UK self-selected sample: A significant proportion of individuals reported drinking more frequently in lockdown, drinking more units per drinking occasion and more frequent heavy episodic drinking

Characterising the patterns of and factors associated with increased alcohol consumption since COVID‐19 in a UK sample. Melissa Oldham et al. Drug and Alcohol Review, March 3 2021. https://doi.org/10.1111/dar.13256

Abstract

Introduction: To examine changes in drinking patterns and to assess factors associated with reported increases in frequency of drinking, units consumed and frequency of heavy episodic drinking (HED) during the UK lockdown.

Methods: Online cross‐sectional survey of 2777 self‐selected UK adults.

Results: Thirty percent of participants reported drinking more frequently in lockdown, 16% reported drinking more units per drinking occasion and 14% reported more frequent HED. For men and women, increased frequency of drinking was associated with being less likely to believe alcohol drinking would lead to greater chance of catching COVID‐19 (men: OR = 0.99, 95% CI = 0.98, 1.00; women: OR = 0.99, 95% CI = 0.99, 1.00) and deterioration in psychological wellbeing (OR = 1.27, 95% CI = 1.04, 1.54; OR = 1.29, 95% CI = 1.11, 1.51); increased unit consumption was associated with deterioration in financial situation (OR = 1.50, 95% CI = 1.21, 1.86; OR = 1.31, 95% CI = 1.05, 1.64) and physical health (OR = 1.31, 95% CI = 1.03, 1.67; OR = 1.66, 95% CI = 1.31, 2.10). Finally, increases in the frequency of HED were associated with deterioration in psychological wellbeing (OR = 1.65, 95% CI = 1.25, 2.18; OR = 1.46, 95% CI = 1.17, 1.82) and being furloughed (OR = 3.25, 95% CI = 1.80, 5.86; OR = 2.06, 95% CI = 1.19, 3.56). Other gender differences were detected, for example, living with children was associated with an increase in units consumed (OR = 1.72, 95% CI = 1.09, 2.73) and the frequency of HED (OR = 2.40, 95% CI = 1.44, 3.99) for men, but not women.

Discussion and Conclusions: In this self‐selected UK sample, a significant proportion of individuals reported drinking more frequently in lockdown, drinking more units per drinking occasion and more frequent HED. There were consistent predictors of increased consumption across men and women, but other gender differences were detected. This study identifies groups that may require targeted support in future lockdowns.

Discussion

About a one‐third of our self‐selected sample, surveyed between 30 April and 14 June 2020, reported drinking more frequently in the first UK lockdown. These rates were roughly equivalent between men and women though the correlates of increased frequency of alcohol consumption differed somewhat by gender. Deterioration in psychological wellbeing and believing that alcohol was unlikely to put them at greater risk of getting or not recovering from COVID‐19 were correlated with increased frequency of drinking for both men and women. Amongst women, increases in the frequency of drinking occasions were also associated with last‐year alcohol reduction attempts and deterioration in physical health. Whereas amongst men, increases in the frequency of drinking were associated with being younger, having a lower baseline AUDIT‐C score, being furloughed, deterioration in living conditions, deterioration in financial circumstances, improvements in social relationships and having fewer pre‐existing health conditions.

In terms of changes in the units consumed per drinking occasion, women were more likely than men to drink the same number of units as pre COVID‐19 (66% vs. 56% of men). Men were more likely than women to drink both more units (19% vs. 14%) and less units (25% vs. 20%). Deterioration in financial circumstances and physical health were associated with increased unit consumption for both men and women. Amongst women, increases in units consumed per drinking occasion were associated with having more alcohol reduction attempts in the last year. Whereas, amongst men, increases in units consumed per drinking occasion were associated with living with children, deterioration in psychological wellbeing and believing that alcohol was unlikely to put them at greater risk of getting or not recovering from COVID‐19.

Finally, the majority (61%) reported no change in the frequency of HED occasions pre‐ and post‐COVID‐19. For both women and men, being furloughed and deterioration in psychological wellbeing were associated with increases in the frequency of HED. Amongst women, increases in the frequency of HED were also associated with being younger, last‐year alcohol reduction attempts and living alone. Amongst men, increases in the frequency of HED were associated with living with children, having a more negative experience of social distancing, deterioration in financial circumstances, improvements in social relationships and being a current smoker.

Implications

This study has important implications in terms of highlighting groups that may need targeted support for alcohol reduction to counteract an increase in drinking during future COVID‐19‐related lockdowns in the UK. There were some consistencies in correlates of increased drinking amongst men and women. Deterioration in psychological wellbeing was one of the most consistent predictors of increases in the frequency of drinking and HED for both men and women. Being furloughed was also a consistent predictor of increases in HED across men and women. This is in line with other literature showing that alcohol consumption increases in economic downturns where unemployment is higher, partly due to increases in leisure time, amongst those who are unemployed [36]. These findings suggest that in future iterations of lockdown, those on furlough or similar schemes may require additional alcohol‐related support. Communications from either the government or employers around the furlough scheme could contain links to resources to help individuals manage their wellbeing and their drinking.

In line with other studies [3], there is also some evidence of gender differences in drinking patterns; units consumed per drinking occasion have polarised more amongst men in that men are more likely than women to be drinking both more and less. Furthermore, the correlates of increases in each drinking pattern are different for men and women. Living with children was associated with an increase in units consumed and the frequency of heavy episodic drinking for men, but not women. Increases in the amount of alcohol consumed on drinking occasions amongst male parents are concerning as HED in particular is likely to impair performance of caring responsibilities. This is in line with some research showing that women disproportionately carry the burdens of increased child care [23, 24], which could explain greater declines in wellbeing amongst women [25, 26].

Deterioration in financial circumstances was consistently associated with increases in all of the drinking measures for men, but only with units consumed for women. Previous research examining the relationships between economic downturns and alcohol consumption also found men were more likely to drink heavily in response to recessions and increased unemployment [36]. This may be due to increased stress in response to traditional gender roles in which men may be more likely to be considered the breadwinner.

Strengths and limitations

A key strength of this study was the variety of measures collected, permitting a detailed analysis of a broad range of potential factors predicting drinking patterns during the start of COVID‐19‐related lockdown in the UK. Furthermore, this survey allowed participants to select the date that they felt COVID‐19 started to affect them, this offers a strength over other studies that rely on using the date lockdown began to signal ‘before’ and ‘after’, in a period of ongoing change. Although the full lockdown began on 24 March, there was advice and knowledge of the virus and its effects in the months before this which may have affected drinking behaviour. For example, pubs in England were closed on 20 March and people were encouraged to work from home or socially distance from 16 March. Furthermore, the collection of data while lockdown was ongoing limits the potential for recall bias, which might be present in retrospective studies. However, this study was not without limitations. As pre‐COVID‐19 drinking is only measured post‐COVID‐19, it may be susceptible to recall bias. The sample was self rather than randomly selected, which reduces the generalisability of these results. Specifically, the results may be more reflective of people who complete online surveys about health than the general population in the UK. Black, Asian and Minority Ethnic people in particular were underrepresented in the study sample, this means that ethnicity was treated as white versus minority ethnic groups. Grouping all Black, Asian and Minority Ethnic participants together in this way does not allow examination of differences between different ethnicities and cultures, which limits the generalisability of these conclusions further. Finally, here we use a non‐validated measure of self‐assessed changes in psychological wellbeing, which may have been interpreted differently by participants.