Skip Navigation LinksHome > January 2014 - Volume 56 - Issue 1 > Sick Leave Days and Costs Associated With Overweight and Obe...
Journal of Occupational & Environmental Medicine:
doi: 10.1097/JOM.0000000000000065
Original Articles

Sick Leave Days and Costs Associated With Overweight and Obesity in Germany

Lehnert, Thomas MSc; Stuhldreher, Nina MSc; Streltchenia, Pawel MSc; Riedel-Heller, Steffi G. MD, MPH; König, Hans-Helmut MD, MPH

Free Access
Supplemental Author Material
Continued Medical Education
Article Outline
Collapse Box

Author Information

From the Department of Health Economics and Health Services Research (Mr Lehnert, Ms Stuhldreher, Ms Streltchenia, and Dr König), Hamburg Center for Health Economics, University Medical Center Hamburg–Eppendorf, Hamburg, Germany; IFB Adiposity Diseases (Mr Lehnert and Dr Riedel-Heller), University Medicine Leipzig, and Department for Social Medicine, Occupational Medicine, and Public Health (Mr Riedel-Heller), University of Leipzig, Leipzig, Germany.

Address correspondence to: Thomas Lehnert, MSc, Department of Health Economics and Health Services Research, Hamburg Center for Health Economics, University Medical Center Hamburg–Eppendorf, Martinistr 52, D-20246 Hamburg, Germany (

This work was supported by the Federal Ministry of Education and Research (BMBF), Germany, FKZ: 01EO1001.

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Web site (

Authors Lehnert, Stuhldreher, Streltchenia, Riedel-Heller, and Konig have no relationships/conditions/circumstances that present potential conflict of interest.

The JOEM editorial board and planners have no financial interest related to this research.

Collapse Box


Objective: To analyze the impact of body mass index on sick leave days and related costs in Germany.

Methods: Cross-sectional analysis of German Socio-Economic Panel data (n = 7990). The relationship between body mass index class and sick leave days was analyzed via analyses of variance (ANOVA) (bivariate) and zero-inflated negative binomial regression models (multivariate).

Results: Body mass index was positively associated with annual sick leave days in the bivariate analysis (P < 0.001). In the fully adjusted zero-inflated negative binomial, overweight women had 3.64, obese women 5.19, and obese men 3.48 excess sick leave days in 2009 (vs normal weight), while excess sick leave days of overweight men were not statistically significant. The extrapolated excess costs in the German working population amount to €2.18 billion (base case).

Conclusions: The absenteeism-related lost productivity costs associated with excess weight are formidable and emphasize the persistent need for health promotion efforts in Germany.

Back to Top | Article Outline
Learning Objectives

* Discuss the new data on associations between overweight and obesity and sick leave days in the German workforce.

* Identify differences in the impact of excess body weight on sick days between men and women.

* Discuss the costs associated with sick days related to overweight and obesity.

Defined by an excess of adipose tissue, obesity is a chronic disease and serious risk factor for a plethora of health problems.1 Among the noncommunicable disease risks, obesity has increased in relative importance over the last 20 years and is now among the leading risk factors for disease in many regions of the world.2 About one third of the world's adult population was overweight or obese (body mass index [BMI], ≥25 kg/m2) in 2008.3 In Germany, around 37% of adults were overweight (25 kg/m2 ≤ BMI < 30 kg/m2) in 2010, while another 23% were obese (BMI ≥ 30 kg/m2).4 Compared with individuals in the normal weight range (18.5 kg/m2 ≤ BMI < 25 kg/m2), those with overweight and obesity have been shown to use more health care services5 and to have higher health care costs (direct costs).6,7 Estimates of total direct costs attributable to overweight and obesity in Germany range from €4.854 billion (in 2002 Euros)8 to €11.359 billion (in 2003 Euros).9

In addition to increased health care costs, overweight and obesity have been linked to decreases in work ability10 and workforce productivity (indirect costs).11 Research has shown that excess weight, particularly obesity, is a significant predictor of presenteeism, absenteeism, disability, and premature mortality.11–15 In contrast to the wealth of international research, few studies have examined the impact of overweight and obesity on indirect costs7–9,16,17 (or indicators thereof15,18) in Germany, the majority of which are top-down cost of illness studies.8,9,16 Solely, a study by Wolfenstetter7 estimated excess weight–related indirect costs on the basis of individual-level survey data.

Probably the most frequently investigated productivity outcome in obesity research is absenteeism, that is, time/days absent from work because of sickness.12,13 Studies from the international literature provide strong evidence that obese employees take more sick leave days, with a trend toward spells of longer duration, and incur higher related costs than comparable normal-weight employees, whereas the evidence for overweight employees was inconclusive.12,13,19 Because of the paucity of empirical evidence for Germany, the aim of this study was to explore the relationship between excess weight and absenteeism, that is, self-reported annual sick leave days in 2009, in a representative sample of German employees, and to provide per-worker and nationally aggregated estimates of excess weight–related sick leave costs.

Back to Top | Article Outline



This analysis is a cross-sectional study based on pooled data from the 2009 and 2010 waves of the German Socio-Economic Panel (SOEP), an annual household panel representative for Germany. Introduced in 1984, the SOEP is the largest and longest running multidisciplinary longitudinal study in Germany, covering a wide range of topics such as housing, income, employment, and education. In addition, the data set contains information on health, making it suitable for analyses in the fields of epidemiology, public health, and health economics. All information is based on self-reports by respondents. More detailed information on the SOEP can be found elsewhere.20 We restricted our analysis to the subsample of respondents aged 18 to 65 years, with BMI ≥ 18.5 kg/m2, who had been employed in 2009. Cases with missing data for the variables of interest (19.61%) were excluded (Figure 1).

Figure 1
Figure 1
Image Tools
Back to Top | Article Outline

The dependent variable in our analysis was sick leave days; the main independent variable was BMI class. Sick leave days were measured as the total number of workdays missed in the year 2009 for health-related reasons. Body mass index was calculated on the basis of self-reported weight and height (not corrected for possible reporting error). According to the International Classification of the World Health Organization,21 we assigned respondents to three BMI classes: normal weight (18.5 kg/m2 ≤ BMI < 25 kg/m2), overweight (25 kg/m2 ≤ BMI < 30 kg/m2), and obese (BMI ≥ 30 kg/m2). In addition to BMI classes, sociodemographic, occupational, and health covariates were included in the analysis.

With regard to sociodemographic characteristics, we constructed—on the basis of school leaving degrees—three educational categories: low (no school degree and secondary school degree), middle (intermediate school degree), and high (technical school degree and upper secondary degree). Marital status was dichotomized, with only married couples living together considered as “married.” Single, divorced, widowed, and separated (but married) individuals were classified as “not married.”

Three variables related to work characteristics were included in the analyses. First, occupational status was dichotomized into “full-time” (full-time employment, vocational training, military service, community service, and sheltered workshop) and “part-time” employment (marginally employed and regular part-time employment). Respondents unemployed or near retirement with zero working hours were excluded from the sample. Second, those employed were classified as working in the primary (agriculture, forestry, and fishery), secondary (manufacturing industry), or tertiary economic sector (services). Third, a variable on autonomy of occupational activity was included (five-point Likert scale, with 1 = low autonomy and 5 = high autonomy). The variable is based on a classification of occupational position and is strongly correlated to the Treiman Prestige Scale.20,22 With respect to the respondents' health status, we included a dichotomous variable indicating current smoking behavior, as well as variables for five (obesity-related) diseases: depression, diabetes, cardiac disease, high blood pressure, and stroke.

Back to Top | Article Outline
Statistical Analyses

First, all variables of interest were compared regarding BMI groups in bivariate analyses, using chi-squared tests and analyses of variance (ANOVA). In a second step, the impact of BMI classes on the number of sick leave days was determined in regression analyses adjusting for different sets of covariates. Because sick leave days were nonnegative integers and the distribution of sick leave days was positively skewed and overdispersed, that is, a large proportion of respondents had no sick leave (44.5% of all cases), ordinary least squares regression was not appropriate. Therefore, we initially used different types of models for count data23–25: a simple Poisson model, a simple negative binomial model, a zero-inflated Poisson model, and a zero-inflated negative binomial (ZINB) model. The ZINB model was chosen over all other models on the basis of the Vuong test.23,26–28

The ZINB is a mixture model, which assumes that the study population latently consists of two subgroups; one group with a high probability of having zero sick leave days and a second group with a high probability of having at least one sick leave day.24 Therefore, the distribution of the outcome “sick leave days” is split in two component distributions, which are analyzed separately.23 In the first part (zero-inflated part), a logistic regression is performed to model the binomial distribution of having “zero days” versus “any sick leave.” In the second part, only the group with at least one sick leave day is considered and the actual count of sick leave days is modeled using a generalized linear model with a negative binomial distribution (count part). For the ease of interpretation, we present the odds of having any sick leave days, and the respective odds ratios in the zero-inflated part (logit model) and the respective number of days for the count part (negative binomial model).23–25

Because of an interaction between BMI classes and sex, we estimated separate models for male and female employees. Furthermore, we reran our models with two different sets of covariates: model 1 adjusted for sociodemographic (age, marital status, and education) and work-related characteristics (autonomy at work, economic sector, and occupational status), and model 2 additionally included health covariates (smoking, diabetes, cardiac disease, high blood pressure, depression, and stroke). Assuming that excess weight may have an additional effect on sick leave days, that is, even when none of the considered secondary diseases were present, we included obesity-related diseases in the analysis. We checked for interactions between BMI classes and each of the diseases, but excluded the interaction terms from the final model, because none of them was significant.

Within each model, the same set of covariates was used to analyze the zero-inflated part and the count part of the model. All models were checked for multicollinearity, using the variance inflation factor, but none of the selected covariates had to be excluded. The level of significance was set at 0.05, and all analyses were carried out with SAS version 9.2 (SAS Institute, Inc, Cary, NC).

Back to Top | Article Outline
Cost Calculation

Annual per capita costs of sick leave–related productivity losses in Germany were calculated by multiplying excess (vs normal weight) sick leave hours with the sex-specific mean hourly labor costs for full- and part-time employees in 2009 (assuming that labor costs approximate employees' marginal productivity).29 Sick leave hours were obtained by multiplying average work hours for full- and part-time employees per day with the number of (significant) excess sick leave days from the fully adjusted model (model 2). Labor costs were obtained by augmenting mean gross wages with ancillary wage costs paid for by employers by a factor of 1.27.30 To provide a conservative estimate, average work hours (women: 5.62 h/d; men: 6.98 h/d) and mean gross hourly wages of unskilled employees (women: €10.89; men: €11.85) were used in the base case.29

Total population costs were calculated by multiplying annual per capita sick leave costs with the total number of overweight and obese male and female employees with sick leave days of 1 or more in 2009. Information on the population aged 18 to 65 years came from the Statistical Yearbook of the German Federal Statistical Office,31 while the employment rate, the sex-specific prevalence of overweight and obesity, and the respective share of employees with sick leave days of 1 or more were based on SOEP calculations. Costs were calculated from a societal perspective, considering lost productivity in paid work (due to absenteeism) only.

Back to Top | Article Outline
Sensitivity Analyses

To evaluate the uncertainty inherent in our total population cost estimates, three input parameters were varied in univariate sensitivity analyses. First, we used sex-specific prevalences of overweight (women: 45.5%) and obesity (women: 21.1%; men: 20.5%) from the German National Nutrition Survey II to extrapolate per capita costs to total population costs. The German National Nutrition Survey II is a representative study where BMI was based on anthropometric measurements of height and weight.32 Second, in addition to the mean gross wages of unskilled employees used in the base case, we varied labor costs, using mean gross wages for two further groups of employees, that is, semiskilled employees (women: €13.78; men: €14.59) and skilled employees (women: €15.85; men: €17.58). Third, excess sick leave days were taken from model 1, in which health-related variables were not controlled for.29 Furthermore, we calculated a worst-case scenario with prevalence data from the German National Nutrition Survey II,32 labor costs of skilled employees,29 and excess sick leave days from model 1. Further information about the input data can be found in the Supplemental Digital Content Table S3,

Back to Top | Article Outline


Sample Characteristics

The final study sample consisted of 7990 individuals aged between 18 and 65 years, who had been employed in 2009, and for whom complete information on all variables of interest was available (Figure 1). Of the total study sample, 37.82% were overweight and 17.13% obese, with marked sex differences. Although 47.51% of men were overweight and 19.13% obese, the corresponding prevalence was much lower in women, with 27.67% being overweight and 15.05% obese. Table 1 provides an overview of sample characteristics for the total sample and across BMI classes. Respondents were on average 45.21 (SD = 10.42) years old; the majority of 72.99% were full-time employed and worked in the tertiary economic sector (57.68%). Body mass index groups significantly differed from one another on most included variables. As expected, overweight and obese employees were on average older, less educated, and had significantly more obesity-related diseases. Obese employees more often held jobs with low autonomy in occupational activity. All variables except for smoking (P = 0.074) and the prevalence of depression (P = 0.074) achieved statistical significance in the bivariate analyses.

Table 1
Table 1
Image Tools
Back to Top | Article Outline
BMI and Sick Leave Days

Body mass index groups significantly differed in sick leave days in the bivariate analyses. The mean number of sick leave days increased from normal weight (7.45 [SD = 20.62]) to overweight (9.77 [SD = 28.84]) and obesity (14.04 [SD = 38.39]). Among respondents with sick leave days of 1 or more, the mean number of sick leave days was 13.67 (SD = 26.37), 18.00 (SD = 37.17), and 23.33 (SD = 47.26) for each of the BMI groups, respectively. On average, women had more sick leave days, with differences being especially pronounced for employees with overweight (Figure 2).

Figure 2
Figure 2
Image Tools

Table 2 displays the relationship between BMI classes and any sick leave (logit model), and mean sick leave days (negative binomial model), adjusted for sociodemographic and work-related variables (model 1), and in addition, for health-related variables (model 2), separately for female and male employees. As can be derived from the logit part of the model, BMI was not significantly associated with having any sick leave days. Even though overweight women and overweight and obese men did have higher odds than normal-weight employees for any sick leave days in both models, differences between BMI groups were not statistically significant. Nevertheless, several of the included variables in the fully adjusted model, for example, age, full-time work, high blood pressure, and depression, were significantly associated with “any sick leave days” (Supplemental Digital Content Tables S1a and S1b,

Table 2
Table 2
Image Tools

The results of the negative binomial models differ by sex. Although both female and male employees with obesity had significantly more sick leave days than employees with normal weight, sick leave days of overweight men did not significantly differ from those of normal-weight men. The positive association between BMI classes and sick leave days was extenuated—compared with the bivariate analysis—after controlling for sociodemographic and work-related covariates (model 1), and decreased further when health-related covariates were additionally adjusted for (model 2). In general, the association between BMI and sick leave days was somewhat more pronounced for female employees, as indicated by the excess sick leave days in Table 2. When sociodemographic and work-related characteristics were adjusted for (model 1), overweight women had 13.72, while obese women had 15.77 absence days, compared with 9.76 days for women with normal weight. This corresponds to 3.96 and 6.01 annual excess sick leave days, respectively. In comparison, male employees with overweight had only 0.75 (not significant) and those with obesity 5.71 annual excess sick leave days. In the fully adjusted model (model 2), overweight women had 3.64, obese women 5.19, and obese men 3.48 excess sick leave days. Mean sick leave days were furthermore associated with age, occupational status, smoking, and depression in female employees, and age, education, autonomy at work, smoking, cardiac disease, and depression in male employees (Supplemental Digital Content Tables S1a and S1b,

Back to Top | Article Outline
Overweight- and Obesity-Related Sick Leave Costs

On the basis of the finding of 3.64 (95% confidence interval [CI]: 1.48 to 5.81) and 5.19 (95% CI: 2.29 to 8.09) excess sick leave days for overweight and obese women, respectively, annual per capita cost of €284 (95% CI: €115 to €454) and €405 (95% CI: €179 to €632) were estimated. Because overweight men did not differ significantly from normal-weight men, costs were calculated for male obese employees only. The excess sick leave days of 3.48 (95% CI: 1.55 to 5.40) for male obese employees translate to annual per capita costs of €367 (95% CI: €164 to €570). Extrapolated to the corresponding population of female and male overweight and obese employees in Germany in 2009, these per capita cost estimates yield total population costs of €2.18 billion, of which nearly two thirds were attributable to women (base case in Table 3).

Table 3
Table 3
Image Tools
Back to Top | Article Outline
Sensitivity Analyses Results

Table 3 depicts the results from the sensitivity analyses. The variation of parameters resulted in attributable total cost between €2.53 billion (using sex-specific overweight and obesity prevalences from German National Nutrition Survery II) and €3.6 billion (using average labor costs of skilled workers), which corresponds to cost increases of +16% and +65% in relation to the base case result. In the worst-case scenario, the attributable costs of €5.42 billion increased by 149% in comparison to the base case, which was disproportionally driven by increased costs for male employees (+183%). More detailed information can be found in the Supplemental Digital Content Table S3,

Back to Top | Article Outline


Main Findings

This study examined the association between BMI classes and sick leave days in a sample of the German working population and calculated attributable per worker and total population costs. Body mass index groups significantly differed in the number of sick leave days in the bivariate relationship, that is, employees with overweight had about 31% more and those with obesity about 88% more absence days than normal-weight employees (full study sample). Compared with men, female employees with excess weight had more sick leave days on average (Figure 2). The positive association between BMI and sick leave days was extenuated after controlling for sociodemographic and work- and health-related covariates in multivariate models, with overweight men not exhibiting significantly more sick leave than those with normal weight in the multivariate models. The annual incremental sick leave days of overweight women (3.64) and obese women (5.19) and obese men (3.48) derived from the fully adjusted ZINB model (model 2) translated into annual per capita costs of €284, €405, and €367 in 2009, respectively. Extrapolated to the total working population in Germany, costs of overweight and obesity as a result of sick leave were estimated at €2.18 billion in 2009 (base case). The univariate sensitivity analyses yielded annual excess weight–related sick leave–attributable costs between €2.53 billion (+16%) and €3.6 billion (+65%).

Back to Top | Article Outline
Comparison to Previous Studies

This is the first study exclusively concerned with the association between excess weight and sick leave (costs) in Germany. This study's findings are generally comparable to those reported in the international literature, where obesity was repeatedly found to be associated with an increased risk and duration of sickness absence, while the evidence for overweight was mixed.12,13 Similar to our findings, many studies moreover observed a more pronounced association for female employees.19,33 Because of differences in study methodology and institutional arrangements, our estimates are not directly comparable to those from studies conducted with populations from other countries. In difference to many prior studies,12,13,19,34,35 we included obesity-related diseases as health covariates in multivariate models. Even though adjusting for health-related variables decreased the effect of BMI on sick leave days (and necessarily costs) only slightly in female employees, the effect in male employees (especially those with obesity) was more pronounced. Hence, irrespective of the presence of secondary diseases, especially employees with obesity had more sick leave days than comparable employees with normal weight. Therefore, excess weight itself (eg, because of aches and pains, dyspnea, sleep disturbances) or possibly mediated through further uncontrolled secondary diseases (eg, musculoskeletal problems or neoplasms) or physical or psychosocial factors of the work environment is associated with sickness absence.

Costs attributable to overweight- and obesity-related sickness absence were reported more rarely.8,9,11,19 Contrastable evidence for Germany is scarce.7–9 A recent prospective cohort study by Wolfenstetter7 used individual-level survey data to estimate indirect costs resulting from sick leave in Germany. The author reported annual per capita sick leave costs of €2271 (€2362 in 2009 Euros) for normal weight, €2826 (€2939 in 2009 Euros) for overweight, and €2830 (€2943 in 2009 Euros) for obese employees (adjusted for age, sex, and socioeconomic status), which corresponds to excess costs of €555 (€578 in 2009 Euros) for overweight and €559 (€580 in 2009 Euros) for obese employees. In comparison to Wolfenstetter, per capita sick leave costs were substantially lower in this study for each of the BMI groups. Using sex-specific sick leave days from model 1, costs for female (male) employees with normal weight, overweight, and obesity amounted to €762 (€1142), €1071 (€1221), and €1231 (€1743), respectively, which corresponds to excess sick leave costs of €309 (€150; not significant) for overweight and €469 (€601) for obese employees. The underlying reasons for the discrepancy in the total sick leave–related per capita costs between both studies remain unclear but could have come about as a result of differences in (excess) sick leave days and/or differences in the wages used to calculate sick leave–related per capita costs for the BMI groups (both of which are not reported). In difference to this study, Wolfenstetter did not extrapolate per capita costs to the total population.

The only available estimates of the total excess weight–related absenteeism costs in Germany come from two top-down cost of illness studies, which reported total population cost of €481 million (€513 million in 2009 Euros)8 and €582 million (€615 million in 2009 Euros).9 These estimates are substantially less than the €2.18 billion we estimated. This stark discrepancy is likely a result of the different methodological approaches used.36 Although the two top-down cost of illness studies applied the concept of population-attributable fraction and used data from various secondary sources to mathematically derive cost estimates, we followed a bottom-up approach, estimating excess weight–related sick leave days directly from individual-level data, based on which costs were extrapolated to the German working population. In accordance with the earlier-mentioned observation, a recent publication concerned with different methods of measuring obesity costs found that reported estimates vary greatly between studies, even in studies using similar methods (eg, population-attributable fraction), and that database studies tended to report higher estimates than population-attributable fraction or modeling studies, which decreased with larger sample sizes and longer observation periods, though.36 Hence, cost estimates of obesity should be scrutinized against the background of the methodology and assumptions of the particular study. Beyond the general inherent methodological and interpretative weaknesses of cost analyses, 36,37 authors have attributed the observed discrepancies in cost estimates of obesity to methodological issues and shortcomings of interdisciplinary obesity research in general, for example, lack of a conceptual definition, lax application of scientific standards of review, tenuous assumption making, and flawed measurements.38 It has been pointed out that these obstacles may have constrained obesity research, from study design choices, operational definitions, selection and measurement of variables, and development and use of measurement tools, and consequently policy decisions for decades.38

Back to Top | Article Outline
Strengths and Weaknesses of the Study

The major strength of this study is the use of data from a fairly large and high-quality sample representative for the German (working) population, which includes comprehensive information on respondents' sociodemography, employment, and health, among others.20 We were, therefore, able to include variables from these domains and thus, control for factors possibly confounding the association between BMI and sick leave days. Nevertheless, because the availability of candidate variables was limited in the 2009/2010 waves of the SOEP data set, the possibility of inadequate or insufficient adjustment for confounding remains. For example, we were not able to control for physical exercise, diet, and sleep duration, because the data set (given the cross-sectional design of this study) did not include adequate measures. Moreover, the role of health-related variables in the causal pathway between BMI and sick leave days is complex, that is, even theoretically it is not always clear how these interact with each other (cause, effect, mediator, moderator).

The present findings are based on pooled data from the 2009 and 2010 waves of the SOEP. Because cross-sectional data do not provide information about the temporal sequence of assumed causes and effects, the effect estimates reported in this study reveal underlying correlations, instead of causal effects between BMI (and other explanatory variables) and sick leave. Hence, reverse causality remains a possibility, that is, sick leave may have modified our explanatory variables (particularly BMI) in some way. The availability of longitudinal analyses showing that baseline obesity is a significant predictor of sick leave in the future sheds doubt on the eligibility of the reverse causality hypothesis, though.7,12 A more plausible possibility is that further unobserved factors may have caused both excess weight and sickness absence.

Confining our sample to respondents employed in 2009 may have limited the generalizability of our effect estimates. Because of the large health burden associated with overweight and obesity, affected individuals may not have been participating in the workforce in 2009 (eg, because of unemployment or disability). A continuous selection process could have led to a “healthy worker effect,” that is, those employed could be healthier and have less sick leave than those not employed. Empirical evidence by Klarenbach et al39 indicates that increasing severity of obesity is associated with lower odds of workforce participation. We analyzed this hypothesis separately for male and female employees and found that BMI classes (adjusted for sociodemographic and health-related variables) were not associated with increased odds of unemployment (Supplemental Digital Content Table S2,

Furthermore, all information in the SOEP is based on respondents' self-reports. Regarding BMI, it has repeatedly been shown that BMI based on self-reports of height and weight is lower than BMI calculated with data from objective measurements.40 Comparing the sex-specific prevalence of overweight and obesity in the SOEP sample with those from two other recent population representative studies in Germany (both of which used anthropometric measurements of height and weight)4,32 reveals that the prevalence of obesity was probably underestimated in women. It follows that some obese women may have been misclassified as overweight and some overweight women as normal weight, which means that the association between BMI and sick leave may have been underestimated (resulting in conservative estimates of incremental sick leave days) for female employees. Similarly, respondents could have underestimated their sick leave days in the previous year as a result of recall problems. Nevertheless, previous studies have shown that self-reported sick leave days correlated well with those from employer records.41–43

Excluding cases with incomplete information on one or more variables (19.61% of all cases) may have resulted in a less-representative sample. A comparison of the final study sample (n = 7990) with the excluded cases (n = 1949) shows that the latter were on average younger, less often married, and had a lower prevalence of overweight and obesity and fewer sick leave days. Although the consequences of excluding cases with incomplete data are not easily assessed, it seems likely that the true effect of BMI class on sick leave days was overestimated as a result. Excluding these cases from the final analyses may have introduced selection and hence increased the uncertainty of our findings.

Another strength of this study is that we used several statistical models that are routinely used for the analysis of count data and compared their respective fits to the observed counts. To appropriately account for the fact that the outcome variable was overdispersed and had a large proportion of excess zeros, we chose a ZINB model, which also showed superior in a formal test (Vuong test).

Back to Top | Article Outline
Future Research and Policy Implication

The results of this cross-sectional study suggest that (overweight and) obese employees in Germany have more sick leave days than comparable employees with normal weight. The associated lost productivity costs are formidable and emphasize the large economic and societal burden accompanied by obesity in Germany.7–9 The findings of the present study should be verified using longitudinal data, including a comprehensive set of sociodemographic and health-related covariates, as well as those related to the physical and psychosocial work environment, because these factors have previously been linked to excess sick leave days in overweight and obese employees.19,24 With regard to the duration of sick leave, it would furthermore be beneficial to distinguish between different lengths of spells in future studies. Because of the general paucity of empirical evidence regarding the economic consequences of excess weight in Germany, future studies should moreover investigate overweight- and obesity-related costs associated with presenteeism (reduced productivity while at work), disability, and premature mortality—preferably using individual-level data.

The current study's findings stress the relevance of health promotion efforts in the work place and beyond. Internationally, various workplace health promotion programs focusing on physical activity, nutrition, and weight loss/maintenance have been developed and evaluated.44–48 Although many programs were shown to help improving or maintaining employee health,46–48 the effects on weight/BMI44 and work ability/productivity and sickness absence were modest.45 The evidence regarding the financial return on investment of worksite health promotion programs surrounding obesity and its behavioral risk factors is inconclusive at this point.49

In contrast to the substantial and growing international evidence on the health and economic effects of workplace health promotion interventions in the field of obesity, comparable studies from Germany are very rare.50 To curtail the possible productivity losses resulting from excess weight in Germany, appropriate interventions/programs should be developed and evaluated regarding their effects on health and productivity outcomes, as well as toward their financial implications.

Back to Top | Article Outline

The data used in this publication were made available to us by the SOEP study at the German Institute for Economic research (DIW Berlin), Berlin.

Back to Top | Article Outline


1. Blüher M. Are there still healthy obese patients? Curr Opin Endocrinol Diabetes Obes. 2012;19:341–346.

2. Lim SS, Vos T, Flaxman AD, et al. A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380:2224–2260.

3. Finucane MM, Stevens GA, Cowan MJ, et al. National, regional, and global trends in body-mass index since 1980: systematic analysis of health examination surveys and epidemiological studies with 960 country-years and 9.1 million participants. Lancet. 2011;377:557–567.

4. Kurth BM. Erste Ergebnisse aus der “Studie zur Gesundheit Erwachsener in Deutschland” (DEGS) [First results from “the German Health Interview and Examination Survey for Adults” (DEGS)]. Bundesgesundheitsbla. 2012;55:980–990.

5. Wolfenstetter SB, Menn P, Holle R, Mielck A, Meisinger C, von Lengerke T. Body weight changes and outpatient medical care utilisation: results of the MONICA/KORA cohorts S3/F3 and S4/F4. Psychosoc Med. 2012;9: Doc09. doi:10.3205/psm000087.

6. von Lengerke T, Reitmeir P, John J. Direkte medizinische Kosten der (starken) Adipositas: ein Bottom-up-Vergleich über- vs. normalgewichtiger Erwachsener in der KORA-Studienregion [Direct medical costs of (severe) obesity: a bottom-up assessment of over- vs. normal-weight adults in the KORA-study region (Augsburg, Germany)]. Gesundheitswesen. 2006;68:110–115.

7. Wolfenstetter SB. Future direct and indirect costs of obesity and the influence of gaining weight: results from the MONICA/KORA cohort studies, 1995–2005. Econ Hum Biol. 2012;10:127–138.

8. Konnopka A, Bodemann M, Konig HH. Health burden and costs of obesity and overweight in Germany. Eur J Health Econ. 2011;12:345–352.

9. Knoll K, Hauner H. Kosten der Adipositas in der Bundesrepublik Deutschland [A health-economic analysis of the total cost burden caused by obesity and the diseases associated with obesity in the Federal Republic of Germany]. Adipositas. 2008;2:204–210.

10. van den Berg TI, Elders LA, de Zwart BC, Burdorf A. The effects of work-related and individual factors on the Work Ability Index: a systematic review. Occup Environ Med. 2009;66:211–220.

11. Trogdon JG, Finkelstein EA, Hylands T, Dellea PS, Kamal-Bahl SJ. Indirect costs of obesity: a review of the current literature. Obes Rev. 2008;9:489–500.

12. van Duijvenbode DC, Hoozemans MJM, van Poppel MN, Proper KI. The relationship between overweight and obesity, and sick leave: a systematic review. Int J Obesity. 2009;33:807–816.

13. Neovius K, Johansson K, Kark M, Neovius M. Obesity status and sick leave: a systematic review. Obes Rev. 2009;10:17–27.

14. Neovius K, Johansson K, Rossner S, Neovius M. Disability pension, employment and obesity status: a systematic review. Obes Rev. 2008;9:572–581.

15. Arndt V, Rothenbacher D, Zschenderlein B, et al. Body mass index and premature mortality in physically heavily working men—a ten-year follow-up of 20,000 construction workers. J Occup Environ Med. 2007;49:913–921.

16. Sander B, Bergemann R. Economic burden of obesity and its complications in Germany. Eur J Health Econ. 2003;4:248–253.

17. Breitfelder A, Wenig CM, Wolfenstetter SB, et al. Relative weight-related costs of healthcare use by children—results from the two German birth cohorts, GINI-plus and LISA-plus. Econ Hum Biol. 2011;9:302–315.

18. Claessen H, Arndt V, Drath C, Brenner H. Overweight, obesity and risk of work disability: a cohort study of construction workers in Germany. Occup Environ Med. 2009;66:402–409.

19. Cawley J, Rizzo JA, Haas K. Occupation-specific absenteeism costs associated with obesity and morbid obesity. J Occup Environ Med. 2007;49:1317–1324.

20. Wagner GG, Frick JR, Schupp J. The German Socio-Economic Panel Study (SOEP)—scope, evaluation and enhancements. Schmollers Jahrbuch. 2007;127:161–191.

21. Jebb SA, Johnstone AM, Warren J, Goldberg GR, Bluck L. Key methodologies in obesity research and practice. In:Williams G, Frühbeck G, eds. Obesity: Science to Practice. Chichester, England: Wiley; 2009:45–78.

22. Ganzeboom HBG, Degraaf PM, Treiman DJ. A standard international socioeconomic index of occupational status. Soc Sci Res. 1992;21:1–56.

23. Atkins DC, Gallop RJ. Rethinking how family researchers model infrequent outcomes: a tutorial on count regression and zero-inflated models. J Fam Psychol. 2007;21:726–735.

24. Taimela S, Laara E, Malmivaara A, et al. Self-reported health problems and sickness absence in different age groups predominantly engaged in physical work. Occup Environ Med. 2007;64:739–746.

25. Afifi AA, Kotlerman JB, Ettner SL, Cowan M. Methods for improving regression analysis for skewed continuous or counted responses. Annu Rev Publ Health. 2007;28:95–111.

26. University of California, Los Angeles. Zero-inflated negative binomial regression. UCLA: Statistical Consulting Group. Available at: Accessed April 9, 2013.

27. University of California, Los Angeles. Poisson regression. UCLA: Statistical Consulting Group. Available at: Accessed February 15, 2013.

28. Vuong QH. Likelihood ratio tests for model selection and non-nested hypotheses. Econometrica. 1989;57:307–333.

29. Federal Statistical Office (Statistisches Bundesamt). Verdienste und Arbeitskosten 2009 [Earnings and Labour Costs 2009]. Fachserie 16, Reihe 2.3. Wiesbaden, Germany: Statistisches Bundesamt; 2010.

30. Eurostat. Social security and other labour costs paid by employer. Available at: Accessed April 12, 2013.

31. Federal Statistical Office (Statistisches Bundesamt). Statistisches Jahrbuch 2011—Für die Bundesrepublik Deutschland mit “Internationalen Übersichten” [Statistical Yearbook 2011—For the Federal Republic of Germany Including “International Tables”]. Wiesbaden, Germany: Statistisches Bundesamt; 2011.

32. Max Rubner-Institut. Nationale Verzehrsstudie II, Ergebnisbericht, Teil 1 [National Nutrition Survey II]. Karlsruhe, Germany: Max Rubner-Institut; 2008. Available at: Accessed March 6, 2013.

33. Ferrie JE, Head J, Shipley MJ, Vahtera J, Marmot MG, Kivimäki M. BMI, obesity, and sickness absence in the Whitehall II study. Obesity. 2007;15:1554–1564.

34. Jans MP, van den Heuvel SG, Hildebrandt VH, Bongers PM. Overweight and obesity as predictors of absenteeism in the working population of the Netherlands. J Occup Environ Med. 2007;49:975–980.

35. Laaksonen M, Piha K, Sarlio-Lahteenkorva S. Relative weight and sickness absence. Obesity. 2007;15:465–472.

36. Bierl M, Marsh T, Webber L, Brown M, McPherson K, Rtveladze K. Apples and oranges: a comparison of costing methods for obesity. Obes Rev. 2013;14:693–706.

37. Roux L, Donaldson C. Economics and obesity: costing the problem or evaluating solutions? Obes Res. 2004;12:173–179.

38. Herbert JA, Allison DB, Archer E, Lavie CJ, Blair SN. Scientific decision making, policy decisions, and the obesity pandemic. Mayo Clin Proc. 2013;88:593–603.

39. Klarenbach S, Padwal R, Chuck A, Jacobs P. Population-based analysis of obesity and workforce participation. Obesity (Silver Spring). 2006;14:920–927.

40. Connor Gorber S, Tremblay M, Moher D, Gorber B. A comparison of direct vs. self-report measures for assessing height, weight and body mass index: a systematic review. Obes Rev. 2007;8:307–326.

41. Voss M, Stark S, Alfredsson L, Vingard E, Josephson M. Comparisons of self-reported and register data on sickness absence among public employees in Sweden. Occup Environ Med. 2008;65:61–67.

42. Fredriksson K, Toomingas A, Torgen M, Thorbjörnsson CB, Kilbom A. Validity and reliability of self-reported retrospectively collected data on sick leave related to musculoskeletal diseases. Scand J Work Environ Health. 1998;24:425–431.

43. Ferrie JE, Kivimaki M, Head J, Shipley MJ, Vahtera J, Marmot MG. A comparison of self-reported sickness absence with absences recorded in employers' registers: evidence from the Whitehall II study. Occup Environ Med. 2005;62:74–79.

44. Anderson LM, Quinn TA, Glanz K, et al. The effectiveness of worksite nutrition and physical activity interventions for controlling employee overweight and obesity: a systematic review. Am J Prev Med. 2009;37:340–357.

45. Rongen A, Robroek SJ, van Lenthe FJ, Burdorf A. Workplace health promotion: a meta-analysis of effectiveness. Am J Prev Med. 2013;44:406–415.

46. Vuillemin A, Rostami C, Maes L, et al. Worksite physical activity interventions and obesity: a review of European studies (the HOPE Project). Obesity Facts. 2011;4:479–488.

47. Jensen JD. Can worksite nutritional interventions improve productivity and firm profitability? A literature review. Perspect Public Health. 2011;131:184–192.

48. Maes L, Van Cauwenberghe E, Van Lippevelde W, et al. Effectiveness of workplace interventions in Europe promoting healthy eating: a systematic review. Eur J Public Health. 2012;22:677–683.

49. van Dongen JM, Proper KI, van Wier MF, et al. Systematic review on the financial return of worksite health promotion programmes aimed at improving nutrition and/or increasing physical activity. Obes Rev. 2011;12:1031–1049.

50. Oberlinner C, Lang S, Germann C, et al. Prevention of overweight and obesity in the workplace. Gesundheitswesen. 2007;69:385–392.

Supplemental Digital Content

Back to Top | Article Outline

Copyright © 2014 by the American College of Occupational and Environmental Medicine


Article Tools