Insulin Coma Therapy: A Case in Point

Insulin coma therapy was a once wildly popular psychiatric treatment for Schizophrenia in which large doses of insulin were repeatedly administered to patients to produce daily comas, continuing on for several weeks.

Consider this 14-year follow-up study on the effectiveness of insulin coma therapy for schizophrenia that was published in the American Journal of Psychiatry in 1955, and stated "We conclude that insulin coma therapy is effective in restoring the schizophrenic patient to his prepsychotic adjustment." If these researchers could study an intervention (that we now, with hindsight bias, know to be obviously ineffective) on 800 patients for 14 years and still conclude that it was effective, we can imagine how many researchers doing studies with much smaller samples and follow-ups would only end up seeing what they want to see. How much of cherished research today would be discredited two or three decades from now?

Here is another study from another prestigious journal JAMA, that was published in 1958. It compared the effectiveness of chlorpromazine, an antipsychotic medication, with insulin coma therapy, and concluded that there was no difference in efficacy between the two, just that chlorpromazine had the advantage of being safer and easier to administer.

By the time insulin coma therapy was fully discredited by more meticulous and controlled studies, most physicians had already stopped doing insulin coma therapy. Not because they thought it was ineffective (they still believed it effective based on their clinical experience) but because they had found a safer alternative in the form of antipsychotic medications.

(Here is a good article on the rise and fall of insulin coma therapy.)

I think this particular example goes a long way to show the limitations of evidence-based medicine. Of course, the alternative is not to rely exclusively on clinical experience, which is even more faulty. The alternative is clinical experience informed by better and more rigorous evidence-based medicine, but it is just a reminder that we are not there yet. It can, in fact, be demonstrated (as in this paper) that most current published research findings are false.


Poonam said…
Very interesting reading. I'm a layperson.Just find you blog very fascinating. Thak you. :-)