Sociodemographic Differences Between Alcohol Use and Sickness Absence: Pooled Analysis of Four Cohort Studies

: Jenni Ervasti, Mika Kivimäki, Jenny Head, Marcel Goldberg, Guillaume Airagnes, Jaana Pentti, Tuula Oksanen, Paula Salo, Sakari Suominen, Markus Jokela, Jussi Vahtera, Marie Zins, Marianna Virtanen

Publisher: OXFORD UNIV PRESS

: 2018

: Alcohol and Alcoholism

: Alcohol and alcoholism (Oxford, Oxfordshire)

: Alcohol Alcohol

: 53

: 1

: 95

: 103

: 9

: 1464-3502

DOI: https://doi.org/10.1093/alcalc/agx079

The U-shaped alcohol use-sickness absence association is more consistent in men than women. Abstinence is a risk factor for sickness absence among people with higher rather than lower SES. Healthy worker effect and health selection may partly explain the observed differences.
Among men (n = 17,285), abstainers (6%), former (5%), current (5%) and consistent (7%) at-risk drinkers had an increased risk of sickness absence compared with consistent low-risk drinkers (77%). Among women (n = 29,229), only abstainers (12%) had a higher risk of sickness absence compared to consistent low-risk drinkers (74%). After adjustment for lifestyle and health, abstaining from alcohol was associated with sickness absence among people with intermediate and high SES, but not among people with low SES.
In a pooled analysis from four cohort studies from three European countries, we demonstrated a U-shaped association between alcohol use and sickness absence, particularly among men. Abstinence from alcohol was associated with increased sickness absenteeism among both sexes and across socioeconomic strata, except those with low SES.
We examined differences in sickness absence in relation to at-risk drinking and abstinence, taking into account potential changes in consumption.
We used individual-participant data (n = 46,514) from four prospective cohort studies from Finland, France and the UK. Participants responded to a survey on alcohol use at two time points 4-6 years apart, and were linked to records of sickness absence for an ~6-year follow-up after the latter survey. Abstainers were those reporting no alcohol use in either survey. At-risk drinkers at T1 were labelled as 'former', at-risk drinkers at T2 as 'current' and at-risk drinkers at both times as 'consistent' at-risk drinkers. The reference group was low-risk drinkers at both times. Study-specific analyses were stratified by sex and socioeconomic status (SES) and the estimates were pooled using meta-analysis.
Conclusions
Results
Short summary
Aims
Methods