Tottenham Report: Gambling harms quantified, but can we trust the figure? By Andrew Tottenham, Managing Director, Tottenham & Co October 13, 2021 at 10:00 pm Last month, Public Health England (PHE), an advisory and support body to the National Health Service and Government, published its review of gambling-related harms; a summary can be found here Gambling-related harms evidence review: summary – GOV.UK (www.gov.uk). PHE came to the conclusion that the “annual economic burden” is approximately £1.27 billion; the 95% confidence range is £841 million to £2.12 billion. This may be the correct figure, it may be more, or it may be less. Unfortunately, this report does not get us to the point where we can confidently say what the actual figure is. The methodology was to look at existing studies of gambling prevalence, problem gambling, and gambling-related harm, as well as some of the responses to consultation exercises carried out by the Gambling Commission and strangely, “an anonymous sample of gambling-related tweets from Twitter”. Anyone who spends time on Twitter knows that tweets represent at best a polarised view of a subsection of society and at worst a toxic cesspit. The intent of the PHE review was to determine what the gambling-related harms are and their prevalence, what the risk factors might be, and what the related costs are, both social and economic. Given the controversy around gambling, it surprises me that there has not been a study long before now that has tried to determine the risk factors and quantify the harms from gambling. A longitudinal study of problem gambling was discussed prior to the implementation of the 2005 Gambling Act to determine a baseline and the impact of the changes from this piece of legislation, but it was dropped due to the cost. Since then, much of the research in the UK has been piecemeal, some carried out by researchers with an outwardly pro- or anti-gambling agenda. Primary research could have been a valuable resource in our understanding of gambling addiction, but the PHE review wasn’t that; instead, it was a meta study, a study of existing studies, some of which they disclosed were of low to moderate quality. PHE being a body with a public-health agenda, which is in and of itself a good thing, it is no surprise that it reached the conclusions that it did. So what is wrong with gambling being part of a public-health agenda? Public-health agendas tend to try to minimise use, either by discouraging the activity/product or by making it extremely difficult to access it. Think cigarettes. It is therefore unsurprising that the small part of their research into perceptions of gambling pointed out that, “With the exception of commercial stakeholders, there was consensus across different types of stakeholders that gambling is a public-health issue and it requires a public-health approach.” It seems strange to me that 54% of the adult population gamble on a regular basis (40% excluding the National Lottery), and yet 0.5% of the population are considered to be problem gamblers and 3.8% at-risk gamblers, and because of this, there is a hue and cry about restricting gambling. By the way, the definition of at-risk gamblers is people who are “typically low- or moderate-risk gamblers, meaning they may experience some level of negative consequences due to their gambling.” Yes, 0.5% of the population are experiencing problems, sometimes with devastating consequences. Would it not be better to understand the real causes of these problems and design and implement effective treatment? The difficulties I have with the PHE’s work are that its methodology is flawed. First, it assigns all of the estimated costs to gambling, when many gambling addicts have other addictions or mental- health issues that may or may not be caused by their disordered gambling. In its conclusion, PHE makes a giant leap from “is related to” to “is caused by”. PHE assessed hundreds of studies before rejecting some as inappropriate, irrelevant, etc. Of the 292, some of which were meta studies, that made it into the review, many had been excluded; only 21 studies were quantitative or quasi-quantitative, while the rest of the studies were mainly qualitative or qualitative in nature and less than 30 of the 292 related to the UK population. How can you assess the impact of gambling on the UK population by reviewing studies of countries with entirely different cultures, social care, and health systems? We already know that rates of problem gambling differ dramatically around the globe. These studies can certainly be informative, but how can they be relevant? One study found a strong correlation between gamblers and tobacco use. Anyone who has spent time in a casino knows this to be true. But no one would be stupid enough to state that gambling causes people to smoke. Correlation does not imply causation, and yet one of the studies reviewed concludes that gambling by young people causes substance abuse. Another study on a larger sample one year later found no such correlation. Are the conclusions of these two studies ignored, one cancelling out the other? And this brings me on to my second criticism, sample sizes. Many of the sample sizes, especially the qualitative surveys, were very low, some as low as 11. I frequently recall my first statistics professor, who always said, “Check the sample size. If the sample size is low, the results are likely to be extremely unreliable.” One of the surprising conclusions was the following, “We had moderate confidence that the following are not risk factors for harmful gambling among children and young people”. The list included the age and age of gambling onset. If we accept this conclusion, as some are readily doing for the financial quantification of harm, what does this mean for policy concerning young adults gambling? And now we come to the big leap. The report is peppered with statements like, “This suggests harmful gambling is related to …”, and “but there was a clear association between gambling at all levels of harm and …”, and “This suggests that gambling may …”, and, “Gambling was linked to …”. When calculating the costs of harm, there was no nuance, no uncertainty; it was an accepted truth that gambling caused these problems and all of the costs were down to gambling. It may be that gambling caused some of the problems identified. It may be that the related problems caused people to gamble in a harmful way. And it may be that an external factor caused both the problem gambling and the other related harms. Unfortunately, this piece of research does not bring us any closer to knowing the answer. The anti-gambling lobby has seized upon the report to justify their position that gambling should be severely restricted. The industry has been muted in its response. Probably because it does not want to be involved in further controversy at such a sensitive time when the report from the 2005 Act Review is being finalised. This has been a missed opportunity. The Department of Culture, Media and Sport, whose responsibility includes gambling, asked the PHE to carry out the review in 2018. Rather than ask a body with a public-health agenda to undertake a review of existing studies in order to quantify the harm caused by gambling, it would have been far better to have asked an independent group of researchers from a reputable university to carry out primary research. Then we might have an answer we all can all rely on.