May 9, 2019

expert reaction to study quantifying the harm of other people’s drinking

Research published in BMJ Open suggests that one in five people in England have been harmed in some way by other people’s drinking over the past year.

Dr Bob Patton, Academic lead in the Drugs, Alcohol and Addictive Behaviors Research Group, University of Surrey, said:

“We know that alcohol is associated with a range of physical and psychological harms to the drinker themselves, but this survey shows that such harm is not just limited to the individual.  The finding that one in five over 16s in England have experienced harm as a consequence of others’ drinking should encourage policy makers to consider ways to raise the bar further when developing strategies to reduce consumption.  Given that those who drink themselves are more likely to be both victim and perpetrator of harms, this should then decrease the health, social and economic costs to our society.

“This research is cross sectional, meaning that it cannot demonstrate that alcohol use is responsible for the harms reported by others; only that it is related to them in some way.  The survey was completed by 4874 people however we do not know how many folk refused to take part, and those that did were limited to individuals who were willing and able to participate, so anyone on holiday, in hospital or residing at one of Her Majesty’s Prisons have been excluded, all of which limits the representativeness of the sample.  As such this study raises more questions than it answers with regard to the factors that influence alcohol related harms to others, however it is an important first step in understanding the costs and consequences to the wider population, and provides good evidence that alcohol related harm is not simply limited to the consumer.”

Dr Joy Leahy, Royal Statistical Society Statistical Ambassador, and statistician, said:

“The authors correctly point out that there is a potential selection bias which is inherent in all national surveys.  One of the main fears with a survey is that there will be bias in the demographics that are willing and available to be interviewed.  This could be particularly important on a topic such as alcohol harm to others, as results could be very different across various demographics.  The authors mention that the sampling method they used was a hybrid of random probability sampling and simple quota sampling – they don’t go into much further detail in this paper, but they reference the Alcohol Toolkit Study.  From my reading of the Alcohol Toolkit Study (BMC Public Health 2015) it appears that this survey has good measures in place to reduce the risk of selection bias, including making sure that quotas were fulfilled for interviews with people across different working statuses, ages and genders.  It therefore looks like this survey is a good representation of the of the full demographic of households across England.  Although, as the authors point out, there are some other demographics, such as homeless people, that are still not represented.

“There could however still be bias due to respondents not completing some or all of the survey – the authors mention for example that they excluded people who hadn’t completed all questions (3.7% of the total respondents) and additional respondents who answered ‘don’t know’ to certain questions.  The authors note that there were some differences between people who answered questions and people who didn’t, which could affect their results.  It’s possible that respondents who did not answer specific questions may not want to disclose or discuss a specific incident – so it might be that the true figure for those experiencing alcohol related harm from others may actually be higher than reported.

“Although this was a large survey (approximately 5000 respondents), we should be careful not to read too much into some of the subgroup analyses.  For example, the press release states “One in five (19%) respondents who reported having been forced or pressurised into something sexual, said this was at the hands of a stranger, but the most commonly cited perpetrator was a co-habiting partner (23%, rising to almost 40% when including partners who lived elsewhere).”  27 respondents reported having been forced or pressurised into something sexual, 5 respondents reported being pressured into sex or something sexual from a stranger, and 6 respondents reported that the perpetrator was someone the respondent was in a relationship with and lived with.  With such limited data on this type of harm, it would not be appropriate to draw conclusions on the most frequent perpetrator of this type of harm from this specific study.”

‘Alcohol-related harm to others in England: a cross-sectional analysis of national survey data’ by Caryl Beynon et al. was published in BMJ Open at 23:30 UK time on Thursday 9 May 2019.

Declared interests

Dr Bob Patton: “1. I have received funding for alcohol research from both the NHS and the Department of Health.  2. I undertake occasional work for Ipsos MORI.”

Dr Joy Leahy: “I’ve no interests to declare.”