We recently conducted research into the effect of alcohol on employment and social outcomes in the UK.
Associations between alcohol and health have been extensively studied (Rehm, Baliunas et al. 2010). Less well understood is the impact of alcohol on employment. Evidence is mixed and effect directions are unclear (Thompson and Pirmohamed 2021, Mangot-Sala, Smidt et al. 2022). Drinking can influence employment outcomes. For instance, it affects people’s ability to hold down a job. There is also evidence supporting effects in the opposite direction, i.e. personal circumstances influencing alcohol consumption. In addition, alcohol-employment associations may be driven by shared upstream factors. For instance, economic recession might cause both unemployment and changes in drinking habits. It is likely all these mechanisms operate together to varying degrees and over different time scales (Hoffmann, Kröger et al. 2019).
Methods
We wanted to find out how alcohol use affects people’s employment patterns, income and deprivation. That’s tricky, because standard research methods often mix up cause and effect – they can’t easily separate whether drinking affects employment, or vice versa. To get around this, we used a method called Mendelian Randomisation (MR). This can be thought of as a way of configuring a natural experiment from observational data (Labrecque and Swanson 2019). This allows a clearer picture of cause and effect, even when other factors present might blur the results.
For our study, we looked at data from over 230,000 people aged 40+ in the UK Biobank (Bycroft, Freeman et al. 2018), all of White British background. We tested how different drinking measures (units per week, and two measures of problematic drinking) influenced nine outcomes related to employment, deprivation, income and education.
Results
Our findings suggested that higher alcohol consumption and problematic drinking increased the risk of living in a more deprived area and may have deleterious effects on employment and income. These effects were more detrimental for men than women. In men, increased alcohol consumption decreased income, increased the risk of not being in paid employment and increased the risk of unemployment. In women, increased alcohol consumption decreased the risk of not being in paid employment, retirement and caring for home/family.
Conclusions
Alcohol plays an important social role in many societies. In the UK, alcohol consumption remains high, 18% of men report drinking more than 8 units on their heaviest drinking day in a week (13% of women exceeded 6 units) (Office for National Statistics UK 2018). Our findings support alcohol consumption having considerable adverse impact by increasing deprivation and deleteriously affecting employment. As drinking behaviour is modifiable, there is potential for substantial reduction in alcohol’s health, social and economic costs through public health initiatives.
Our finding that effects were more detrimental for men than women is intriguing. This suggests mechanisms may differ for men and women. Being flippant, one could say women drink to socialise (and this helps their employability), whereas men socialise to drink. However, another explanation is that effects differ across consumption levels. Low levels of drinking may have beneficial effects on the outcomes we looked at, whereas heavy drinking may have detrimental effects. This would also fit with our findings as women drink markedly less than men.
Our study should be considered in the light of potential limitations. Alcohol consumption is strongly socially patterned and results based on white 40-70 year old members of the UK Biobank cohort may not generalise well to other UK population strata, to non-UK populations or to other eras. Despite the large sample size, our MR analyses were weakly powered (indeed our analyses were restricted to White British for this reason). Low power makes our findings more vulnerable to bias arising from assumption violations.
Written by Dr Desmond Campbell, Research Associate, Health Economics and Health Technology Assessment, University of Glasgow.
All IAS Blogposts are published with the permission of the author. The views expressed are solely the author’s own and do not necessarily represent the views of the Institute of Alcohol Studies.