Blinding us with Nonscience
The government decides, from time to time, that it needs to justify its intended course of action with science. But is that science or “nonscience”, selective statistics dressed up to look like science? Yes, we should reduce air pollution from vehicles and discourage excessive alcohol consumption; policies of this nature do not need spurious justification by bogus science. The SAGE committee has provided statistical extrapolations of their guesses of people’s reactions in hitherto unknown circumstances. This is known as “behavioural science” even though, because the conventions of science (theory – test – revise theory etc.) are not followed, it is not science at all.
This week saw the announcement that the sale of petrol and diesel motor vehicles will be banned from 2030. Fair enough. However it is based on nonscience published by Public Health England (PHE) in 2018. The authors are anonymous but the source is “the UK Health Forum (UKHF), in collaboration with Imperial College (the School of Public Health and the Business School), [which has] built on the UKHF’s existing flexible microsimulation model.” Perhaps unfairly, UKHF has been described as “a slush fund for 'public health' activists to lobby for the usual assortment of paternalistic anti-market interventions in lifestyle choices.” PHE seems to have come to the same conclusion because it withdrew funding from UKHF in the year following their document. UKHF then closed down.
PHE and the Department for Health have long commissioned work from the Centre for Social Marketing at the University of Stirling which, coincidentally, endorsed whatever PHE wanted to do. The Centre is headed by Professor Gerard Hastings who “also conducts critical marketing research into the impact of potentially damaging marketing, such as alcohol, tobacco and fast food promotion.” He authored “Europe’s only social marketing textbook: Social Marketing: Why Should the Devil have all the Best Tunes?”
The extent to which he who paid the piper called the tune should always be taken into account in assessing the credibility of academic research. Volkwagen notoriously fiddled the research they gave regulators. Imperial College has a strong relationship with the Department of Health and Social Care and there is nothing wrong with that. The funding for its School of Public Health research is not apparent from its annual report but, overall, government and health authorities are its third largest source.
In terms of provenance, the Imperial Environmental Research Group (ERG) under Professor Kelly is indeed impressive. The 2018 study would have been more convincing, however, if it had simply come from named ERG authors in the conventional way, and, better still, been published after peer review in a leading academic journal. As it is, we only have a mongrel.
Three areas are taken: Lambeth (high pollution), South Lakeland (low pollution) and England as a whole. South Lakeland is the area around Kendal, South Cumbria, with life expectancies about the national average. Life expectancy in Lambeth is much the same: 78.4 for men and 83.5 for women. It has been rising steadily, possibly due to air pollution.
The baselines in the PHE research were the years 2010 and 2015. Curiously, background pollution in South Lakeland seems to have declined over the five years but that was not commented upon. Fine particulates are treated as being 100% in the low category for South Lakeland with zero in the high category. Vice versa for Lambeth and one third in each of the three (high, medium and low) categories for England. Nitrogen dioxide variation was less extreme for South Lakeland. No discussion of these odd findings and assumptions.
The basic, and undeniable, thesis is that air pollution causes related diseases and consequential premature deaths. Table 7 makes some heroic estimates of the costs of primary, secondary and social care and medication.
Their conclusion (p.49) is:
“Between 2017 and 2025, the total cost to the NHS and Social Care of air pollution in England is estimated to be £1.60 billion for PM2.5[fine particulates] and NO2 [nitrogen dioxide] combined (£1.54 billion for PM2.5 and £60.81 million for NO2) where there is robust evidence for an association between exposure and disease. If we include the costs for diseases where there is less robust evidence for an association, then the estimate is increased to an overall total of £2.81 billion for PM2.5 and £2.75 billion for NO2 in England between 2017 and 2025.”
Curiously, the extension to 2035 comes earlier (p.29):
“From 2017 to 2035 it is predicted that 3,242 cumulative incidence cases per 100,000 population will be attributable to PM2.5 exposure in Lambeth, compared with 861 cases per 100,000 population in South Lakeland and 2,248 per 100,000 population in England. This represents a total NHS and social care cost of £9.41 billion, £80.26 million and £7.45 million per population for England, Lambeth, and South Lakeland respectively.”
One would expect the cost attributed to pollution would be the England level minus the South Lakeland (low) level but this subtraction does not seem to have been made.
Reviewing this paper was hard work as the authors seem to have immersed themselves in detail and failed to address three big picture questions:
The scenarios needed to be compared with a careful consideration of what would have happened without the pollution. People do die of the same diseases, e.g. heart attacks, when pollution is not a factor. And people do die prematurely, notably in the poorer areas of Lambeth, due to relative deprivation. Waiting in winter for London buses is a major health risk.
No attempt was made to disentangle causation from correlation.
It is hard to verify data extrapolated 20 years and no attempt was made to do so.
In short, when government says it is merely following the science: it may be blinding itself or they may just be trying to blind us with nonscience.