BMJ Open (May 2023)

Latent classes of unhealthy behaviours and their associations with subsequent sickness absence: a prospective register-linkage study among Finnish young and early midlife employees

  • Jouni Lahti,
  • Tea Lallukka,
  • Ossi Rahkonen,
  • Anne Kouvonen,
  • Jatta Salmela,
  • Noora Kanerva

DOI
https://doi.org/10.1136/bmjopen-2022-070883
Journal volume & issue
Vol. 13, no. 5

Abstract

Read online

Objectives Unhealthy behaviours are associated with increased sickness absence (SA), but few studies have considered person-oriented approach in these associations. Using latent class analysis, we examined clustering of unhealthy behaviours among Finnish municipal employees and their associations with subsequent SA.Design A prospective register-linkage study.Setting Unhealthy behaviours (low leisure-time physical activity, non-daily fruit and vegetable consumption, insufficient sleep, excessive alcohol use and tobacco use) were derived from the Helsinki Health Study questionnaire survey, collected in 2017 among 19- to 39-year-old employees of the City of Helsinki, Finland.Participants A total of 4002 employees (81% women) of the City of Helsinki, Finland.Primary outcome measures The questionnaire data were prospectively linked to employer’s SA register through March 2020. Associations between latent classes of unhealthy behaviours and subsequent SA (1–7 days/8+ days/all lengths) were examined using negative binomial regression.Results Among women, a three-class latent class model was selected: (1) few unhealthy behaviours (84%), (2) excessive alcohol and tobacco use (12%) and (3) several unhealthy behaviours (5%). Women belonging to classes 2 and 3 had increased SA rates compared with those in class 1, regardless of the length of SA spells. Among men, a 2-latent class model was selected: (1) few unhealthy behaviours (53%) and (2) several unhealthy behaviours (47%). Men belonging to class 2 had increased rates of 1–7 days’ SA compared with men in class 1.Conclusions This study suggests that preventive actions aiming to reduce employees’ SA should consider simultaneously several unhealthy behaviours. Targeted interventions may benefit of identifying the clustering of these behaviours among occupational groups.