Following the posting of The Madness of Psychiatry, there has been a flurry of activity in the twittersphere with Louis Appleby, the UK’s suicide czar posting:
Figure 1 shows a set of fluctuating suicide rates year on year for veterans admitted to medical beds. The rates are higher than national suicide rates but these rates and their fluctuations are in keeping with what might have been expected in a set of younger men. The increases in the late 1950s and early 1970s may mirror the effects of the Korean and Vietnam wars, or perhaps other social factors or they may be entirely random.
Figures 2 and 3 are strikingly different to Figure 1. Figure 2 does not show the expected fluctuations linked to social factors or any randomness. It shows a steady rise in suicide rates in those who have been hospitalized for a mental condition. Until 1955 the rates are identical to the rates found in those hospitalized for a general medical condition.
But as of 1955, they start climbing in an uninterrupted fashion. The rises and falls we see in Figure 1 that might or might not be linked to social factors such as the Korean war are not there. This can be seen clearly in Figure 2 when the two sets of figures are superimposed and again in Figure 3 which show admissions to psychiatric beds on their own.
Why the bifurcation in 1955? This was the year of the introduction of chlorpromazine. Year on year after 1955 a greater number of tranquilizers (antipsychotics / neuroleptics) like chlorpromazine were consumed by veterans with mental health problems as an ever greater number of these drugs were marketed. These drugs were given to veterans who were depressed, anxious or psychotic – they were not as might be thought now restricted to veterans who were schizophrenic.
The question for Batman in the face of rising suicides in the US military, what would he do now.
What’s with the Question Marks?
In 2001, Khan et al had an article in the American Journal of Psychiatry which showed the figures for Risperdal, Zyprexa and Seroquel below. Intrigued by the high rates of suicide on Zyprexa, but even more so by the question marks where the missing data should be, I checked the FDA reviews for this drug and found unlike other reviews for other drugs the data was indeed missing. I have added extra data for Sertindole, and Geodon.
What if we tell him we didn’t collect the data on suicides?
I wrote to Lilly asking for the data and had the following response:
Dear Dr Healy
Thank-you for your letter concerning suicide attempts in clinical trials with olanzapine, which was forwarded to us by Harry Owens. I am sorry you did not find our previous letter on this subject helpful.
Your question was referred to our parent company in the USA, but unfortunately the specific data you requested are not available.
Yours sincerely – Dr A Simpson, Medical Director
The previous letter had intimated the data didn’t exist, which was scarcely credible. This was a time when the company according to documents later obtained under FOI were doing everything they could to get me or my junior doctor to prescribe Zyprexa, not showing me the data made it impossible to prescribe their drug.
Informed consent and Zyprexa
I wrote to the then Minister for Health Alan Milburn to see if he could help as he had responsibility for clinical trials in the UK, which could not both involve Zyprexa and informed consent if these data were not available. I never got a reply. Not even an acknowledgement. This was a time when Lilly it is rumored were threatening to pull out of the UK if Zyprexa were not written prominently into the Guideline for Schizophrenia, which was the very first NICE Guideline.
I later got to see the data. The number of suicidal acts on Zyprexa was higher than the number on Risperdal.
These data do not include suicides and suicidal acts that might have been triggered by withdrawal. In a 1993 FDA approval review of Risperdal, Andrew Mosholder, one of the agencies reviewers, notes that:
“The sponsor reports no instance of risperidone abuse or dependence. Withdrawal phenomena were not formally assessed after patients discontinued risperidone. Several patients committed suicide within one month of discontinuing risperidone, however, it does not seem reasonable to attribute this to withdrawal, given the absence of other indications of a risperidone withdrawal syndrome and the fact that schizophrenia is known to be a risk factor for suicide”.
These clinical trial data are ambiguous. They are not good quality data. There is no adjustment for patient exposure, and in this case some adjustment is called for but there is no way to undertake it. There should also be data for suicidal acts during the withdrawal period but these data are not included.
As they stand the data show a statistically significant increase in risk. This doesn’t mean antipsychotics cause suicide, it means that in these trials they caused a significantly greater number of suicides and suicidal acts than happened on placebo.
Although the data are poor, you might have thought journals would be interested. Far from it. They are not prepared to publish, even though this is the best we have and there is not a journal editor who does not trumpet clinical trial data as the gold standard.
In fact clinical trials are close to useless when it comes to suicide. It is easy to design a study of a drug known to cause suicide that would show a reduced rate of suicide compared to placebo (See Healy 2012). Clinical trials function instead for public health officials and journal editors as a bureaucrat’s tool to avoid exercising judgement. When they pose problems like the data here do – better they remain unpublished.
The Bureaucrat Will See You Now
Louis Appleby and other officials take a public health or vaccination perspective. This is essentially anti-medical when it comes to drugs. It accepts casualties without attempting to mitigate the risks. There is not supposed to be any questioning of drugs any more than there might be of vaccines. When officials like this talk about benefit risk ratios, they mean benefit on a public scale that includes job creation or the relocation of a company like Lilly out of the UK rather than the good of patients. The benefit is moreover ambigous – Zyprexa would in fact have an effect for GPI – dementia paralytica – tertiary syphilis, but using it might get in the way of us discovering that in fact penicillin would be a real benefit. The language has been perverted – an effect that suits some interest groups has been termed a benefit when it may well not be for many.
If bureaucrats of this type had run the automobile industry from the start they would never have allowed brakes in cars – and it would be close to impossible to persuade them that good drivers can in fact drive quicker if they have brakes than if they don’t. If you want a good driver or a good doctor these days, well now that’s getting harder and harder to find here in North Korea.