The relationship between overweight and obesity, and sick ... - Nature

International Journal of Obesity (2009) 33, 807–816 & 2009 Macmillan Publishers Limited All rights reserved 0307-0565/09 $32.00 www.nature.com/ijo

REVIEW The relationship between overweight and obesity, and sick leave: a systematic review DC van Duijvenbode1, MJM Hoozemans2, MNM van Poppel1 and KI Proper1 1 Department of Public and Occupational Health and the EMGO Institute for Health and Care Research, VU University Medical Center Amsterdam, Amsterdam, The Netherlands and 2Research Institute MOVE, Faculty of Human Movement Sciences, VU University Amsterdam, Amsterdam, The Netherlands

Objective: To investigate the relationship between overweight and obesity and sick leave. Design: Systematic, qualitative review. Literature search: A search in diverse databases was performed. Studies were considered as relevant if they were longitudinal in design and investigated the relationship between overweight and obesity and sick leave. Data extraction: The methodological quality of the studies was evaluated with a quality criteria list. To draw conclusions, a bestevidence synthesis was applied. Results: Thirteen studies were included. Four out of seven found overweight to be a predictor of long-term sick leave, whereas the remaining three showed a positive trend, but did not observe significance. Of the five studies investigating the relationship between overweight and short-term sick leave, inconsistent results were shown. Seven out of eight studies investigating the relationship with long-term sick leave found obesity as a significant predictor of long-term sick leave. In contrast, there were inconsistent results between the five studies examining the relationship between obesity and short spells of sick leave. Conclusions: Although this review found inconclusive evidence for a relationship between overweight and sick leave, a clear trend was discerned in that overweight was a predictor of especially long spells (47 days) of sick leave. As regards obesity, there was strong evidence for a positive relationship with sick leave because of the consistent finding that obesity was a significant predictor of long-term sick leave. International Journal of Obesity (2009) 33, 807–816; doi:10.1038/ijo.2009.121; published online 16 June 2009 Keywords: overweight; relation; sick leave; literature study; levels of evidence

Introduction Over the past decades, the prevalence of overweight (body mass index, BMI X25 kg/m2) and obesity (BMI X30 kg/m2) has increased dramatically.1–4 Overweight and obesity have an enormous public health impact. For instance, they are linked to an increased risk of type 2 diabetes mellitus, coronary heart disease, stroke, hypertension, gallbladder disease, some forms of cancer, sleep apnoea, and osteoarthritis.1 Owing to increased morbidity, overweight and obesity are associated with increased health care costs. Correspondence: Dr KI Proper, Department of Public and Occupational Health and the EMGO Institute for Health and Care Research, VU University Medical Center Amsterdam, Van der Boechorststraat 7, 1081 BT Amsterdam, The Netherlands. E-mail: [email protected] Received 15 December 2008; revised 1 May 2009; accepted 10 May 2009; published online 16 June 2009

Estimates vary from 1 to 7% of the total health care cost in the developed world.5–7 By way of illustration, from a systematic review, Thompson and Wolf6 suggested that obesity accounted for 5.5–7.0% of the national health expenditure in the United States. A few years later, using an econometric model, Finkelstein et al.7 estimated the obesity-attributable health care expenditures in the United States at state level and reported the costs to be $75 billion in 2003, which averaged by about 6% of the adult medical expenditures. In addition, obesity-related morbidity involves considerable indirect costs, that is costs owing to lost productivity, such as sickness absence. For example, a Chinese case study found that the indirect costs were often far more significant than the health care costs, with indirect costs owing to loss of productivity ranging from 3.6 to 8.7% of the gross national product in China in 2000 and 2025, respectively.8 This was supported by Katzmarzyk and Janssen9 who estimated the total costs of obesity in Canada

Relationship between overweight and sick leave: a review DC van Duijvenbode et al

808 at $4.3 billion, with two-thirds of the costs involving indirect costs. In addition, from a study among Swedish women, it was concluded that approximately 10% of the total cost of productivity loss owing to sick leave and disability pensions was related to obesity and obesity-related diseases.10 A recent review also concluded that BMI was significantly associated with disability pension.11 The rising prevalence of overweight and obesity, the associated increased risk of morbidity, and the high direct and indirect costs are a motivation to investigate the relationship of overweight and obesity with sick leave. Although some studies have investigated the relationship between overweight and obesity and sick leave, as far as the authors are aware, no review has been published yet that primarily examines this relationship. Instead, the earlier reviews on a similar topic (eg costs of obesity) have briefly discussed the potential for a relationship between overweight, obesity, and sick leave.12,13 In addition, earlier literature studies did not apply a systematic approach in the literature search or use a quality assessment and bestevidence synthesis. The aim of this review was to investigate the relationship between overweight and obesity and sick leave using a systematic approach in literature search, selection, and quality assessment.

Methods A computerised literature search was carried out in PubMed, Embase, PsycINFO, and SPORTDiscus databases. The search strategy consisted of two elements related to the independent variables (e.g. ‘overweight’, ‘obesity’, ‘body weight’, ‘BMI’, ‘waist circumference’, and ‘skinfold thickness’) and the outcome variables (e.g. ‘absenteeism’, ‘sick leave’, and ‘sickness absence’). In addition, references of key publications were searched, as were references of all included studies. The authors’ personal database was also searched.

Table 1

Inclusion criteria Publications found in the PubMed, Embase, PsycINFO, and SPORTDiscus databases were screened by two reviewers (DvD and KP) to meet the following inclusion criteria: The study population was used. The study examined the relationship between a measure of overweight or obesity and a measure of sick leave. The study design was longitudinal. The study had to be a full peer-reviewed report published in English, Dutch, or German. The study had to be published between 1987 and April 2009.

Quality assessment Two reviewers (DvD and KP) independently scored the quality of each study according to a standardised set of predefined criteria (Table 1). The criteria were adapted from Arie¨ns and Hayden.14,15 Each quality criterion was rated positive [1] or negative [0], with a negative score in the case of an inadequate performance or an unclear or missing description of the item. For each study, a total quality score was calculated by counting the number of items scored positively. A study was considered as high quality if least 51% was scored positively; otherwise a study was considered as low quality. The results of the scorings of both reviewers were compared and differences were discussed during a consensus meeting. If, after discussion, the reviewers could not agree, a third person (MH) made the final decision. Levels of evidence To draw conclusions as to the relationship between overweight and obesity and sick leave, a best-evidence synthesis was applied consisting of four levels of evidence and taking into account the methodological quality and the outcomes of the studies selected (i.e. inconsistent outcomes). The strength of evidence was assessed according to the rating

Criteria list for assessment of methodological quality of longitudinal studies14,15

Item definition Study purpose Study population

Independent variable

Dependent variable

Analysis and data presentation

International Journal of Obesity

A. Positive, if a specific, clearly stated purpose was described. B. Positive, if the main features (description of sampling frame, distribution by age and sex) of the study population were stated. C. Positive, if the participation rate at baseline was at least 80%. D. Positive, if the response after 1 year of follow-up was at least 80%, or if the non-response was not selective. E. Positive, if data on the independent variables (e.g. weight and waist circumference) were measured with standardised methods of acceptable quality and not by self-report. F. Positive, if data on sex, age, and smoking or alcohol use were collected and used in the analyses. G. Positive, if the exposure was assessed at a time before the occurrence of the outcome. H. Positive, if data on sick leave were collected using standardised methods of acceptable quality. Data were collected from company or insurance data and not by self-report. I. Positive, if data on sick leave were collected for at least 1 year. J. Positive, if the statistical model used was appropriate for the outcome studied and the measures of association estimated with this model are presented (including confidence intervals). K. Positive, if the study controlled for confounding. L. Positive, if the number of cases in the multivariate analysis was at least 10 times the number of independent variables in the analysis.


809 system of Hoogendoorn et al.16 and consisted of the following four levels: Strong evidence: consistent findings in multiple highquality studies. Moderate evidence: consistent findings in one highquality study and at least one low-quality study, or consistent findings in multiple low-quality studies. Insufficient evidence: only one study available. Inconclusive evidence: inconsistent findings in multiple studies. Findings were considered consistent when at least 75% of the studies showed results in the same direction, which was defined according to significance (Po0.05).

Results Search and selection A total of 1228 citations from PubMed, Embase, PsycINFO, and SPORTDiscus were screened. Out of these, a total of 13 longitudinal studies met the inclusion criteria.17–29 One publication30 described a study that was described partly in two earlier publications.27,28 Although the most recent publication slightly differed from the other publications in that it combined the study populations of the two earlier ones and added another oil refinery, it did not provide additional data that were relevant to this review. It was, therefore, decided to exclude the 2008 publication. The characteristics of these studies are presented in Table 2. The most important reason for exclusion was the fact that study titles or, next, abstracts did not report on a relationship between overweight or obesity and sick leave. Further, many studies were not longitudinal in design. Studies were also excluded because they investigated the relationship between overweight or obesity and disability pension or presenteeism, which was not the outcome of interest in this review.

Methodological quality assessment The scoring of the 13 longitudinal studies led to a disagreement of 12%. The reviewers could not reach consensus about one study, and thus the third reviewer was consulted. Table 3 shows the outcome of the quality assessment. All studies except for one27 were of high quality.

Study characteristics The study population of the prospective studies varied from 255 to 21 419 workers. The follow-up time of the studies varied from 1 to 10 years. All studies used BMI as independent variable, with one study using both BMI and waist circumference.24 Weight was evaluated by self-report in half of the studies. Generally, the same BMI cut-off points were used to define overweight or obesity, namely 25–29.9 and X30 kg/m2, respectively. In one study, however, obesity

was defined as a BMI X27.2 kg/m2 for men and 26.9 kg/m2 for women, which represented a value of 20% above the ideal body weight as defined by recommendations.27 All except one study23 had an objective measure of sick leave. Sick leave data were mostly collected from company records. Some studies distinguished between short- and long-term sick leave, but differed in the operationalisation. For example, Labriola et al.23 and Moreau et al.24 applied a cutoff point of 7 days to distinguish between short- and longterm sick leave, whereas another study applied a cut-off point of 3 days of sick leave.22 Another study examined the relationship between overweight and obesity with multiple measures of sick leave, including very short (1–2 days), short (3–7 days), medium (8–21 days), and long (X22 days).21 In total, eight studies investigated the relationship between overweight and/or obesity with long-term sick leave,19–24,28,29 of which five studies also investigated the relationship with shorter spells of sick leave.20–24 The remaining five studies did not distinguish between duration of sick leave; they investigated the relationship between overweight and obesity and total duration of sick leave.17,18,25–27 Further, two studies investigated the relationship with a specific measure of sick leave, namely absence owing to respiratory complaints and musculoskeletal problems, respectively.17,18 All other studies used sickness absence irrespective of the cause.

Overweight The relationship between overweight and sick leave was investigated in eight high-quality longitudinal studies.19–25,28 Table 4 presents an overview of the high quality studies showing either positive or no relationship with sick leave. Seven of these examined the relationship with long-term sick leave, with five of them also investigating the relationship with short-term sick leave.20–24 The remaining study, that did not distinguish between duration of sick leave, found overweight to be significantly related with sick leave.25 As regards the relationship between overweight and long-term sick leave, four studies found a significant relationship in that overweight workers were at increased risk of long-term sick leave.20–24 Despite a positive trend, the remaining three studies did not observe significance.19,23,28 For example, the study of Christensen et al.19 among 5366 Danish workers found a hazard ratio of 1.38 (95% CI: 0.97–1.95) for the onset of long-term sickness during the 18-month follow-up period. Owing to inconsistency, there was inconclusive evidence of a significant relationship between overweight and long-term sick leave. This conclusion was also true for the relationship with shorter spells of sick leave, as results between the studies were inconsistent. As mentioned, some studies investigated the relationship with both short- and long-term sick leave. For example, Ferrie et al.20 showed that overweight was a significant predictor of both short and long spells of sickness absence among 10 308 British civil servants, male, and female. In International Journal of Obesity

810


Table 2

Summary and methodological quality score of longitudinal studies on the relationship between overweight and obesity and sick leave MQS

Study population

Follow-up

Measurement of overweight/obesity

Measurement of sick leave

Results

17

83%

326 (out of 370) male welders (mean age 41.5), metal workers (mean age 39.8), and office clerks (mean age 39.9) from a construction company of large objects (bridges, oilrigs) in The Netherlands; 251 after follow-up

2 years

Weight and height evaluated by self-report

Absence due to respiratory sickness, reported to the administration office

Absence due to respiratory complaints was not associated with BMI

18

83%

283 (out of 312 ) male welders (mean age 41.5) and metal workers (mean age 39.8) involved in fabrication of large constructions in The Netherlands; 222 at follow-up

2 years

Weight and height evaluated by self-report

Absence due to musculoskeletal problems, reported to the administration office

BMI was not associated with sick leave in a multivariate model

19

75%

5366 Danish employees aged 18–69; 2580 men and 2440 women at follow-up

18 months

Weight and height evaluated by self-report Normal: BMI 18.5–24.9 kg/m2 Overweight: BMI 25.0–29.9 kg/m2 Obesity: BMI X30.0 kg/m2

Data from national register on social transfer payments

Hazard ratios (HR) for onset of long-term sickness during the follow-up Normal weight women: HR ¼ 1.00 Overweight women: HR ¼ 1.38 (CI ¼ 0.97–1.95) Obese women: HR ¼ 1.54 (CI ¼ 0.87–2.75) Normal weight men: HR ¼ 1.00 Overweight men: HR ¼ 1.38 (CI ¼ 0.95–1.99) Obese men: HR ¼ 1.31 (CI ¼ 0.75–2.92)

20

83%

10 308 (out of 14121) employees working in 20 civil service departments in London (UK); 2564 female (mean age 45.3) and 5853 male (mean age 44.1) at follow-up

7 years (mean)

Weight and height measured by physical examination Reference: BMI 21.1–22.9 kg/m2 Overweight: BMI 25.0–29.9 kg/m2 Obesity: BMI X30 kg/m2

Data from civil service pay centre; 47 days of absence required a medical certificate

BMI as a predictor (rate ratios ¼ RR) of short and long spells of sickness absence Short spells in overweight women: RR ¼ 1.15 (CI ¼ 1.05–1.26) Short spells in obese women: RR ¼ 1.15 (CI ¼ 1.02–1.29) Long spells in overweight women: RR ¼ 1.39 (CI ¼ 1.22–1.57) Long spells in obese women: RR ¼ 1.52 (CI ¼ 1.30–1.76) Short spells in overweight men: RR ¼ 1.13 (CI ¼ 1.05–1.21) Short spells in obese men: RR ¼ 1.19 (CI ¼ 1.06–1.33) Long spells in overweight men: RR ¼ 1.16 (CI ¼ 1.04–1.29) Long spells in obese men: RR ¼ 1.49 (CI ¼ 1.27–1.75)

21

75%

1284 employees (29% female) aged 18–59 (mean age 36), working in industrial, administrative, and service sector; 1191 at follow-up

3 years

Weight and height measured by physical examination Normal weight: BMI o25 kg/m2 Overweight: BMI 25–29.9 kg/m2 Obesity: BMI 430 kg/m2

Data from company records Very short-term absenteeism: 1–2 Short-term absenteeism: 3–7 days Medium long-term absenteeism: 8–21 days Long-term absenteeism: 422 days

No significant differences were found for mean frequency of very short-term absenteeism No significant differences were found for mean frequency of short-term absenteeism Compared to normal weight, overweight and obesity had significant (Po0.005) higher mean frequency of medium long-term absenteeism


Reference number

Table 2 (continued) Reference number

MQS

Study population

Follow-up



Results

Compared to normal weight, obesity had significant (Po0.005) higher mean frequency of long-term absenteeism Compared to normal weight, obesity had significant (Po0.05) higher total mean frequency of absenteeism Compared to normal weight, obesity had significant (Po0.05) higher mean total duration (in days) of absenteeism At follow-up, 1452 middle-aged men and 5386 middle-age women employed at the city of Helsinki

2.9 years (mean)

Weight and height evaluated by selfreport Normal weight: BMI o25 kg/m2 Overweight: BMI 25–29.9 kg/m2 Obesity: BMI 430 kg/m2

Data from company records Short spells: 1–3 days Long spells: 43 days (medical certificate was required)

Short spells in normal weight men: RR ¼ 1.0 Short spells in overweight men: RR ¼ 1.11 (CI ¼ 0.94–1.31) Short spells in obese men: RR ¼ 1.32 (CI ¼ 1.06–1.63) Long spells in normal weight men: RR ¼ 1.0 Long spells in overweight men: RR ¼ 1.42 (CI ¼ 1.17–1.72) Long spells in obese men: RR ¼ 1.64 (CI ¼ 1.64–2.62) Short spells in normal weight women: RR ¼ 1.0 Short spells in overweight women: RR ¼ 1.10 (CI ¼ 1.02–1.18) Short spells in obese women: RR ¼ 1.48 (CI ¼ 1.37–1.61) Long spells in normal weight women: RR ¼ 1.0 Long spells in overweight women: RR ¼ 1.27 (CI ¼ 1.17–1.38) Long spells in obese women: RR ¼ 1.77 (CI ¼ 1.61–1.94)

23

58%

Random sample of 5574 Danish employees aged 18–64 years; 1672 women and 2120 men at follow-up

5 years

Weight and height evaluated by self-report Normal weight: BMI 18.5–24.9 kg/m2 Overweight: BMI 25.0–29.9 kg/m2 Obesity: BMI X30 kg/m2

Self-reported sick leave of 12 months before follow-up survey Low sick leave: p6 days High sick leave: 47 days

BMI as risk factor for sick leave above mean (6.27 days in the last 12 months) Normal weight: reference Overweight: OR ¼ 1.01 (CI ¼ 0.82–1.24) Obesity: OR ¼ 1.57 (CI ¼ 1.09–2.25)

24

83%

21 419 (out of 44 623) employees, aged 35–59, from 25 large companies or public administrations in Belgium. At follow-up, 15 557 males and 4906 females

1 year

Weight, height, and waist circumference was measured by physical examination Normal weight: BMI o25 kg/m2 Overweight: BMI 25–29.9 kg/m2 Obesity: BMI X30 kg/m2 Waist circumference cut-off points in men Normal: o94 cm Overweight: 94–101.9 Obesity: X102 Waist circumference cut-off points in women

Data from company records Calendar days of absence form work with medical certification (partial days off from work were counted as full days) Short spell: p7 days Long spell: 47 days High sick leave: X75th percentile

BMI as a predictor (OR) of short spells of sickness absence: Normal weight: reference Overweight men: OR ¼ 0.99 (CI ¼ 0.90–1.08) Obese men: OR ¼ 1.09 (CI ¼ 0.96–1.25) Overweight women: OR ¼ 1.04 (CI ¼ 0.87–1.23) Obese women: OR ¼ 1.04 (CI ¼ 0.83–1.32) BMI as a predictor (OR) of long spells of sickness absence: Overweight men: OR ¼ 1.05 (CI ¼ 1.02–1.29) Obese men: OR ¼ 1.52 (CI ¼ 1.30–1.78) Overweight women: OR ¼ 1.06 (CI ¼ 0.87–1.29) Obese women: OR ¼ 1.42 (CI ¼ 1.11–1.82) BMI as a predictor (OR) of high (XP75) incidence of sickness absence:


67%

811


22

812


Table 2 (continued) Reference number

MQS

Study population

Follow-up



Results

Normal: o80 cm Overweight: 80–87.9 Obesity: X88

25

75%

11 728 (out of 74 060) employees of Duke University Health System and Duke University (69.5% women, 44.5% 15–34 years, 48.1% 35–44 years, 7.5% X55 years)

9 year study period with a dynamic follow-up time

Weight and height measured by physical examination Normal: BMI 18.5–24.9 kg/m2 Overweight: BMI 25–29.9 kg/m2 Obesity class I: BMI 30–34.9 kg/m2 Obesity class II: BMI 35–39.9 kg/m2 Obesity class III: BMI X40 kg/m2

Data on lost workdays from university occupational health service administration

Normal weight: RR ¼ 1.00 Overweight: RR ¼ 3.43 (CI ¼ 2.81–4.17) Obesity class I: RR ¼ 3.39 (CI ¼ 2.74–4.20) Obesity class II: RR ¼ 5.43 (CI ¼ 4.34–6.80) Obesity class III: RR ¼ 8.04 (CI ¼ 6.42–10.07)

26

92%

255 (out of 226) female student nurses, aged 18–25, in the UK; 185 students at follow-up

33 months

Weight and height measured by physical examination

Data from school records. Absence p3 days of sick leave, NOT requiring a medical certificate Sickness 43 days of sick leave, requiring a medical certificate

BMI was related to sick leave in regression analysis. ¼ 0.0014 (Po0.025), with a explained variance of 14%

27

50%

1910 male (mean age 37.6) and 377 female (mean age 29.5) employees of a Louisiana (US) oil refinery

6.4 years (mean)

Weight and height evaluated by self-report Obesity in men: BMI X27.2 kg/m2 Obesity in women: BMI X26.9 kg/m2

Data from company records Illness absence days of 46 days. Pregnancy and childbirth-related absence excluded

Average duration of absence in not-obese women: 4 days Average duration of absence in obese women: 5 days Average duration of absence in not-obese men: 3.8 days Average duration of absence in obese men: 7.4 days

28

58%

2097 male and 453 female employees of a Texas (US) oil refinery. Mean age at end of study period was 46 years

10 year study period with dynamic follow-up time

Weight and height measured on physical examination Normal weight: BMI o25 kg/m2 Overweight: BMI 25.0–29.9 kg/m2 Obesity: BMI X30 kg/m2

Data from company records Absences lasting 6 days or more

Absence frequencies per 100 male employees was 20.3, 22.4, and 28.8 days for, respectively, normal weight, overweight and obese men Absence frequencies per 100 female employees was 13.3, 22.2, and 47.7 days for,


Overweight men: OR ¼ 1.12 (CI ¼ 1.01–1.25) Obese men: OR ¼ 1.50 (CI ¼ 1.30–1.73) Overweight women: 1.16 (CI ¼ 0.96–1.42) Obese women: 1.63 (CI ¼ 1.27–2.09) Normal waist circumference: reference Waist circumference as a predictor (OR) of short spells of sickness absence: Overweight men: 1.04 (CI ¼ 0.94–1.15) Obese men: 1.07 (CI ¼ 0.96–1.19) Overweight women: 1.06 (CI ¼ 0.88–1.28) Obese women: 1.09(CI ¼ 0.90–1.32) Waist circumference as a predictor (OR) of long spells of sickness absence: Overweight men: 1.13 (CI ¼ 1.00–1.28) Obese men: 1.50 (CI ¼ 1.32–1.71) Overweight women: 1.14 (CI ¼ 0.92–1.41) Obese women: 1.33 (CI ¼ 1.08–1.63) Waist circumference as a predictor (OR) of high (XP75) incidence of sickness absence: Overweight men: 1.13 (CI ¼ 1.01–1.26) Obese men: 1.48 (CI ¼ 1.31–1.67) Overweight women: 1.20 (CI ¼ 0.97–1.50) Obese women: 1.55 (CI ¼ 1.25–1.91)

Relationship between overweight and sick leave: a review DC van Duijvenbode et al Compared with not-obese, obese women had a RR of 1.3 (CI ¼ 1.0–1.6) for having long-term sick leave

Abbreviations: BMI, body mass index; CI, confidence interval; MQS, methodological quality score; NS, not significant; OR, odds ratio; RR, rate ratio.

Company records Long-term sick-listing was defined as sick-listed 28 days or more requiring a medical certificate 29

92%

5424 (out of 9000) women from Sweden employed in the public sector

3 years

Weight and height evaluated by selfreport Not obese: BMI o30 kg/m2 Obesity BMI X30 kg/m2

respectively, normal weight, overweight and obese women Obese women had significant (Po0.05) higher absence frequencies compared with normal weight women Number of workdays lost per male employee were 6.8, 7.9, and 10.5 for, respectively, normal weight, overweight and obese men Number of workdays lost per female employee were 7.7, 15.1, and 21.8 for, respectively, normal weight, overweight and obese women Obese men and women had significant (Po0.05) greater workdays lost compared with normal weight men and women

Measurement of sick leave MQS Reference number

Table 2 (continued)

Study population

Follow-up


Results

813 Table 3

Quality scores of the longitudinal studies

First author (year of publication)

A

B

C

D

E

F

G

H

I

J

K

L

Total

Alexopoulos (2007) Burdorf (1998) Christensen (2007) Ferrie (2007) Jans (2007) Laaksonen (2007) Labriola (2006) Moreau (2004) Østbye (2007) Parkes (1986) Tsai (1997) Tsai (2005) Vinga˚rd (2005)

1 1 1 1 1 1 1 1 1 1 1 1 1

1 1 0 1 1 0 0 0 1 1 1 1 1

1 1 1 0 0 0 0 0 0 1 0 1 1

1 1 0 1 1 1 0 1 0 1 0 0 1

0 0 0 1 1 0 0 1 1 1 0 1 1

1 1 1 1 1 0 1 1 1 1 1 1 1

1 1 1 1 0 1 1 1 1 1 0 0 1

1 1 1 1 1 1 0 1 1 1 1 1 0

1 1 1 0 1 1 1 1 0 1 0 0 1

0 0 1 1 0 1 1 1 1 0 0 0 1

1 1 1 1 1 1 1 1 1 1 1 0 1

1 1 1 1 1 1 1 1 1 1 1 1 1

83% 83% 75% 83% 75% 67% 58% 83% 75% 92% 50% 58% 92%

their study, odds ratios (OR) varied from 1.13 to 1.39 with slightly greater ORs for long spells of sickness absence than for short spells.20 On the other hand, some studies found different results for the separate sick leave measures investigated. To illustrate, in a heterogeneous Dutch working population, Jans et al.21 showed positive relationships between overweight and both medium-term (8–21 days) and long-term (422 days) sick leave, but not for very short (1–2 days) and short-term (3–7 days) sick leave. In addition, Laaksonen et al.22 found no relationship between overweight and short spells (1–3 days) among men, whereas overweight male and female employees were at increased risk of sick leave of more than 3 days; ORs were 1.42 and 1.27 for men and women, respectively.22

Obesity The relationship between obesity and sick leave was investigated in 10 longitudinal studies,19–25,27–29 of which one was of low quality.27 The low-quality study observed differences in average absence duration of 3.6 and 1.0 days between obese and non-obese male and female employees of an oil refinery.27 Eight high-quality studies investigated the relationship with long-term sick leave,19–24,28,29 of which five also investigated the relationship with shorter spells.20–24 A study among civil servants showed that obesity either at age 25 or phase 1 (age 35–55 years) increased long spells (47 days) of sickness absence by about 60% in both men and women, as well as increased short spells of absence in men by 25%.20 Another study among Swedish women working in the public sector showed that during the 3-year follow-up period, obese women were at increased risk (OR: 1.3, 95% CI: 1.0–1.6) of long-term sick leave (X28 days) after adjustment compared with their non-obese counterparts.29 In the studies investigating the relationship with long-term sick leave, a consistent pattern was seen in that all except one19 found a significant relationship between obesity and longterm sick leave. The study without statistical significant results, however, found a positive trend for the relationship International Journal of Obesity


814 between obesity and the onset of long-term sickness for men and women (OR: 1.31, 95% CI: 0.75–2.92 and OR: 1.54, 95% CI: 0.87–2.75, respectively).19 In conclusion, using our best-evidence synthesis, we found strong evidence for a positive relationship between obesity and long-term sick leave. As regards the relationship between obesity and short-term sick leave, and there was inconsistency between the studies identified, the conclusion was inconclusive evidence. Despite a positive trend, two of the five could not detect statistical significance between obesity and short-term sick leave.21,24 Although Jans et al.21 showed that obese workers were on average 14 days a year more absent than their normal-weight co-workers, no relationship with the frequency of very short-term and short-term absenteeism was observed. In addition, Moreau et al.24 examined the relationship of obesity measured by both BMI and waist circumference with different measures of sick leave, that is short spells (p7 days), long spells (47 days), and high incidence (XP75). They found consistent results with obese male and female employees, defined by BMI and waist circumference, being at increased odds for long spells of sickness absence as well as for high sick leave incidence. In contrast, non-significant relationships were found for obesity and sickness absence of 7 days or fewer.24

Table 4 Overview of high-quality longitudinal studies showing a positive (+) or no relationship between overweight and obesity and sick leave Relationship

Reference number

Overweight +

20 21 22 24 25

0

19 21 22

Discussion The aim of this review was systematically to assess the evidence for a relationship between overweight and obesity and sick leave. As described in Introduction section, investigating this relationship is of relevance for several reasons. As is well known, obesity has an enormous public health and consequently economic impact. Although diverse studies have been conducted in the past few decades, as far as we are aware no systematic review of the available literature taking into account the quality of the studies was performed. Moreover, insight into this relationship is of International Journal of Obesity

((very) short sick leave) (men: short sick leave)

23 24

(men and women: short sick leave, women: long and incidence of sick leave)

28

Obesity +

20 21

(medium and long sick leave)

22 23 24

(men and women: long and incidence of sick leave)

25 28

(men and women: duration of sick leave, women: frequency of sick leave)

29

0

19 21 24 28

Relationship of BMI and sick leave Three prospective studies of high quality examined the relationship with BMI as a continuous variable and a measure of sick leave. Two of these specified the cause of sick leave, that is absence owing to respiratory complaints17 and musculoskeletal problems,18 and did not detect a relationship between BMI and sick leave. The third study investigated the relationship between objectively measured body weight and height and sick leave, regardless of the cause.26 From their regression analysis, they found a positive relationship with relative body weight accounting for 14% of the explained variance in sickness absence over a 33-month period.26 Given these contradictory results, inconclusive evidence was found for a relationship between BMI and sick leave.

(medium and long sick leave) (men: long sick leave, women: short and long sick leave) (men: long and incidence of sick leave)

((very) short sick leave) (men and women: short sick leave) (men: frequency of sick leave)

BMI (continuous variable) 26 + 17 0 18

Note: In case of different (contradictory) results within one study, the results of separate analyses are written between brackets.

interest to those bearing (part of) the costs. Despite differences in social security systems between countries worldwide, the employer bears a substantial part of the indirect costs and/or direct costs. Thus, from the employer’s perspective, insight into the effect of obesity on sick leave, the latter being an important contributor of the indirect costs, is of relevance, and may help policymakers, such as employers, justify implementation of health promotion programmes. On the basis of the studies identified and our best-evidence synthesis, we concluded ‘inconclusive evidence’ for overweight to be related to either long-term or short-term sick leave. This conclusion is, however, positively relevant because it was the result of inconsistent findings between the studies in that not all studies observed statistical significant differences in sick leave between overweight workers and normal-weight workers. In all the studies lacking statistical significance, however, a positive trend was discerned in that overweight workers were at increased risk of sick leave compared with their normal-weight counterparts. It can, therefore, be questioned why some studies found statistical significance, whereas others did not.


815 Although possible explanations are expected to be found in study characteristics, such as sample size or follow-up period, looking at the studies in detail, we could not detect the cause for the ‘inconsistency’. It also seemed that the effect sizes of studies finding statistical significance did not visibly differ from those with lack of statistical significance. In this respect, it may be that when ‘clinical’ relevance instead of statistical significance is considered, results could tend to the conclusion of a positive relationship between overweight and sick leave. For obesity, there was consistency with obesity being a predictor of sick leave, and long-term sick leave in particular. Thus, we concluded ‘strong evidence’ for obesity and longterm sick leave. Our conclusions were not entirely in line with earlier reviews on a similar topic, reporting an association between overweight or obesity and sick leave.12,13,31 Plausible explanations for the different conclusions are thus differences in the purpose of the review, and thereby the method of the review. For example, we included studies with a longitudinal design only, whereas earlier reviews included a lower number of studies, which were predominantly crosssectional in design.12,13,31 The remarkably low number of longitudinal studies included in the earlier reviews could be because the present review was performed a few years later, next to the primary purpose of the review. Namely, they were interested in other outcomes, such as the economic impact of obesity. As this review specifically focused on the relationship between overweight and obesity and sick leave, and because of the extensive search strategy applied, we identified more studies than earlier reviews. The systematic approach in the literature search, as well as in quality assessment and conclusion as to the evidence, are among the strengths of this review. Moreover, we distinguished between overweight and obesity, as well as between short- and longterm duration of sick leave. Separate analysis of short-term and long-term sick leave is meaningful from the hypothesis that obese persons are more chronically ill, and consequently have more health care consumption as well as more absence from work owing to sickness than their non-obese counterparts. Although our review confirmed the hypothesis that obese workers have more sick leave, and especially longer-term sick leave, so far this relationship has not been examined in a review similar to the present one. From a review in 2006 summarising the literature on the relationship between obesity and the four most prevalent disabling conditions totalling over 50% of all disability benefit claims (i.e. arthritis, mental health disorders, learning disabilities, and back ailments), we know that there is indeed a strong association between obesity and each of the four conditions.32 To date, however, it is still unknown whether the presence of certain disease conditions among obese workers leads to more sick leave. Although our review showed a relationship between obesity and especially longer-term sick leave, the question as to the mechanism between obesity and sick leave remains unanswered. Future research examining

the possible mediating role of chronic diseases in the relationship between obesity and sick leave would thus be interesting. There are also some limitations of the present review. Although diverse strategies were used to find all relevant published studies available, the possibility of selection and publication bias cannot be excluded. On the basis of our extensive search strategy and the identification of multiple studies with both a significant positive and a lack of significant relationship between overweight and/or obesity and sick leave, however, we believe the possibility of these types of bias to be small. Further, quality assessment was carried out with a standardised set of predefined criteria adapted from existing quality criteria lists.14,15 Despite the six domains of potential bias as defined by Hayden et al.15 being taken into account, there may still be bias owing to our scoring method. Namely, we scored negative in case of an unclear or missing description of the quality criterion item. This method can in fact lead to a lower-quality score than by asking the authors. We decided not to approach authors for a number of reasons, however, inter alia because of the probability that some authors cannot be located. Moreover, as all studies, except one, were already of high quality, asking the authors to check the scorings would not have led to substantially different conclusions. Studies were considered of high quality when the study scored 51% or higher. Although this cut-off point is often used in systematic reviews, it is an arbitrary cut-off point, and results could be sensitive to change in the cut-off point. Sensitivity analysis showed, however, that changing the cut-off point to 40 or 60% would not affect the results. The inconsistent results between the studies identified in this review could be explained by numerous factors, including differences in (number of) study population, the measurement of overweight and obesity, or the measurement of sick leave. As there was heterogeneity between the studies included, we decided not to pool the data and perform a meta-analysis. The difference in measures of sick leave made a comparison between the results and especially a meta-analysis particularly difficult. In conclusion, although this review found inconclusive evidence for the relationship between overweight and sick leave, a clear trend was discerned in that overweight was a predictor of especially long spells (47 days) of sick leave. As regards obesity, this review concluded with a strong evidence for a relationship with long-term sick leave. From the studies included, obesity seemed to be a significant predictor of long-term sick leave. Therefore, and considering the public and economic impact of obesity, interventions aimed at the prevention of further weight gain or the treatment of overweight and obesity are considered as relevant, both from the individual and the employer perspective. As overweight and obesity are caused by an imbalance between energy expenditure and energy consumption, an intervention aimed at the promotion of physical activity and a healthy International Journal of Obesity


816 diet is likely to fight against obesity, and subsequently against sick leave in obese workers. To date, there is no such lifestyle intervention specifically targeting overweight and obesity and evaluating its effect on sick leave. Despite this, we believe such an intervention, that is one aimed at the promotion of a healthy energy balance, to be effective in the prevention and treatment of obesity, and thereby to reduce obesity-attributable sick leave.

References 1 World Health Organization (WHO). Obesity: Preventing and Managing the Global Epidemic. Report of a WHO Consultation. World Health Organization: Geneva, 2004. 2 Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, Flegal KM. Prevalence of overweight and obesity in the United States, 1999–2004. JAMA 2006; 295: 1549–1555. 3 Cameron AJ, Welborn TA, Zimmet PZ, Dunstan DW, Owen N, Salmon J et al. Overweight and obesity in Australia: the 1999– 2000 Australian Diabetes, Obesity and Lifestyle Study (AusDiab). Med J Aust 2003; 178: 427–432. 4 Visscher TLS, Seidell JC. Time trends (1993–1997) and seasonal variation in body mass index and waist circumference in the Netherlands. Int J Obes Relat Metab Disord 2004; 28: 1309–1316. 5 Seidell JC. Societal and personal costs of obesity. Exp Clin Endocrinol Diabetes 1998; 106 (Suppl 2): 7–9. 6 Thompson D, Wolf AM. The medical-care cost burden of obesity. Obes Rev 2001; 2: 189–197. 7 Finkelstein EA, Fiebelkorn IC, Wang G. State-level estimates of annual medical expenditures attributable to obesity. Obes Res 2004; 12: 18–24. 8 Popkin BM, Kim S, Rusev ER, Du S, Zizza C. Measuring the full economic costs of diet, physical activity and obesity-related chronic diseases. Obes Rev 2006; 7: 271–293. 9 Katzmarzyk PT, Janssen I. The economic costs associated with physical inactivity and obesity in Canada: an update. Can J Appl Physiol 2004; 29: 90–115. 10 Narbro K, Jonsson E, Larsson B, Waaler H, Wedel H, Sjostrom L. Economic consequences of sick-leave and early retirement in obese Swedish women. Int J Obes Relat Metab Disord 1996; 20: 895–903. ¨ ssner S, Neovius M. Disability 11 Neovius K, Johansson K, Ro pension, employment and obesity status: a systematic review. Obes Rev 2008; 1, [E-pub ahead of print]. 12 Schmier JK, Jones ML, Halpern MT. Cost of obesity in the workplace. Scand J Work Environ Health 2006; 32: 5–11. 13 Colditz GA. Economic costs of obesity and inactivity. Med Sci Sports Exerc 1999; 31 (Suppl): 663–667. 14 Arie¨ns GA, van Mechelen W, Bongers PM, Bouter LM, van der Wal G. Physical risk factors for neck pain. Scand J Work Environ Health 2000; 26: 7–19. 15 Hayden JA, Cote P, Bombardier C. Evaluation of the quality of prognosis studies in systematic reviews. Ann Intern Med 2006; 144: 427–437.


16 Hoogendoorn WE, van Poppel MNM, Bongers PM, Koes BW, Bouter LM. Systematic review of psychosocial factors at work and private life as risk factors for back pain. Spine 2000; 25: 2114–2125. 17 Alexopoulos EC, Burdorf A. Prognostic factors for respiratory sickness absence and return to work among blue collar workers and office personnel. Occup Environ Med 2001; 58: 246–252. 18 Burdorf A, Naaktgeboren B, Post W. Prognostic factors for musculoskeletal sickness absence and return to work among welders and metal workers. Occup Environ Med 1998; 55: 490–495. 19 Christensen KB, Lund T, Labriola M, Bultmann U, Villadsen E. The impact of health behaviour on long term sickness absence: results from DWECS/DREAM. Ind Health 2007; 45: 348–351. 20 Ferrie JE, Shipley MJ, Vahtera J, Marmot MG, Klarenbach S. BMI, obesity, and sickness absence in the Whitehall II Study. Obesity 2007; 15: 1554–1564. 21 Jans MP, Heuvel van den 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. 22 Laaksonen M, Piha K, Sarlio-Lahteenkorva S. Relative weight and sickness absence. Obesity 2007; 15: 465–472. 23 Labriola M, Lund T, Burr H. Prospective study of physical and psychosocial risk factors for sickness absence. Occup Med (Lond) 2006; 56: 469–474. 24 Moreau M, Valente F, Mak R, Pelfrene E, de SP, De BG et al. Obesity, body fat distribution and incidence of sick leave in the Belgian workforce: the Belstress study. Int J Obes Relat Metab Disord 2004; 28: 574–582. 25 Østbye T, Dement JM, Krause KM. Obesity and workers’ compensation. Results from the Duke Health and Safety Surveillance System. Arch Intern Med 2007; 167: 766–773. 26 Parkes KR. Relative weight, smoking, and mental health as predictors of sickness and absence from work. J Appl Psychol 1987; 72: 275–286. 27 Tsai SP, Gilstrap EL, Colangelo TA, Menard AK, Ross CE. Illness absence at an oil refinery and petrochemical plant. J Occup Environ Med 1997; 39: 455–462. 28 Tsai SP, Wendt JK, Ahmed FS, Donnelly RP, Strawmyer TR. Illness absence patterns among employees in a petrochemical facility: impact of selected health risk factors. J Occup Environ Med 2005; 47: 838–846. 29 Vingard E, Lindberg P, Josephson M, Voss M, Heijbel B, Alfredsson L et al. Long-term sick-listing among women in the public sector and its associations with age, social situation, lifestyle, and work factors: a three-year follow-up study. Scand J Public Health 2005; 33: 370–375. 30 Tsai SP, Ahmed FS, Wendt JK, Bhojani F, Donnelly RP. The impact of obesity on illness absence and productivity in an industrial population of petrochemical workers. Ann Epidemiol 2008; 18: 8–14. 31 Aldana SG, Pronk NP. Health promotion programs, modifiable health risks, and employee absenteeism. J Occup Environ Med 2001; 43: 36–46. 32 Ells LJ, Lang R, Shield JPH, Wilkinson JR, Lidstone JSM, Coulton S et al. Obesity and disabilityFa short review. Obes Rev 2006; 7: 341–345.