In 1966 Eysenck criticized the effectiveness of psychotherapies, arguing that spontaneous remission alone was responsible for the individual’s improved condition - in other words, even if the individual did not have therapy, he or she would have improved, simply through a natural process of recovery, just as with a common cold. Eysenck, like so many others, struggled with many fundamental questions about how to evaluate the effectiveness of therapy. Eysenck suggested ways to measure the effectiveness of therapy - for example, using control groups.

An important question is who decides whether the therapy was successful. Therapists are unlikely to state that their therapy was ineffective, especially if the client has spent a lot of time and money on the sessions. In general, when therapists report on the success of their own therapy, they lack objectivity. Another issue is self-reporting by the client. This assumes that the client is in a position to judge his or her own progress and actually understands the techniques used by the therapist. Family and friends could report on progress made by the client, but sometimes they may be part of the problem and not really objective observers.

In order to study effectiveness, psychologists make use of outcome studies, which focus on the result - did patients show improvement or not? It is simple to say that all of the participants in an outcome study suffered from depression, but it is highly unlikely that they all experienced the disorder to the same degree. Although the therapists all practiced the same type of therapy, therapy is not a standardized procedure. Therapy is highly individualized and personal. It is not a linear sequence of techniques - assessment of the client is ongoing, and treatment techniques are continually adapted to meet needs as they surface.

Most research on effectiveness uses randomized double-blind placebo-controlled studies, where some of the participants are given the treatment, others are given a fake treatment (the placebo), and neither the researchers nor the participants know which is which until the study ends; thus, they are both "blind."

However, a single study is not enough to determine the effectiveness of a treatment.  For that reason, researchers carry out a meta-analysis.

Smith and Glass (1977) carried out a meta-analysis of 375 controlled studies of psychotherapy; they found that on average, a client that receives therapy is better off than 75% of untreated individuals. The researchers also concluded that all types of therapy are to some extent effective and that it may not be the specific kind of therapy that makes a difference. This is perhaps the same effect that is observed in the placebo effect - that is, just the fact that a person has a treatment is beneficial.

The same positive results were found in a more recent meta-analysis based on 147 studies and several thousand participants by Cuijpers et al. (2011). The aim of the study was to investigate the overall impact of psychotherapy in the treatment of adult depression compared to no treatment. They found that all forms of psychotherapy were found to be superior to not receiving therapy. Both of these meta-analyses challenge Eysenck's theory of spontaneous remission.

The effectiveness of treatments for depression

There is an enormous market for antidepressant drugs. In the UK for example, the National Health Service spent £270 million on antidepressants in 2011. But are the drugs effective? Would CBT be an equal or even better alternative?

Elkin et al. (1989) carried out one of the best-controlled outcome studies in the treatment of major depression, conducted by the National Institute of Mental Health. This study included 28 clinicians who worked with 250 patients diagnosed as having major depression. Individuals were randomly assigned to treatment using either an antidepressant drug, interpersonal therapy (a short-term therapy that focuses on interpersonal relationships), or cognitive-behavioural therapy. In addition, a control group was given a placebo pill. The clients that received either a placebo or the drug treatment were part of a double-blind design so that neither the patients nor the doctors knew which was which. All patients were assessed at the start, after 8 and 12 weeks of treatment, and after 18 months.

The results showed that just over 50 percent of patients recovered in all three of the treatment conditions. Only 29 percent recovered in the placebo group. The drug treatment produced faster results, but the study showed no overall difference in the effectiveness of CBT, IPT, and drug treatment. In other words, it did not matter which treatment patients received, all the treatments had positive results.

Kirsch and Sapirstein (1998) carried out a meta-analysis, covering 2318 patients who had been randomly assigned to either antidepressant medication or a placebo in 19 double-blind clinical trials. They found that seventy-five percent of the improvement in the drug group also occurred when people were given dummy pills with no active ingredient in them. A later study by Kirsch et al (2008) found no significant difference between placebo and SSRIs, except in the case of the most severely depressed patients. The researchers argue that the placebo effect may account for any observed improvement.

In the USA, all studies conducted by pharmaceutical companies must be filed with the Food and Drug Administration (FDA) and they rate the findings as either positive, neutral, or negative. Researchers can use these files to conduct meta-analyses that include all studies, also those with negative results. One of the problems with most meta-analyses is that they depend on published work but studies selected for publication of clinical trials may not tell the whole story and therefore lead to unrealistic estimates of drug effectiveness. Turner et al (2008) argue that one of the problems of studies of the effectiveness of anti-depressants is in fact publication bias. As a result, the results of the published studies could exaggerate the effectiveness of SSRIs like Prozac. The research team examined what happened to the data from antidepressant studies conducted between 1987 and 2004. There were 74 FDA-registered studies in all. FDA rated 38 studies as having positive findings and 37 of these studies were published. The remaining 36 studies were rated as having either questionable (12 studies) or negative findings (24 studies).3 of these studies were published as not positive but 11 studies were published with results that seemed positive even though they had been rated as questionable or negative by FAD. The remaining studies were not published. Overall, the researchers found that 94% of studies actually published reported positive outcomes; when published and unpublished studies were included for analysis, the percentage of positive outcomes decreased to 51%.

Results like these are difficult to interpret but it could seem that journals are less likely to publish research with negative results. According to the researchers, the published literature gave an effect size nearly one third larger than the effect size found in the FDA data. However, they warn that even though they cannot determine the exact cause for the observed publication bias the consequences of selective publication can lead doctors to make inappropriate prescription decisions that may not be in the best interest of their patients.

Before you think that this is a conspiracy by the drug companies, it is important to point out that this is a problem in all areas of psychology. It is referred to as the file drawer effect - that is, a bias in the scientific literature where only successful or positive results of the research are published, and studies where the null hypothesis is retained are not published. This then ends up misrepresenting the overall research in the field.

Finally, research by Riggs et al. (2007) studied the effectiveness of CBT in combination with either a placebo or an SSRI. The study was a randomized double-blind study with 126 adolescents, aged 13 - 19, who suffered from depression. Many of the participants were recruited from the social service and juvenile justice systems.

The researchers found that the adolescents in the study complied with the treatment to a large extent. After the study, the participants were rated by a physician who found that 67 percent of the patients in the CBT group + placebo group and 76 percent of the patients in the SSRI drug + CBT were judged as “very much improved” or “much improved” after being treated for four months. The researchers concluded that treatment with drugs and CBT is effective but that treatment with a placebo and CBT is almost as effective. The participants’ self-reports after the study showed that depression had decreased and so had the other behavioural problems. The researchers argued that teaching the adolescents cognitive and behavioural techniques helped them to manage negative thoughts and feelings that could trigger substance use. Treatment could start with CBT alone, and if the patient does not respond to that, a drug from the SSRI group could be added.

Checking for understanding

References: Abnormal