Evaluating ECT effectiveness

Tania Gergel writes a very personal account in BJPsych of the benefits she has experienced from having ECT treatment. She discusses her treatment in the context of the academic literature about ECT and acknowledges that not all ECT recipients have had such positive gains.

I just wanted to pick up what Gergel says about her reasons for dismissing arguments challenging the effectiveness of ECT (see previous post). No-one is denying that ECT can have a placebo effect and I’m not wanting to undermine her faith in her treatment. 

Gergel argues that ECT would not have continued for 80 years if it wasn’t effective. But apparently successful interventions have been removed from medical practice when they have later been considered to do more harm than good. Examples in psychiatry would include leucotomy for psychosis and malarial treatment of dementia paralytica. The originators of both treatments were awarded the Nobel prize at the time because their discoveries were mistakenly thought to be advances in knowledge. In fact Cerletti was also nominated for the Nobel prize for ECT treatment of schizophrenia and manic-depression. ECT is no longer generally seen as indicated for schizophrenia.

Medicine has always exploited the placebo effect. Dramatic treatments like ECT may seem to be needed in desperate personal situations. Of course ECT was first introduced without anaesthesia and muscle relaxants, and such unmodified ECT continues to be given in some parts of the world. 

Doctors’ belief in their treatment and patients’ faith in it is a powerful mutually reinforcing combination (see eg. my book chapter). Randomised controlled trials were introduced to try to move on from the bias of such personal claims. The trouble is that there is still bias in clinical trials (see eg. my webpage). Even if there is a statistically significant difference between ECT and sham ECT in controlled trials, this difference could be because trials are unblinded, for example, through side effects of ECT treatment. For example, trial subjects may become aware that they have had real ECT because of the initial ECT side effects, such as headache. In which case if unblinded, expectancy effects could still be a self-fulfilling prophecy leading to an overestimate of treatment effectiveness. 

In general, we should be sceptical about the effectiveness of medical interventions because research methods are malleable enough to lead to exaggerated positive claims for treatment benefit (see eg. previous post). Individual experiences like that of Gergel, however well expressed in the BJPsych article, do not really solve this issue.