Article Text

Download PDFPDF

The attribution of work environment in explaining gender differences in long-term sickness absence: results from the prospective DREAM study
  1. Merete Labriola1,
  2. Kari Anne Holte2,
  3. Karl Bang Christensen3,
  4. Helene Feveile4,5,
  5. Kristina Alexanderson6,
  6. Thomas Lund1
  1. 1Danish Ramazzini Centre, Department of Occupational Medicine, Regional Hospital Herning, Herning, Denmark
  2. 2Department of Social Science and Business Development, International Research Institute of Stavanger, Stavanger, Norway
  3. 3Department of Biostatistics, University of Copenhagen, Copenhagen, Denmark
  4. 4National Research Centre for the Working Environment, Copenhagen, Denmark
  5. 5Statistics Denmark, Copenhagen, Denmark
  6. 6Department of Clinical Neuroscience, Karolinska Institute, University of Stockholm, Stockholm, Sweden
  1. Correspondence to Merete Labriola, Danish Ramazzini Centre, Department of Occupational Medicine, Regional Hospital Herning, Herning 7400, Denmark; meretelabriola{at}me.com

Abstract

Objectives To identify differences in risk of long-term sickness absence between female and male employees in Denmark and to examine to what extent differences could be explained by work environment factors.

Methods A cohort of 5026 employees (49.1% women, mean age 40.4 years; 50.9% men, mean age 40.2 years) was interviewed in 2000 regarding gender, age, family status, socio-economic position and psychosocial and physical work environment factors. The participants were followed for 18 months in order to assess their incidence of long-term sickness absence exceeding 8 consecutive weeks.

Results 298 workers (5.9%) received sickness absence compensation for 8 weeks or more. Women had an excess risk of 37% compared to men, when adjusting for age, family status and socio-economic position. Physical work environment exposures could not explain this difference, whereas differences in psychosocial work environment exposures explained 32% of the differences in risk of long-term sickness absence between men and women, causing the effect of gender to become statistically insignificant. The combined effect of physical and psychosocial factors was similar, explaining 30% of the gender difference.

Conclusion Differences in psychosocial work environments in terms of emotional demands, reward at work, management quality and role conflicts, explained roughly 30% of women's excess long-term sickness absence risk. Assuming women and men had identical working conditions would leave the larger part of the gender difference in long-term sickness absence from work unexplained.

  • Register data
  • sickness absence
  • men
  • women
  • work environment
  • epidemiology
  • gender
  • longitudinal studies
  • sickness absence
  • international occupational health

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

What this paper adds

  • Although women in western countries have higher rates of long-term sickness absence than men, the reasons for this difference are not fully known.

  • No studies have investigated work environment exposures as a cause of this difference in a general working population sample.

  • Differences in psychosocial work environment explained roughly 30% of women's excess long-term sickness absence risk; physical work environment exposures were not relevant.

  • Emotional demands, reward at work, management quality and role conflicts are especially important in explaining the gender gap in long-term sickness absence risk.

Introduction

It has been demonstrated that women have a higher risk of sickness absence than men,1–6 and gender is often mentioned as an important determinant of sickness absence. A few studies have focussed on the effect that measures reflecting working conditions have on this difference.1 2 7–9 However, only Laaksonen and colleagues1 and Smeby and colleagues2 have specifically addressed the work environment as a potential cause of gender differences in long-term sickness absence, using data from a cohort of municipal employees in Helsinki and from a Norwegian urban population, respectively. Long-term sickness absence has been labelled an emerging public health problem,10 11 and longer absence periods exceeding 3–4 weeks in one study12 and 8 or more weeks in another13 have been shown to affect future labour market attachment in terms of disability pension. No studies so far have examined to what extent excess risk for longer sickness absence spells among women in the general working population can be explained by work environment exposures. The aim of this study was to identify differences in the risk of long-term sickness absence defined as 8 or more weeks of sickness absence, in a population sample based on the general population of Denmark, and to examine to what extent differences could be explained by, potentially changeable, work environment factors.

Methods

A random sample of 11 437 people living in Denmark in 2000 were contacted and 8583 (75%) responded. Of these, 5366 were 18–64 years of age and had worked as employees for at least 2 months before the questionnaire-based baseline interview. Data collected included information on gender, age, socio-economic position and work environment in 2000.5 Data on long-term sickness absence, defined as receiving sickness absence compensation for at least 8 consecutive weeks during 18 months of follow-up, were obtained by linkage to a national register of social transfer payments (DREAM).5

Psychosocial work environment factors were measured with 14 items, combined into five scales measuring emotional demands, demands for hiding emotions, management quality, role conflicts and reward. These five dimensions were selected as previous studies on the same population have shown them to be significantly associated with increased risk of long-term sickness absence.5 Physical work environment factors were assessed using nine questions, combined into four indices measuring extreme bending or twisting of the neck or back, working mainly standing or squatting, physical workload in terms of lifting or carrying loads, and pushing or pulling loads. These four indices were selected as previous studies on the same population have shown them to be significantly associated with increased risk of long-term sickness absence.14 Family status was assessed using a variable with four categories (single without children, single with children, cohabitating without children, cohabitating with children). Based on employment grade, job title and education, respondents were classified into five socio-economic groups; I: executive managers and/or academics; II: middle managers and/or 3–4 years of further education; III: other white-collar workers; IV: skilled blue-collar workers; and V: semi-skilled or unskilled workers.

Overall, 5026 subjects had complete data and were included in the study: 2560 men with a mean (SD) age of 40.2 (11.4) years and 2466 women with a mean (SD) age of 40.4 (11.0) years.

The outcome long-term sickness absence was defined as the first event of 8 or more consecutive weeks of sickness absence in the follow-up period from 1 January 2001 to 30 June 2002.

The Cox proportional hazards model was used to calculate HRs and 95% CIs. Those who died, emigrated or retired were censored, but were considered to be under risk until the time of death, emigration or retirement. Those receiving another social benefit, for example those on maternity leave, were not considered to be under risk during that period. The analysis was conducted in five steps. First, HRs adjusted for age and family status were computed. In a second step HRs were adjusted for age, family status and socio-economic position. A third step introduced the four indices of physical work environment exposures into the model. A fourth model step substituted physical work environment exposures with the five scales measuring psychosocial work environment exposures, while the fifth and last step introduced physical and psychosocial work environment factors combined. All analyses were carried out using SAS v 9.1.

Results

During the 18-month follow-up period, 127 men (5.0%) and 171 women (6.9%), giving a total of 298 subjects (5.9%), received sickness absence compensation for 8 weeks or more.

Table 1 shows the HRs of sickness absence of 8 weeks or more for women and men during an 18-month follow-up period.

Table 1

Hazard ratios for onset of sickness absence of 8 or more consecutive weeks during a 18-month follow-up period

In the initial model 1, taking into account the effects of age and family status, women had an excess risk for long-term sickness absence of 41% (HR 1.41, 95% CI 1.12 to 1.78). Adjusting further for socio-economic position caused the estimate to decrease by 10% (HR 1.37, 95% CI 1.08 to 1.75). Introducing the four indices for physical work environment exposures caused a 11% increase in the estimate for the effect of gender, yielding an HR in model 3 similar to that in model 1 (HR 1.41, 95% CI 1.10 to 1.80). A similar step, substituting the physical work environment indices with the five scales assessing psychosocial work environment exposures, caused a 32% decrease in the effect of gender from model 2 to model 4, leaving the differences in risk of long-term sickness absence between men and women statistically insignificant (HR 1.25, 95% 0.98 to 1.60). A final model step 5, estimating the combined effect of psychosocial and physical work environment exposures, caused a 30% decrease in the effect of gender compared to model 2 (HR 1.26, 95% CI 0.98 to 1.63). Removing sickness absence possibly related to pregnancy did not affect the estimates of sickness absence according to gender, or the explanatory effect of work on gender differences in sickness absence (not shown).

Discussion

After controlling for age, family status and socio-economic position, this study found a 37% excess risk among women of experiencing a period of 8 or more consecutive weeks of sickness absence during a 18-month follow-up period. Physical work environment factors did not explain this difference but did tend to increase gender differences slightly. Psychosocial work environment factors in terms of measures capturing emotional demands, reward at work, management quality and role conflicts explained 32% of the excess risk, and physical and psychosocial factors combined explained 30% of the excess risk for absences of 8 weeks or longer.

The included measures of work environment do not encompass all dimensions of work environment, and consequently a further effect of unmeasured work environment factors cannot be ruled out. If working conditions are not measured properly, their effects are likely to be underestimated. This could be the case in the study by Smeby and colleagues, where excess risk of long-term sickness absence among women could not be explained by work-related factors, general health or mental distress.2 Another reason for this finding could be differences in the definition of sickness absence: Smeby and colleagues define sickness absence as all spells exceeding 16 days, whereas the present study focuses on sickness absence periods exceeding 8 weeks. In any case, the present study disagrees with the concluding remarks by Smeby et al that “[f]actors explaining the gender divide should be sought elsewhere [than in the workplace]”. Even after taking into account socio-economic position, which also captures job function, psychosocial work environment factors explained 32% of gender differences in risk of long-term sickness absence. Physical work factors could not explain the gender difference, but did seem to increase the gap slightly. This is in line with the findings of Mastekaasa et al.9 A possible explanation for this trend is the fact that men are more exposed to hard physical work than women.14

More directly compatible as regards outcome are the findings by Laaksonen and colleagues who defined long-term sickness absence as periods exceeding 60 days, which is comparable to the outcome of this study.15 However, Laaksonen et al found no excess risk among women for these longer spells, and consequently no analysis to explain a gender difference was carried out. The findings of the present study disagree with those of Laaksonen et al, possibly because their study was based on a population of public sector employees from one city and only included subjects 40 years of age or older, whereas the present study was based on a sample of the general working population, and encompassed all occupations, a broader age span and a less skewed gender distribution (the municipal sectors are typically female-dominated in the Nordic countries). Laaksonen et al found gender differences for shorter absence spells, but these differences could not be attributed to work environment exposures. However, the measures of work exposure in the analyses underlying the conclusions by Laaksonen et al are crude compared to those in the present study.

In another study, Laaksonen and colleagues found that differences between occupations held by women and men explained a substantial proportion of the female excess in sickness absence, whereas workplace had little impact when differences in occupations were taken into account. A considerable part of the occupation effect was explained by vertical differences in social class between male and female occupations.1 However, as the authors themselves note, attributing the entire impact of occupation to working conditions is a profound overestimation. It seems that the choice between using occupation level data or individual level exposure data is a question of either over- or underestimating the effect of working conditions on gender differences in sickness absence. Furthermore, the explanatory value of occupational group is limited by the fact that generalisation of findings from one country to another can be problematic: work content, working conditions and work organisation might have very specific characteristics for an occupation in one country, but different characteristics in the ‘same’ occupation in another country.

When seeking explanatory factors outside the workplace, one has to consider differences in biology, for example reproduction-related health problems.7 Removing sickness absence that might be related to pregnancy did not affect the estimates of sickness absence according to gender, or the explanatory effect of work on gender differences in sickness absence. This implies that the difference in long-term sickness absence risk between men and women was not related to long-term sickness absence due to pregnancy. Also related to differences in biology are gender related reactions to work related exposures, for example shift work. There are indications that female shift workers are more likely to report menstrual irregularity and longer menstrual cycles compared to non-shift workers.16 However, the research on reproduction health and sickness absence is inconclusive.7 The present study is controlled for age, family status and socio-economic position. However, family status does not provide any information about how domestic responsibilities that might cause differences in non-work related strain, are shared within families. For women, the double exposure of home and work responsibilities is indicated as a risk factor for common health symptoms,17 but is unlikely to cause health problems leading to at least 8 weeks of work disability.

The present study does not measure the effect of biology, reproductive health problems or non-work related strain, but the overall conclusion that approximately 30% of gender differences in risk of long-term sickness absence could be explained by differences in psychosocial work environment, more specifically emotional demands, reward at work, management quality and role conflict, indicates that 70% of the gender difference in risk of long-term sickness absence is related to factors that lie outside the workplace.

References

Footnotes

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.