In pursuing a public health approach to gambling, the Gambling Commission has doubled down on Public Health England’s dodgy data. Dan Waugh holds the regulators to account.
There are few clearer indications of an attempt to suppress personal freedom than the coupling together of three otherwise benign words – ‘public’, ‘health’ and ‘approach’. While public health is a well-established and important field in disease prevention, its insinuation into myriad aspects of modern life often goes unexplained (as Christopher Snowdon and others have repeatedly warned). Those who stake a public health claim to particular issues (or, just as often, ‘crises’) rarely explain just why they should be considered so; or what in fact the public health response might constitute or be expected to achieve. In a growing number of western societies, those three little words appear to have acquired the creed-like power to justify almost any intrusion on the citizenry.
Recent developments in the public policy debate on gambling (occasioned in part by the government’s drawn-out review of legislation), have shed light on what the expansion into everyday life of revisionist approaches to public health might in fact entail.
In September 2021, Public Health England (PHE) published a series of reports under the umbrella of its evidence review of gambling-related harms. The review was largely constituted of rehashed data from NHS Health Surveys with the only substantively new addition being PHE’s estimate that harmful gambling was associated with £1.27billion per year in economic and social costs (including 409 excess deaths from suicide). The publication of these estimates met with outrage in both the national news media and the Houses of Parliament.
The only problem was, PHE’s figures had been manufactured. Faced with a dearth of hard data, the research team attempted to extrapolate from a series of one-off studies, based on non-representative (and often quite small) samples – many drawn from outside the UK. The result was a report replete with obscure calculations, methodological errors and a disregard for comorbidities, counter-factual scenarios or social and economic benefits.
In March 2022, Cieo exposed the problems with the PHE claims. Later that month, the Gambling Commission met with researchers to listen to concerns about the report. The market regulator had supported PHE’s findings and might therefore have been expected to know something that others did not. As it turned out, they did know quite a bit – but nothing that validated PHE’s claims. With the Office for Health Improvement and Disparities (which had subsumed the PHE team) unwilling to meet and with little prospect of illumination from the regulator (‘not a priority,’ it claimed), the researchers turned to the Freedom of Information Act.
A request to the DHSC to publish the calculations on which PHE had based its estimates met first with obfuscation and then flat refusal (‘we would be required to adapt and format our analyses for external use which would imply considerable time’). A request to the Gambling Commission to release copies of correspondence between the regulator and PHE yielded a small number of heavily redacted communiques – an implausibly thin harvest given the regulator’s involvement. When challenged, the Commission admitted that the disclosure had been partial and, eventually, another batch of redacted documents was released.
The Commission’s correspondence is revealing. Exchanges early in 2021 show keen anticipation at the prospect of PHE report’s publication, which it considered would have a bearing on the outcome of the Gambling Act Review. In May, an executive director at the regulator wrote ‘we are currently developing our advice to the government on the gambling review and are keen to take into account your work as we develop that.’
By September 24th, PHE, preparing to publish, was looking for cheerleaders. The Commission duly obliged, agreeing to provide a ‘supportive quote’ despite the fact that – at this point – it had not actually seen any of the analysis and so presumably did not know what it would be supporting. Later that day, the regulator received a draft copy of the PHE press release, which included the following passage:
Health leaders from organisations such as X and X are joining PHE in calling for a public health approach to gambling focusing on prevention, early intervention and treatment. This approach is similar to how we tackle tobacco consumption or unhealthy food consumption, and would require cross-government support.
Reinventing prohibition
Here was a clear signal that the public health establishment considered gambling – rather than disordered or harmful gambling – to be the issue in need of resolution. Given that the government plans to stub out smoking by 2030, it smacked of prohibition as an end-game.
The Gambling Commission suggested that references to tobacco be excised on the grounds that ‘it’s [sic] inclusion may see some stakeholder pushback’ as ‘many consider that tobacco is not a suitable comparison to gambling as it is harmful to all’. Critically, the regulator’s caution stemmed from concerns about how the tobacco reference would be construed rather than what it revealed about PHE’s agenda (an agenda that was concealed when the passage was dutifully removed).
By June – prompted by concerns raised by both Cieo and the Racing Post – a handful of journalists started asking questions, including whether the Commission had undertaken any analysis to establish the reliability of PHE’s cost estimates. The regulator’s response then was (and still is):
as the gambling regulator it is not part of our statutory role to investigate, or provide a detailed assessment, of the methodologies used by other public bodies in fulfilling their statutory roles.
While perhaps technically correct, this turned out to be deeply misleading. It may be that the Commission was not required to assess the PHE claims – but that did not mean that they had not done so.
A further batch of Freedom of Information Act disclosures revealed that the Commission had reviewed PHE’s work, not once but twice. The first of these reviews was carried out between November and December 2021. In it, the Commission expressed ‘surprise’ that ‘the £1.27bn cost of the gambling industry has been promoted by PHE as one of the main findings from the entire review’; and speculated that,
this may be due to its status as being one of the few conclusions that is ‘new’, a desire to emphasise the need for further research to strengthen the evidence base or intending to ensure gambling is considered as a public health issue. (1)
The reviewer went on to state ‘I don’t believe that even PHE have [sic] confidence that £1.27bn is an accurate estimate of those costs’; and seemed particularly sceptical about the suicide claims, observing that:
considering this section accounts for nearly half the gambling-related costs identified in the report, it is not particularly lengthy and is based on a single study in a different jurisdiction.
Finally, the Commission highlighted the absence of any attempt to estimate benefits from gambling (cost analysis, unleavened by consideration of benefits is no guide to policy) but indicated long odds that this would ever be undertaken, writing that,
a full economic evaluation of the gambling industry could be slightly awkward politically as it would fund work to identify the benefits of gambling but – to me – feels like the most balanced approach.
To this, the reviewer added: ‘it’s likely that we would be critical of any industry-funded effort to estimate the benefits of the gambling industry.’ A regulator that had been happy to endorse – unseen and unchecked – the claims of Public Health appeared to express an a priori bias against research funded by its own licensees. This asymmetry has serious potential consequences. If the Commission is systemically inclined to believe gambling’s detractors and disbelieve its advocates then the public consultations that have notionally justified so many of the rule changes in recent years may in fact have been little more than bureaucratic theatre.
Manipulating the figures
The second review appears to have taken place in April 2022 – as a direct consequence of issues raised by researchers the month before. The Commission’s reviewer notes that concerns about the reliability of the PHE suicide estimates ‘seem valid’ and astutely observes that ‘the reality is that reliable data does not exist’. In a troubling coda, he or she adds that,
if investigated in future, the [suicide mortality] multiplier may be lower but the [problem gambling] rate may be higher than 0.4% to produce a broadly similar number.
In other words, the regulator appeared to suggest that while PHE had overstated the rate of suicide among people with a ‘problem gambling’ classification, its overall estimate of the number of suicides might be retrieved if the reported rate of problem gambling could be shown to be higher. Given that the Commission is in the process of taking direct control of official prevalence reporting (previously conducted as part of the NHS Health Survey), this could be construed as manipulation in contemplation. (2)
The documents disclosed under the Freedom of Information Act show not only that the regulator did review PHE’s work – but also that it expressed strong concerns about its reliability. This makes what happened next all the more remarkable. In May 2022, the Gambling Commission’s Advisory Board for Safer Gambling (ABSG) published its Year Three Progress Report on the National Strategy to Reduce Gambling Harms. In it, the ABSG (whose chair had been a member of the PHE Expert Reference Group) cites the PHE work as a ‘strength’ of the strategy; and made a specific point of stating that the suicide statistic ‘should act as a significant catalyst towards [sic] action’. By this point, the Commission was aware that the suicide statistic was ‘unreliable’; yet it permitted the ABSG assessment to stand. (3) There is no evidence that the regulator brought its concerns about PHE’s work to the Department for Digital, Culture, Media and Sport (its sponsor) or to the DHSC (PHE’s sponsor). Certainly, none of the documents released under the Freedom of Information Act suggest that this occurred.
Sunlight is the best disinfectant
By the start of August, a cover-up involving one government ministry and two state agencies was beginning to crack. On 2nd August, the DHSC finally admitted (in response to repeated Freedom of Information Act requests) that PHE had made a fundamental ‘mistake’ in its analysis of suicide prevalence. On the same day, the Health minister, Maggie Throup provided an answer to a parliamentary question submitted by the Conservative MP, Scott Benton. (4) She revealed that her department would conduct a review of the PHE estimates and that it would publish, in full, the underpinning calculations at some point in the Autumn.
We may be in for further twists and turns before this saga ends. In the meantime, we are left with three key questions to consider:
Should the government department responsible for the publication of such a flawed report be permitted to review and re-publish without external scrutiny?
Should attempts to cover up the problems with the report by both the DHSC and the Gambling Commission be independently investigated (and what might be the consequences of failing to do so)?
How much weight should be accorded to related research projects, including PHE’s Delphi study (which proposed, amongst other things, ‘standardised design and packaging for gambling products so that no brands; colours; imagery; corporate logos and trademarks are displayed’) and the University of Sheffield’s forthcoming study of maximal treatment costs? (5)
Amid all the questions, we have at least answered the question that we started with. Those three little words, ‘public health approach’ – used within a gambling context at least – appear to signify a quite different triptych of incompetence, manipulation and deceit.
(1) PHE took pains to point out that its cost estimates were ‘associated with’ (but not necessarily caused by) harmful gambling. By December however, the market regulator had shifted seamlessly from correlation to causation. PHE’s characterisation of costs associated with people with an ‘at risk’ or ‘problem gambling’ classification had now become ‘costs of the gambling industry’.
(2) In May 2022, the Gambling Commission published results from an experimental Pilot Gambling Prevalence Survey which reported a problem gambling rate around three times higher than NHS Health Surveys. Subsequent analysis of this work has revealed significant flaws.
(3) The Commission has refused to disclose documents relating to the ABSG’s consideration of the PHE review.
(4) The answer, when it came, was a whole seven weeks overdue.
(5) A valid area for research but with high potential risk of manipulation and political interference.