Anyone working in an insulin unit in the 1950s would not recognise Dr Pimm's account of the results of their treatment, ^{Reference Pimm1} or details of what it involved. The patients received daily and increasing doses of insulin, rising to many hundreds of units, for a 6-week period. The depth of the resulting hypoglycaemic coma was determined by the patient demonstrating a Babinski response over a period of 15 min. They were then revived by ingesting glucose.

I worked in the insulin unit at Newcastle General Hospital from 1956 to 1959, when I was senior registrar to Sir Martin Roth. Insulin treatment was reserved for people experiencing their first attack of schizophrenia, and from memory I would say half made a complete remission and another 25% improved. Nobody thought that we were effecting a cure, but remissions lasted about 2 years. One woman relapsed 9 years after her treatment. Of course there were dangers, but in those days the alternative was incarceration in a locked ward in a Victorian asylum, with little hope of rehabilitation or discharge.

Martin Roth was an intellectual giant, but also a man who was perspicacious and compassionate, and who would not have contemplated using such a treatment if he did not think it effective. The depth of the coma seemed to me to be critical in terms of remission. A few patients did not regain consciousness when given glucose, but usually ‘came out of it’ after some hours, although there was the occasional death. Very occasionally, a patient who was clearly psychotic who had an ‘irreversible coma’ on recovery was greatly mentally improved. These days, people find this difficult to believe, but I witnessed it on one occasion. I find it inconceivable that a multitude of psychiatrists, working in Europe and North America over 25 years, would not have noticed that the treatment they were giving was having no effect, when it clearly was, if only for a limited period. The real question was not whether insulin worked but how did insulin work.

I have no wish to minimise the success of Dr Bourne's crusade, but what made insulin units redundant was the realisation that the new antipsychotic drugs actually worked, and at last, we had an effective, cheap and long-lasting method of managing a seemingly incurable disease. This was generally accepted by 1960.