Dr Richard Asher made many excellent contributions to critical thinking as part of medical practice in the UK; one of the most entertaining of these was: Straight and Crooked Thinking In Medicine (pdf). Read it and savour it for the many useful and well-expressed observations. I’m going to highlight two that came up recently in different contexts (I’m separating them out with a rule but they are quotations and it seems easier to read than the standard blockquote).
Use of Vitamins
If I had to suggest one particular field of therapy which needs to come under the scrutiny of statistical clinical trial it is the use of vitamins. More than two million pounds’ worth of vitamins are prescribed each year under the National Health Service, and evidence that they do much good is very scanty. Are we not accepting arguments of this kind: (1) people deprived of vitamin B have sore tongues, (2) therefore everyone with a sore tongue needs extra vitamin B; (1) people deprived of vitamin C heal their wounds slowly, (2) therefore everyone with a wound needs vitamin C. These arguments alone are not rational-for instance, people deprived of salt for long enough may become mentally confused, but not everyone who becomes mentally confused needs salt.
The laity, too, have got the idea that as vitamins are essential to life any food that is full of vitamins is extra good for them. Subconsciously, I think they have adopted a sort of mathematical crooked argument that A is good, therefore 2A is twice as good and infinity x A is infinitely good. This is a dangerous argument. Cobalt, for instance, is (in minute quantities) essential to life, so is sulphur, yet a diet saturated with these substances might prove fatal…
But to return to vitamins the reason why they provide such a glorious field for the polytherapist and the commercial chemist is that they are more or less harmless (apart from very heavy doses of vitamin D). Therefore any doctor can prescribe them with the argument, “If you have vitamin deficiency,.they will do you good; if you haven’t vitamin deficiency, they can’t do you harm therefore logically we might as well prescribe them for everyone.” This argument is dangerous for these reasons: (1) by extending the argument we would find ourselves giving every member of the population an enormous number of vitamins and drugs; (2) if people are given vitamins they receive an impression that their diet is defective; (3) the cost to the country is enormous; and (4) people receiving a form of treatment (even if it is prophylactic) are more liable to believe they are unhealthy.
In our attitude to preventive medicine we have got to be careful that too much of the “can’t do harm might do good” argument does not lead to an excess of prevention. We must not be so busy preventing disease that we have no time to enjoy freedom from it…[pg. 461]
Another perversion of thought applicable to statistics is to believe that a statistical relationship between two things implies a causal relationship between them. It does not do so. It is very probable that statistics would show erythroblastosis foetalis to be twenty times less common among the offspring of opium smokers. This is no argument for encouraging opium smoking in the maternity wards. It depends on the fact that opium is largely a Chinese habit and that 99% of Chinese are rhesus-positive, and so cannot produce erythroblastosis foetalis. Similarly, if rheumatic fever is significantly more common in those who have sucked dummy teats in childhood, it may only mean that rheumatic fever is commoner among the poor, and so is dummysucking. If A varies with B, it does not show that A causes B. [pp. 461-2.] [Asher R. Straight and Crooked Thinking In Medicine (pdf). Br Med J. 1954 Aug 21;2(4885):460-2.
Asher might be a little wary of the news that a recent pilot initiative to reduce the number of frequent attenders at a GP’s surgery, 9 frequent attenders consented to a meeting with the manager of their local MIND at which:
[r]esolutions involved a range of ideas including volunteer work, subsidised reflexology, anxiety groups and subsidised counselling.
Those ideas are largely unexceptional but it might depend on whether reflexology had been presented as ‘a nice foot massage’ or as a therapeutic intervention that ‘might unblock energy flows’ or ‘correct imbalances’.
Several of Asher’s observations have had a profound influence and contributed to the groundswell of support for evidence-based medicine. However, it is with some tristesse that I must acknowledge that Asher would probably be very annoyed that these strictures and observations are as relevant today as when he made them more than 50 years ago. No piece that quotes Asher would be complete without his best-known observation: Talking Sense.
It is important to realise that ideas are much easier to believe if they are comforting, and that many clinical notions are accepted because they are comforting rather than because there is any evidence to support them. Just as we swallow food because we like it not because of its nutritional content, so do we swallow ideas because we like them and not because of their rational content.