Did you realise that if everyone in the UK switched off just one light bulb we could close down an entire power station? Think of the savings we could make!
Or, to put it another way, if we were to reduce our domestic electricity consumption by turning off a light bulb, thus bringing about a 0.5% reduction in demand we might be able to reduce national standby power generation by 0.5% which is equivalent to shutting down one out of two hundred power stations. Somehow, it is less impressive that way, isn’t it?
This is what I call the Metric Shift Illusion. You convert from one basis of measurement to another, in order to impress your audience. Virtually every grant proposal begins with the phrase: “XXX thousand people suffer from Condition YYYY at a cost of ZZZZ billions”.
The Wellcome Trust, no less, source of succor to destitute academics, has used this ploy in a Press Release:
A new report estimates that disorders of the brain cost the UK almost £113 billion per year, more than the GDP of New Zealand.
This is supposed to make you go “Wow! Big problem. These people need funding”. New Zealand is an entire country (and a very scenic and pleasant one, I am reliably told) so presumably the problem, brain disorders, is the size of that whole country, or something like it, and particularly deserving of our attention.
“The cost of brain disorders is 3126 Euros per person per year” is less stirring, though more informative, if we are sure it is accurate.
However, when you read the paper, the whole confection falls apart. Rather than “brain disorders” being tumours, abscesses and the like, they have thrown in age-related dementias and psychiatric disorders, such that their top 5 disorders are: headache, anxiety disorders, sleep disorders, mood disorders and somatoform disorders. I am all for recognising that we need to understand the brain, but I think this is pushing things a bit far. Headaches? There is usually little to be seen on a scan, and most of the treatments have a large psychological component. Anxiety disorders? Again, a very large psychological component. This is a very loose and inclusive re-definition of brain disorders.
The authors are explicit as to how they reached their total numbers: “a prevalence-based approach which multiplied the total number of UK persons affected by a disorder in a 12-month period (2010), with their mean cost in the same year”. A little later they confess: “the total amounted to roughly 45 million diagnosed cases, ranging between 26,000 cases of brain tumour and 18 million cases of headache”. So, this paper is a study of headaches and anxiety, and some other stuff. Does the Trade Descriptions Act apply to the titles of papers?
The crux of this paper is to understand how they calculated costs.
“Indirect costs associated with patients’ productivity losses constituted by far the largest component of the total cost comprising 46.4% of the overall cost, whereas the remainder of the cost was divided into 26.8% each for direct non-medical and direct healthcare costs.” So, what assumptions did they make in calculating productivity losses? Were they age corrected? Did they assume that the stated reasons for absence from work could be trusted? Did they assume that people at work are actually working? These are tricky matters, with much scope for approximations and rough assumptions. Commonly, staff reductions lead to increased productivity.
Figure 3 suggests the costly disorders are Dementia, mood disorders, psychosis, addiction and anxiety disorders. This certainly seems likely. My gripe is about the calculation of indirect costs. To give the authors their due, they have not added any in the case of dementia, mental retardation, and child disorders. The calculation of non-medical and implied social costs is difficult, and much affected by current economic realities. The only expensive “loss of production” disorders are those which affect irreplaceable key workers, and there are few of those. For many disorders there are only too many people willing to step in to take the job. Much more detail about the assumptions would have made the whole paper more convincing. (This applies not just to this paper, but to the general methodology of indirect costs). The whole point of open economies is that they are elastic. It is easy to find substitutes (either persons or products) for most requirements. Any loss of health in executives over 40 leads to faster promotion for executives under 40. When one mobile phone company goes down the other takes over their market share. And so on.
However, if we concentrate on direct healthcare costs, anxiety, strokes, addiction, psychosis and mental retardation are particularly expensive to treat.
The paper has interesting material, but New Zealand was best left out of it.
Some Other Popular Metric Shift Illusions
“If we were to increase income tax by just one penny, we would have billions to spend on infrastructure/sick donkeys/apprentices/invading foreign countries/stimulating the economy”. This last example is particularly silly, because a penny drawn from the productive economy is a certain loss, and the other schemes are uncertain gains, of equal magnitude at best, commonly much less.
“If everyone on earth held their breath for 10 seconds, think of the tons and tons of oxygen we could save. And if only we could get everyone on earth to hold their breath for 10 minutes………..”
By the way, don’t turn off your lamp bulb. For a very small cost you can avoid falling down stairs. People falling downstairs cost us billions and billions and billions.