therapeutic effect of antidepressants is due to the placebo effect.

Analysis of studies published in scientific journals

In 1995, Kirsch and Sapirstein analyzed 40 clinical trials involving a total of more than 3,000 depressed patients. They were interested in the mean improvement that followed treatment with antidepressants, psychotherapy, placebo, and no treatment. The results of their meta-analysis are presented in the diagram.

effectiveness of drug treatment

As you can see, there is a big difference between no treatment and placebo, and relatively little between placebo and antidepressants. It is also worth paying attention to the strength of the effect. In meta-analyzes, an effect size of 0.5 is considered medium and an effect size of 0.8 is considered large. Here, even the placebo hit a force of almost 1.2.

The results presented in the graph show that only 25% of the therapeutic effect can be attributed to the active substances of the drugs and the rest is the placebo effect and partly the effect of the spontaneous improvement that sometimes occurs. While not much, the difference between drugs and placebo is statistically significant nevertheless. Why then did Kirsch and Sapirstein consider drugs to be just a placebo effect? A more detailed analysis led them to this conclusion.

Side effects increase the placebo effect

Kirsch and Sapirstein found that all the drugs tested in the studies they analyzed had almost the same effect on improvement – from 24% to 26%. This is an unusual situation. They expected a greater spread – that some drugs would be more effective, others less. However, it turned out that the older and newer generation antidepressants are as effective. What’s more, certain medications that are not antidepressants, such as sedatives, stimulants, antipsychotics, herbal remedies, and thyroid medications, which in one study were administered to depressed patients without thyroid problems, have been shown to be equally effective in treating depression.

Researchers noted that the common feature of drugs that give greater improvement than placebo is an easily noticeable side effect. According to them, this 25% success rate can also be explained by the placebo effect. During clinical trials, each patient is informed that they will be randomly selected to receive a placebo or a group that will use the drug. You will also be informed about any possible side effects that may occur while taking the test. Although no one will tell the patients before the end of the test whether they are taking the drug or a placebo, if they experience side effects that they previously told might be likely to occur, they guess that they are taking the drug, and this enhances the placebo effect.

In support of his thesis, Irving Kirsch describes the history of the drug Emend, which was introduced to the market in 1998. Its main goal was to prevent nausea and vomiting in cancer patients receiving chemotherapy, but first clinical trials also showed improvements in the treatment of depression. The drug was supposed to be a breakthrough in the treatment of this disease as it had significantly fewer side effects than other antidepressants. However, it was not possible to prove the advantage of this drug over placebo in further clinical trials.

Another evidence of the placebo effect is the fact that in tests that compare the effects of two different antidepressants (so the participants are sure that they are taking the drug), 60% of patients improve, while in those comparing the drug with a placebo (participants are not sure if they are taking the drug), antidepressants only help 46% of patients. Moreover, a strong correlation has been shown between the effectiveness of SSRI antidepressants (e.g. prozac) and the strength of side effects experienced by the patient.

In order to avoid the possibility of the participant guessing whether he/she is given a drug or a placebo, sometimes active placebo is used. It is a substance that is not expected to have any therapeutic effect, but has drug-like side effects. An active placebo has been used in several studies of antidepressants. It was a drug used, among others. in diseases of the digestive system, side effects of which included dry mouth, insomnia, headaches and nausea, and therefore some of the effects that antidepressants can also cause. Of the 9 such studies Kirsch found, only 2 found a statistically significant difference between the drug and placebo.

Analysis of all pharmaceutical companies’ research

The analyzes carried out by Kirsch and Sapirstein sparked heated discussions in the scientific world. So 10 years later, together with Thomas Moor of the Washington University Department of Public Health and Medical Services, they decided to repeat the analysis with more studies. To this end, they asked the FDA (Food and Drug Administration) to provide the results of all the studies they have on the effectiveness of the six most popular antidepressants.

The Food and Drug Administration is the institution that regulates the introduction of new drugs to the market in the United States. Pharmaceutical companies are required to submit the results of all clinical trials they have conducted to it. It is commonly known that scientific journals are more likely to publish results of studies in which the hypothesis was confirmed than those that “did not come out”. This applies not only to antidepressants, but also to other areas of medicine and psychology. However, the FDA has the results of all studies, including unpublished ones.

So it is not surprising that an analysis of the results of studies submitted by the FDA showed that the drugs were less effective than placebo than the analysis of only published studies. It turned out that the effect of the active substances of drugs is only 18%. The mean difference in improvement between those taking the drug and those taking the placebo was 1.8 points on the Hamilton scale. This is a popular depression scale with a maximum of 51 points. According to the NICE (National Institute for Health and Clinical Excellence) guidelines, for a test result to be considered clinically significant, the difference between the drug and the placebo should be at least 3 points.

Moreover, almost all studies conducted by pharmaceutical companies involved people with severe depression. An analysis of studies published in peer-reviewed journals that include both drug-funded and independent research findings showed that antidepressants are more effective for severe depression than for moderate and mild depression. So, if pharmaceutical companies have conducted studies primarily on people with severe depression, they should be able to demonstrate greater effectiveness of drugs, which has not happened.

Further analysis showed that it is not true that the effectiveness of drugs increases with the severity of depression. Improvement with medication is the same regardless of the severity of depression, but with the severity of the disease, the improvement with placebo decreases, which is why there is a greater difference between medications and placebo in severe depression. It is not known exactly why this is so. Kirsch suggests that perhaps patients with severe depression, due to the fact that they have frequently taken various antidepressants before, are more likely to recognize whether they are in the placebo group or in the drug group. If they recognize that they are in the placebo group, its potency decreases.

How to demonstrate effectiveness that is not there

In addition to the bias in publishing studies (publishing only “successful”), which I wrote about above, Kirsch and his colleagues also discovered other tricks that pharmaceutical companies use to prove the effectiveness of drugs.

It includes:

Publishing the same research multiple times. To keep this from being noticed, they change study authors, make minor changes to the data, or present slightly different data to scientific journals and slightly different to drug regulatory agencies.
Publishing only selected data. For example, one study was conducted in 245 patients and showed a difference between the drug and a placebo of 3 Hamilton points, and the published study report only spoke of 27 patients and thus showed a difference of 15 points.
Conducting meta-analyzes, not taking into account all the studies carried out, but only “successful” ones.
Mixing unfavorable results with favorable ones, so that the summary result of the study is statistically significant.

Why did the FDA and similar institutions introduce ineffective drugs to the market?

The FDA and other drug regulatory agencies have all research results reported by pharmaceutical companies, not just those published in scientific journals. They know how many studies have failed to show the effectiveness of antidepressants. So why do they authorize the marketing of drugs that are no better than placebo?

The main reason is the criterion a drug must meet in order to be approved. Well, pharmaceutical companies must present the results of at least two clinical trials that have shown that the drug is more effective than placebo. Therefore, if they carry out up even 100 studies and only 2 of them prove the drug’s effectiveness, it will meet the approval criterion. A meta-analysis of 100 such studies will of course show no effectiveness.

In addition, a conflict of interest in drug regulatory agencies may also be relevant. These agencies are responsible for ensuring that only effective and safe drugs are released for sale, but pharmaceutical companies must pay a fee for the drug to be approved. The funds from these fees constitute 40% of the budget of the US Food and Drug Administration, and in the case of the European Medicines Agency it is as much as 70%.

Is Irving Kirsch right?

Despite the many arguments and detailed analyzes carried out by Kirsch and his colleagues, he still has many opponents who point to errors in his analyzes and incorrect conclusions. Critical articles have been published, among others in The International Journal of Neuropsychofarmacology, Annals of General Psychiatry and at PsychologyToday.com. Given the controversy of his results, it is not surprising to me that many people disagree with his results. I am personally convinced by Kirsch’s conclusions and I believe that even if the effectiveness of antidepressants is higher than his analyzes showed, it is still very low compared to placebo. However, I understand those who have different opinions and trust his critics more.

If you are also convinced by Kirsch’s arguments, and you are currently taking antidepressants, and after reading this article, you want to give them up, don’t do it without consulting your doctor. Irving Kirsch himself recommends in his book not to stop taking antidepressants alone. Rapid discontinuation of SSRI medications causes deterioration in about 20% of patients.