Objectives This study aims to investigate the role of lifestyle factors in relation to the presence and degree of productivity loss at work and sick leave.
Methods A cross-sectional study recruited 10 624 workers in 49 companies in the Netherlands in 2005–2009. Productivity loss at work was measured on a 10-point scale indicating how much work was actually performed on the previous workday. Sick leave was measured by asking how many days in the past 12 months workers were off work due to health problems. Logistic regression analyses were applied to study the association between obesity and lifestyle behaviours and both outcome measures.
Results Obesity was associated with the presence of sick leave (OR 1.25) and prolonged duration (OR 1.55). Insufficient physical activity (OR 1.12) and smoking (OR 1.17) were also associated with the presence of sick leave. Smoking (OR 1.45), obesity (OR 1.29) and insufficient fruit and vegetable intake (OR 1.22) were associated with the degree of productivity loss at work. The combined population attributable fractions of lifestyle factors for sick leave and the higher levels of productivity loss at work were above 10%.
Conclusions Lifestyle-related factors, especially smoking and obesity, were associated with the presence and duration of sick leave and degree of productivity loss at work. More than 10% of sick leave and the higher levels of productivity loss at work may be attributed to lifestyle behaviours and obesity. Hence, primary interventions on lifestyle may have a noticeable contribution to maintaining a productive workforce.
- Productivity loss
- sick leave
- cross sectional studies
- sickness absence
Statistics from Altmetric.com
What this paper adds
This study investigates the role of lifestyle factors in relation to the presence and degree of productivity loss at work and sick leave.
Lifestyle-related factors, especially smoking and obesity, are associated with the presence and duration of sick leave and the level of productivity loss at work.
More than 10% of sick leave and the higher levels of productivity loss at work may be attributed to lifestyle behaviours and obesity.
Primary interventions on lifestyle may contribute to maintaining a productive workforce.
With ageing populations, there is a need to keep the ageing workforce healthy and productive. The American College of Occupational and Environmental Medicine moved the health agenda forward by focussing on prevention in the workplace.1 Many employers offer health promotion programmes to their workers, and it has been estimated that over 80% of worksites with 50 or more employees and almost all large employers offer some kind of health improvement programme.2 Obesity and unhealthy lifestyle behaviours are increasingly being linked with productivity loss at work and sick leave, which lead to elevated indirect costs.3–5 Productivity loss at work due to impaired health has an impact on future sick leave and on future general health.6 7 Health-related determinants of sick leave and productivity loss at work are well studied,4 8 9 but the relationship with lifestyle factors is less clear.
A higher prevalence of sick leave was found among obese workers,10–12 and obesity also predicted long-term sick leave.13 The association between obesity and productivity loss at work is less convincing, with some cross-sectional studies reporting a positive association between obesity and productivity loss at work,14–16 whereas other studies did not find any association.17 18 The findings on whether workers who have a physically active lifestyle have less sick leave and higher productivity at work are inconsistent. Some studies did not find an association,14 17 while others reported an inverse association between physical activity and sick leave19 or productivity loss at work.16 In a systematic review, limited evidence was found for the effectiveness of physical activity programmes in the workplace regarding sickness absence, while no evidence was found for productivity loss at work.20 For smoking and productivity loss at work, both positive21 and negative associations17 have been reported. Possible explanations for these contradictory findings are the differences in outcome measurements. Often studies only focus on the presence of productivity loss at work and sick leave, ignoring the level of productivity loss at work and sick leave duration. The assessment of productivity loss at work is described as an important issue to be addressed in research.22
It could be hypothesised that obese workers and workers with unhealthy lifestyle behaviours have longer sick leave durations and a higher level of productivity loss at work compared to workers with a healthy body weight and healthy lifestyle behaviours. In order to investigate whether lifestyle factors are associated with the presence and degree of productivity loss at work and sick leave, a large cross-sectional study among various companies in the Netherlands was conducted. This study aims to investigate: (1) the role of lifestyle factors in relation to the presence of productivity loss at work and sick leave; (2) the associations of lifestyle factors with different levels of productivity loss at work and sick leave durations;, and (3) whether similar lifestyle and health factors are related to productivity loss at work as well as sick leave.
The study population consisted of workers in 49 companies in the Netherlands in 2005–2009. Companies from a wide range of sectors participated, that is commercial services (41%), non-commercial services (37%), industrial manufacturers (18%) and construction (4%). These companies had commissioned an occupational health organisation to launch a programme to investigate the work ability of their workforces and as part of this programme a questionnaire survey was conducted on lifestyle factors, health, work demands, work ability and productivity. Companies participating in this programme invited all their workers to participate. The occupational health organisation sent an invitation to all eligible workers by regular mail, and provided workers with an individualised password to fill out the questionnaire on a secure website. Written informed consent was obtained from all participants at the time of enrolment. Complete data on productivity loss at work, sickness absence and lifestyle factors were collected for 10 624 workers and made available to the Erasmus Productivity Loss at Work database (ELPW database). The response rate varied from 9% to 95% across companies with an overall response rate of 57%.
Productivity loss at work
The main outcome of this study was productivity loss at work, measured with the quantity scale of the Quantity and Quality (QQ) method.23 Respondents were asked to indicate how much work they actually performed during regular hours on their most recent regular workday as compared with normal. The quantity of productivity was measured on a scale from 0 (nothing) to 10 (regular quantity). The outcome was dichotomised so that those scoring 0–9 were classified as reporting productivity loss at work and compared with those who scored a 10. Individuals reporting productivity loss at work were further categorised into three levels (10%, 20% and ≥30% productivity loss at work corresponding to the scores 9, 8 and 0–7 on the 10-point scale).
Sick leave was derived from the Work Ability Index (WAI).24 Participants were asked to indicate on a 5-point ordinal scale on how many days in the past 12 months they were not able to work due to health problems. A dichotomous variable for sick leave (yes/no) was created. Those individuals reporting sick leave in the past 12 months were further categorised into three sick leave categories: 1–9 days, 10–24 days and 25 or more days.
Lifestyle and health factors
Height and weight were collected by questionnaire. Body mass index (BMI) was calculated by dividing body weight in kilograms by the square of body height in metres and used to define subjects as normal weight (BMI <25 kg/m2), overweight (BMI 25–30 kg/m2) or obese (BMI ≥30 kg/m2). Physical activity, fruit and vegetable intake, smoking and alcohol consumption were measured with single yes/no questions. Those who reported they were physically active for at least 30 min per day were considered to be in agreement with the recommendation on moderate physical activity.25 Fruit and vegetable intake was defined as eating fruit and vegetables 5 days per week or more, as a proxy estimate for compliance with the guidelines for sufficient fruit and vegetable intake.26 Smoking was assessed as current smoking status, and alcohol use as drinking 10 or more glasses per week. The number of diagnosed diseases was derived from the WAI, and measured using a list of 13 broad categories of diseases ever diagnosed by a physician, and categorised into no disease, the presence of one disease, and the presence of two or more diseases.24
Work demands were estimated based on the expert judgement of the employer. Each job title in the study population was characterised as a job with primarily physically demanding activities or as a job with primarily mental tasks. This job exposure matrix was linked to job title at the individual level, and, hence, for each worker his/her job was categorised into a mentally or physically demanding job.
Descriptive statistics were used to describe the characteristics of the study population. Cohen's κ was calculated to estimate the association between sick leave and productivity loss at work. Logistic regression analyses were used to explore associations between the dependent variables presence of productivity loss at work and presence of sick leave and the independent variables individual characteristics, lifestyle and health factors and work demands. The odds ratios (OR) were estimated as a measure of association with corresponding 95% confidence intervals (95% CI). First, univariate associations were explored, and all variables that were statistically significant (p<0.05) in the univariate analyses were investigated in a multivariate analysis. To increase comparability between the models for productivity loss at work and sick leave, any variable with a statistically significant association with one outcome in the univariate analysis was also included in the multivariate model for the other outcome. The absence of productivity loss at work and the absence of sick leave were defined as reference categories. Second, multinomial logistic regression analyses were performed to study the associations between the individual, lifestyle and health factors, and work demands with the degree of sick leave or productivity loss at work. This analysis was restricted to employees with sick leave or productivity loss at work. The lowest levels of productivity loss at work (10%) and sick leave (1–9 days) were defined as reference categories. All analyses were carried out in 2009 with the Statistical Package for Social Sciences v 15.0 for Windows. Population attributable fractions (PAFs) were calculated for significant lifestyle factors related to productivity loss at work or sick leave, using the formula PAF=Pe(OR−1)/(1+Pe(OR−1)), in which Pe is the prevalence in the study population.
Table 1 shows that 44% of subjects reported productivity loss at work during the previous workday and 56% lost at least 1 work day because of sick leave in the past 12 months. No association at the individual level was found between productivity loss at work and sick leave (Cohen's κ=0.07). The mean age of the study population was 43.8 years (±9.9 years), ranging from 18 to 68 years. The mean BMI of the respondents was 25.4 kg/m2 (±3.9 kg/m2). All lifestyle factors were inter-related, except physical activity and alcohol intake. The prevalence of the different disease categories ranged from 3% (hereditary diseases) to 77% (musculoskeletal disorders). BMI was associated with the presence of disease: 69% of normal weight workers had at least one disease, 75% of overweight workers had at least one disease, and 83% of obese workers had at least one disease. More workers in physically demanding jobs were obese compared to workers with mentally demanding jobs (13% vs 9%).
Determinants of the presence of productivity loss at work and sick leave
Lifestyle factors were associated with both the occurrence of productivity loss at work and sick leave (tables 2 and 3). Obese workers reported sick leave more often than workers with a normal body weight (OR 1.27, 95% CI 1.11 to 1.46). No statistically significant association was found between obesity and productivity loss at work (OR 1.05, 95% CI 0.92 to 1.19). Insufficient physical activity was associated with productivity loss at work (OR 1.08, 95% CI 1.00 to 1.17) and with sick leave (OR 1.19, 95% CI 1.10 to 1.29) in the univariate analyses, but in the multivariate analysis remained statistically significant only for sick leave (OR 1.12, 95% CI 1.03 to 1.21). The presence of sick leave in the past 12 months was also associated with smoking, the presence of diseases, younger age and working in a mentally demanding job.
The presence of productivity loss at work was associated with insufficient fruit and vegetable intake, drinking 10 or less glasses of alcohol per week, the presence of diseases and younger age. The PAFs for the presence of sick leave due to obesity, insufficient physical activity and smoking were 2.7%, 4.5% and 4.1%, respectively, resulting in a combined PAF of 10.9%. The PAF for the presence of productivity loss at work due to insufficient fruit and vegetable intake was 3.9%.
Determinants of the level of productivity loss at work
Table 4 shows the multivariate associations with the level of productivity loss at work. Obesity (OR30% 1.29, 95% CI 1.00 to 1.65) and insufficient fruit and vegetable intake (OR30% 1.22, 95% CI 1.04 to 1.43) were associated with more productivity loss at work. Smoking was associated both with 20% productivity loss at work (OR20% 1.25, 95% CI 1.04 to 1.50) and with 30% or more productivity loss (OR30% 1.45, 95% CI 1.21 to 1.73). Compared with the univariate analyses there were modest changes in these estimates (<10%) in the multivariate analysis. The combined PAF for obesity, insufficient fruit and vegetable intake and smoking increased with the degree of productivity loss at work, to 7.4% for 20% productivity loss and 18.6% for 30% or more productivity loss at work.
Determinants of the duration of sick leave
Table 5 shows the multivariate associations with sick leave duration. Among obese workers, sick leaves of 10–24 days (OR10-24 1.66, 95% CI 1.33 to 2.07) and 25 days or more (OR25+ 1.55, 95% CI 1.22 to 1.95) were more prevalent than among individuals with a normal body weight. Smoking was associated with 10–24 days of sick leave (OR10-24 1.30, 95% CI 1.10 to 1.53), whereas drinking more than 10 glasses of alcohol per week was inversely associated with 25 or more days of sick leave (OR25+ 0.70, 95% CI 0.55 to 0.89). Except for obesity (with a decreased OR of 21% for 10–24 days of sick leave, and 27% for 25 or more days), there were modest changes (<10%) in the odds ratios of the lifestyle factors after adjustment for each other and other factors. The combined PAF for overweight, obesity and smoking was 20.3% for 10–24 days off work due to health problems, and 13.5% for 25 or more days of sick leave.
Lifestyle factors as well as health factors were associated with the presence of sick leave and productivity loss at work. Obesity and smoking were associated with a higher level of productivity loss at work and with more days off work due to health problems. The combined population attributable fractions for sick leave and for a higher level of productivity loss at work due to overweight and lifestyle behaviours were above 10%.
In our study population, the prevalence of obesity among men (11.2%) was similar to the estimated prevalence of obesity in the Dutch population, but for women a slightly lower prevalence was observed (10.4% vs 12.4%). Obesity was associated with both the presence and duration of sick leave. Although it was not associated with the presence of productivity loss, within the productivity loss group relatively more obese workers than workers of normal body weight had the highest level of productivity loss. No consistent associations were found for workers with a BMI between 25 and 30 kg/m2. Our finding that obesity was associated with sick leave is consistent with a recent systematic review concluding that there is strong evidence that these workers are at increased risk for taking sick leave.11 Recent reviews and studies also showed more long-term sick leave in obese workers.10 11 13 For the lifestyle behaviours, modest associations with productivity loss at work and sick leave were found. Our results are comparable with the findings of a recent prospective study on health-related behaviour and sickness absence reporting the strongest associations for smoking and obesity.12 The lifestyle factors were inter-related, but odds ratios for the lifestyle behaviours in the multivariate analyses were comparable with the odds ratios in the univariate analyses. More than 10% of the sick leave and higher levels of productivity loss at work can be attributed to obesity and unhealthy lifestyle behaviours. Among workers with longer sick leave durations and more productivity loss at work, lifestyle factors became increasingly important. Hence, primary interventions on lifestyle may make a noticeable contribution to maintaining a productive workforce.
There are indications that interventions can counterbalance the negative influence of lifestyle factors on sick leave. In a meta-analysis it was found that some workplace physical activity interventions can improve worksite outcomes such as sick leave.27 In a literature review, Matson Koffman and colleagues found that employers who invested in comprehensive worksite health promotion programmes can improve cardiovascular health in employees and yielded a US$3 to US$6 return on investment for each dollar invested over a 2- to 5-year period.28
Although we found no association between smoking and the presence of productivity loss at work, in the multinomial analyses the association between smoking and the level of productivity loss at work was stronger for the higher levels of productivity loss at work. In the literature contradictory findings have been reported, with both positive and negative associations for smoking with productivity loss at work.17 21 These differences could be due to focus on the presence of productivity loss at work, as shown in an earlier study of a subset of the available data,17 versus an analysis on the degree of productivity loss at work.21
An important question is which employee populations benefit the most from worksite health promotion interventions targeting weight and unhealthy behaviours.29 In this study longer sick leave duration and a higher level of productivity loss at work were found among workers with physically demanding jobs. However, for obese workers no consistent higher productivity loss at work or longer sick leave duration was found among obese workers in physically demanding jobs compared to obese workers in mentally demanding jobs (data not shown). Schulte and colleagues found several, mostly cross-sectional, studies showing an association between work-related factors such as job stress and shift work and obesity.30 A recent study showed that workers with a BMI above 35 kg/m2 experienced more difficulty with job-related physical tasks than participants with overweight or mild obesity.15 Psychosocial work demands were also found to be related to lifestyle behaviours and obesity.29 Based on the results in the current study, the prevention of unhealthy lifestyle behaviours and obesity are important in both physically and mentally demanding jobs. The modifying role of lifestyle and obesity in relationships between work-related factors and sick leave and productivity loss at work remains an interesting topic for future research.
There was no correlation between productivity loss at work and sick leave. Productivity loss at work was not measured as health-related productivity loss, but as overall productivity loss at work in order to avoid response bias. As shown in the tables, health factors were more strongly related to sick leave than to productivity loss at work. There is evidence that workers who become sick or feel a little sick, first go to work and may suffer productivity loss at work because of reduced health. This might be an explanation for the weaker associations of lifestyle and health factors with productivity loss at work compared to the same associations with sick leave.
A major limitation of this study is the cross-sectional study design, which does not permit further explanation with respect to causality. As mentioned before, obesity was found to predict long-term sick leave.11 For the other lifestyle factors it is difficult to judge causality. Another limitation is the reliance on self-reported productivity loss at work and sick leave, partly due to lack of instruments to measure productivity loss objectively, especially for knowledge-based occupations.31 The method used for the assessment of productivity loss at work showed significant correlations between self-reported productivity and actual work output (r=0.48) among floor layers.32 A disadvantage of this method is that productivity is assessed during the previous regular workday and does not take into account the expected fluctuations in productivity loss within workers across workdays. This unknown daily fluctuation will have contributed to random measurement error and, thus, attenuated the observed associations. A third limitation is the method for the assessment of work demands at job level and the assessment of weight and height. Work demands were estimated based on the expert judgement of employers and not by self-report of the employees or by a panel of expert raters. This might cause some misclassification. Both weight and height were based on self-reports, and height has been found to be over-reported by both men and women.33 However, as Spencer and colleagues34 concluded, self-reported height and weight data are valid for identifying relationships in epidemiological studies. A fourth limitation is the variability in response levels across companies. The response level was lower in large companies, in commercial services companies and among blue collar workers. However, using a cut-off of 80% response, no significant differences were found in sick leave and productivity loss at work between companies with high and low response levels, and response level was also not statistically significant when included in the univariate analyses. Therefore, we think that this source of selection bias will not have influenced the results to a major extent. Finally, the definitions of insufficient fruit and vegetable intake and of alcohol use do not correspond with the current national recommendations. Since drinking one (women) or two glasses (men) of alcohol per day is most likely beneficial to health, this could be an explanation for the inverse association with the presence of productivity loss at work and with long-term sick leave. For sick leave it has been reported that sickness absence was most common among heavy drinkers, and least common among moderate drinkers.12 Because educational level and physical or mental job demands are closely related, we did not adjust for education level in the analyses.
In conclusion, lifestyle behaviours, and especially smoking and obesity, were associated with the presence and duration of sick leave and the level of productivity loss at work. More than 10% of sick leave and the higher levels of productivity loss at work may be attributed to lifestyle behaviours and obesity. Hence, primary interventions on lifestyle may have a noticeable contribution to maintaining a productive workforce.
Funding This study was partly funded by ZonMw, The Netherlands Organization for Health Research and Development (project number 62300039).
Competing interests Jan Plat is employed at PreventNed, which collected the questionnaires as described in this paper. All analyses for the article were supervised and performed by personnel who are not part of PreventNed.
Provenance and peer review Not commissioned; externally peer reviewed.
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.