"Middle-class over-50s have become a generation of problem drinkers," the Mail Online reports – a headline that actually has little basis in fact.
This follows the analysis of more than 9,000 adults aged over 50 from the English Longitudinal Study of Ageing. It found that over-50s falling into a "higher-risk drinking" category were more likely to have middle-class traits such as high educational achievement, better self-rated health, and being socially active.
Higher-risk drinking was defined as drinking more than 50 alcohol units per week (equivalent to five or more bottles of wine) for men, and over 35 units per week (three-and-half bottles of wine) for women.
The Mail's headline got the wrong idea though, because only 3-7% of over-50s drank at these "higher-risk" levels. While an obvious issue of concern, it a stretch to say this is a generation of problem drinkers.
There were also significantly different patterns between men and women. For example, higher-risk drinking was linked to higher income, but only in women.
These uncertainties aside, the study does reinforce the fact that alcohol misuse and the risks of drinking too much do not respect class boundaries. You can do just as much damage by drinking champagne to excess as you can by drinking cheap cider.
Where did the story come from?
The study was carried out by Professor José Iparraguirre from the Research Department of Age UK and was also funded by Age UK.
The Mail, The Daily Telegraph, The Times and The Guardian reported the research facts accurately, although none discussed any limitations associated with the research. All papers carried useful quotes from independent experts. For example, the Mail included a quote from Professor Sir Ian Gilmore, chairman of the Alcohol Health Alliance, who said: "Whilst it may be true that middle-class drinkers are able to offset some health problems because of healthier diets and lifestyles more generally, the risks of serious health harms are still significant. For example, even low levels of drinking increase the risk of developing cancer."
What kind of research was this?
This was an analysis of a longitudinal survey of ageing, estimating what risk factors might be linked to harmful alcoholic drinking in over-50s.
A longitudinal study involves repeated measures over time, so is great for measuring changes in drinking behaviour. One of the disadvantages is that they tend to rely on self-reported survey estimates of alcohol intake, which can be an unreliable reporting method. Some people may deliberately under-report their drinking habits out of embarrassment or social pressure. Others might under or overestimate them by accident through not knowing how many units are in their drinks. Heavy drinkers may forget how much they drank over the course of their sessions.
By using lots of people, over and underestimates should even out to give a relatively accurate picture of what’s going on, but this is never perfect.
What did the research involve?
Researchers analysed the drinking habits of 9,251 adults aged over 50, looking for links between their drinking habits and their income, lifestyle and social situation.
The drinking and other lifestyle information came from responses to the English Longitudinal Study of Ageing. This gathered data from a representative sample of UK men and women aged 50 and over from 2008 to 2011.
Their definition of harmful drinking used the highest risk category from NICE guidelines. This is called "higher-risk drinking" and describes men drinking more than 50 alcohol units per week, or women drinking more than 35 units per week. For men, this is equivalent to five or more bottles of wine a week, or 16 pints or more of strong lager, and equivalent to three-and-half bottles of wine, or 11 pints or more of strong lager for women.
They used two sources of alcohol unit measures, to see if that made any difference to the findings. The first calculated that:
- one pint of normal-strength beer was equivalent to two units
- a 175ml glass of wine is equivalent to two units
- a 250ml glass of wine is equivalent to three units
The second, using the drinkaware website, calculated that:
- one glass of wine was equivalent to three units
- one pint of beer was equal to three units
The analysis estimated how the risk of harmful drinking was influenced by:
- lifestyle (diet, smoking and physical activity levels)
- self-reported health (poor to excellent)
- marital status
- caring responsibilities
- children in the home
- social isolation
The researchers looked for a link between heavier drinking and people quitting the longitudinal survey. They found no link, suggesting that people dropping out were not an important issue.
What were the basic results?
The results showed different patterns for men and women:
- The risk of women being in the higher-risk drinking category reduced steadily from the ages of 50 to 90.
- By contrast, men’s risk peaked in their mid-60s, before declining.
- For both sexes, reporting better health was linked to increase alcohol consumption.
- Achieving higher educational attainment and smoking were linked to being in the higher-risk drinking category.
- Income was linked to higher-risk drinking in women, but not men.
- Having a job had no link overall. But retirement increased the chances of drinking at higher risk levels for women.
- Being single, separated or divorced was linked to being in the higher-risk drinking category, although only for men.
- Loneliness and depression are not associated with higher-risk drinking.
- Having caring responsibilities reduces the probability of being at higher risk for women.
Some of the analysis looked at how likely people were to enter the higher-risk category from lower drinking levels over a two-year period. This found:
- For women, being younger and having a higher income increased the probability of becoming a higher-risk alcohol drinker over time.
- For men, not eating healthily, being younger and having a higher income increase the probability of becoming a higher-risk alcohol drinker
How did the researchers interpret the results?
The researchers attempted to draw some themes from their many individual results: "… we can sketch – at the risk of much simplification – the problem of harmful alcohol drinking among people aged 50 or over in England as a middle-class phenomenon: people in better health, higher income, with higher educational attainment and socially more active are more likely to drink at harmful levels."
They say the concept of: "successful ageing […] embraces components such as non-smoking, greater physical activity, more social contacts, better self-rated health and absence of depression, among others." And that their results show: "generally speaking, people aged 50 or over ageing 'successfully' in England are more at risk of drinking at harmful levels".
This study showed that higher-risk drinking was linked to a number of factors the researchers described as "middle-class", like higher educational attainment, being socially active and good ratings of health.
Professor Jose Iparraguirre, author of the research, said in the Guardian: "Because this group is typically healthier than other parts of the older population, they might not realise that what they are doing is putting their health in danger".
There are a few reasons to be cautious with these findings.
The study produced a lot of results, so there is a risk some were chance findings. This is particularly relevant, as the analysis focused on higher-risk drinking. Of the large number of people taking part in this survey, only a small chunk (3-7%) fell into this category. Analyses based on these smaller numbers are more likely to give chance findings.
Also, the study only tracked people for a maximum of three years, which isn’t particularly long. Studies tracking drinking behaviour over longer periods of time might show different patterns.
The study used a representative group of UK older adults, which is a strength. However, we can’t be sure this paints a totally realistic picture across the UK, as there may be geographical variation.
The researchers tell us that heavy drinking in older age is linked to death in the short term. This means there was a risk of higher-than-normal numbers of older moderate drinkers, as heavy drinkers might have died earlier. Usefully, the researchers re-ran the stats using an age cut-off of 70. This showed no difference to the age 90 cut-off used for the main analysis, meaning this wasn’t an important influencing factor.
Rosanna O’Connor, director of Alcohol Drugs and Tobacco at Public Health England, said in the Guardian: "Around one in five adults regularly drinks at levels that can damage their health, leading to serious, but preventable, conditions such as stroke, some cancers, depression and liver disease. Many are unaware of the harm caused, especially from drinking frequently throughout the week."
The NHS Health Check, which is available to everyone in England aged 40-74, includes an alcohol risk assessment and advice for those whose drinking may be putting their health at risk.
Analysis by Bazian
Edited by NHS Website
Links to the headlines
Mail Online, 24 July 2015
The Independent, 24 July 2015
Sky News, 24 July 2015
The Daily Telegraph, 24 July 2015
The Guardian, 24 July 2015
Daily Express, 24 July 2015
The Huffington Post, 24 July 2015
Links to the science
BMJ Open. Published online July 23 2015