International Archives of Occupational and Environmental Health

, Volume 86, Issue 8, pp 887–899

Sick leave due to back pain in a cohort of young workers

Authors

    • Department of Public Health and SurveillanceScientific Institute of Public Health
    • Section of Occupational, Environmental and Insurance Medicine, Department of Public HealthKatholieke Universiteit Leuven
  • A. Burdorf
    • Department of Public Health, Erasmus MCUniversity Medical Center Rotterdam
  • G. Crombez
    • Department of PsychologyUniversity of Ghent
  • G. Verbeke
    • Department of Public Health, Biostatistical CentreKatholieke Universiteit Leuven
  • R. Masschelein
    • Section of Occupational, Environmental and Insurance Medicine, Department of Public HealthKatholieke Universiteit Leuven
  • Ph. Mairiaux
    • Occupational Health and Health Education Unit, Department of Public HealthUniversity of Liège
  • G. F. Moens
    • Section of Occupational, Environmental and Insurance Medicine, Department of Public HealthKatholieke Universiteit Leuven
    • External Service for Prevention and Protection at Work IDEWE
  • The BelCoBack Study Group
Original Article

DOI: 10.1007/s00420-012-0824-y

Cite this article as:
Van Nieuwenhuyse, A., Burdorf, A., Crombez, G. et al. Int Arch Occup Environ Health (2013) 86: 887. doi:10.1007/s00420-012-0824-y

Abstract

Purpose

Evidence on risk factors for sick leave from prospective studies in work settings is limited. Furthermore, most available studies focused on workers with substantial low back disorders. These studies consistently report that physical work factors constitute a hindrance to work. However, it remains unclear whether the same risk factors are relevant in workers with less severe conditions or in early phases of the development of back pain. Therefore, this article aims to study risk factors for the occurrence of sick leave due to low back pain (LBP) among young workers with no or a modest history of back pain.

Methods

Participants were 716 young healthcare or distribution workers with no or minimal antecedents of LBP in the year before inclusion. We investigated the role of potential physical, psychosocial and individual risk factors at baseline on the occurrence of sick leave due to LBP 1 year later. To this purpose, we used Cox regression with a constant risk period.

Results

Six per cent (95 % CI 4.1–7.6) of the workers reported sick leave 1 year later; they accounted for 12 % of the sick-leave days independent of cause. A non-stimulating psychosocial work environment turned out to be the strongest risk factor for sick leave due to LBP (RR 6.08; 95 % CI 1.42–26.07). Physical factors were not predictive.

Conclusions

In the early phases of back pain and in less severe conditions, the main benefit of interventions lies in targeting the organisation and design of jobs to create a challenging professional environment.

Keywords

Low back painSick leavePsychosocial work environmentProspectiveYoung workers

Introduction

Back pain is a major health and economic problem in Western industrialised societies. It is a primary reason for health care use and work inability and leads to substantial costs for society and industry. Cost-of-illness studies of back pain in the Netherlands, Sweden and the U.K. have concluded that the main burden imposed by back pain is related to production loss attributable to work absenteeism. These costs supersede the health care costs (Ekman et al. 2005; Van Tulder et al. 1995; Maniadakis and Gray 2000).

In contrast to knowledge about risk factors for low back pain (LBP), knowledge about risk factors for sick leave from back disorders remains unsatisfactory (Alexanderson and Norlund 2004). First, the number of prospective studies is limited and the resulting evidence on the role of physical, psychosocial and individual factors is unclear. Many researchers assumed that a focus upon risk factors for the development of LBP would automatically prevent subsequent sick leave. Recent evidence has shown that this assumption may not be valid. It seems that different sets of factors may have to be addressed in the prevention of LBP and in the prevention of sick leave (Ijzelenberg et al. 2004; Gheldof et al. 2005; Ijzelenberg and Burdorf 2005). Therefore, there has been a call for longitudinal studies focusing on sick leave (SBU 2004). Second, almost all studies have focused on study populations with a considerable proportion of workers with a history of LBP. The magnitude of and the risk factors for back-related sick leave in individuals with no or limited antecedents of LBP are largely unknown.

Some evidence exists on factors that predict the duration of sick leave in workers in the beginning of a LBP-related sick leave episode, that is, radiating pain, high levels of disability and social dysfunction, social isolation, being an older female, and receiving a high level of compensation. With regard to work-related factors, patients with low back pain at the highest risk for long-term absence are workers doing heavy physical work. For the psychosocial factors, however, the evidence remains inconclusive. In spite of well-known effects of history of low back pain on recurrences of back pain, history of LBP does not influence the duration of sick leave due to LBP (Steenstra et al. 2005). History of LBP was reported as prognostic for a more frequent drop out from work by Wasiak et al. (2004). Similarly, musculoskeletal pain and combinations of pain predicted sickness absence spells among municipal workers during a three-year follow-up study (Kääriä et al. 2012). Apart from the effect of previous pain, the evidence of the respective roles of physical, psychosocial and individual risk factors on the frequency of sick leave remains inconclusive.

In Belgium, the compulsory social health insurance covers the entire population and is organised by sickness funds. When an employee starts sickness absence, he/she send a medical certificate of the treating physician to the employer who pays the first 2–4 weeks of work incapacity. If work incapacity continues, the patient applies for a social security benefit by sending another medical certificate to the medical adviser of the sickness fund. Claim assessment, follow-up evaluation and the decision about benefit entitlement are done by the medical adviser. There is no time limit for coverage (Du Bois and Donceel 2008).

Furthermore, every employer is obliged by law to organise occupational health care for the employee. Most employers hire the services of an external occupational health care service. A few large companies have an internal occupational health care service. The main role of the occupational physician is to prevent occupational diseases and accidents. The occupational physician is hardly involved in the sick leave process. It is only recently that employees on sick leave have been legally entitled to contact the occupational physician to discuss return to work options (Tiedtke et al. 2012).

In a sample of young workers with no or a modest history of back pain, we aimed
  1. 1.

    To describe the occurrence, the duration and the frequency of sick leave due to LBP,

     
  2. 2.

    To investigate the effect of work-related and individual factors on the occurrence of sick leave due to LBP, and

     
  3. 3.

    To assess the impact of each identified risk factor on the occurrence of sick leave among the workers exposed to one of these factors and among the entire study population.

     

Materials and methods

Subjects and methods

The BelCoBack Study (Belgian Cohort Study on Low Back Pain) is a prospective study. Methods have been described in detail before (Van Nieuwenhuyse et al. 2006). In summary, in 2000 and 2001, participants were recruited among the employees of four healthcare institutions and two distribution companies throughout Belgium, and baseline measurements were obtained. Employment in either sector is associated with elevated risks for LBP. However, the choice of these two sectors (and not, for example, the construction sector) was practical: the collaboration partner IDEWE, as an important occupational health service, recruited the participants among the employees of their clients, many of which are active in the health care and distribution sectors. The recruitment took place as a result of the annual medical examination by the occupational health physician. In Belgium, such an examination is obliged by law for workers exposed to occupational risks (Royal Decree May 28th 2003). To minimise dropout, only workers with a tenured position or equivalent were included. Furthermore, to reduce the influence of age and of prior episodes of LBP, participants had to be no older than 30 years at the time of intake and had to have been free of episodes of LBP of seven or more consecutive days during the 12 months before intake in the study. Of 1,672 eligible employees, 1,200 (72 %) agreed to participate. However, during a first contact, 159 were excluded because they did not meet the last inclusion criterion, leaving a sample of 1,041 workers. Of those 1,041 workers, 972 (93 %) completed the questionnaire at baseline.

One year later (2001–2002), participants were requested again to fill in a questionnaire. Of the 972 workers who responded at baseline, 800 (82 %) returned the questionnaire. Questionnaires at baseline and at follow-up were distributed within the companies by the research assistants. The majority of workers filled in the questionnaire at home and sent it back to the research assistants. In case of non-response, two reminders were sent.

For the longitudinal analyses described in this paper, a cohort was identified of 851 employees with a minimal experience of at least 2 months in their function at intake. An interval of at least 2 months was considered sufficient to appreciate the work constraints in a function. The questionnaire at 1 year of follow-up was available for 716 of these 851 workers (response of 84 %, loss to follow-up of 16 %).

The study protocol was approved by the local commission for medical ethics, and an informed consent was given by all included employees before their participation in the study.

Data collection

Questionnaires at baseline

At baseline, self-reported questionnaires were used to register factors that may be related to low back disorders, that is, (1) physical load at work and during leisure time, (2) psychosocial work characteristics and (3) individual variables.

Questions on current physical workload (Somville and Mairiaux 2003) addressed (1) the duration of working in awkward postures, (2) the duration of exposure to whole body vibration, (3) the intensity and, where indicated, the frequency of manual materials handling such as lifting, carrying, pushing, or pulling of loads, (4) static work postures (that is, standing and sitting for long periods) and (5) ability to change posture regularly. Duration, frequency and intensity were rated on three- or four-point ordinal scales. Furthermore, we addressed the seniority in the current function and the working schedule (percentage of employment, day or night duty). Additional questions on (at least weekly) sporting activities, engagement in construction and embellishment work at home, and on motor vehicle driving outside the work (km/year) served to assess the physical load during leisure time.

Psychosocial work characteristics were evaluated with the 43-item Job Content Questionnaire (Karasek and Theorell 1990). The different items were measured on four-point Likert scales, ranging from ‘completely disagree’ to ‘completely agree’, yielding a sum score for each dimension. Based on the Demand-Control-Support model of Karasek and Theorell, the following dimensions were taken into account: skill discretion (six items), decision authority (three items), psychological job demands (five items), supervisor and co-worker support (four items each), job insecurity (five items), and job dissatisfaction (five items). For the analyses, the psychosocial work characteristics were categorised into tertiles.

Individual variables included the following: (1) age, sex, language and educational level as demographic factors; (2) smoking behaviour, body mass index, perceived general health and complaints of the neck, back, upper or lower limbs in the year before inclusion as health-related factors; and (3) pain-related fear, catastrophising about pain, negative affectivity and somatisation as psychological factors. The questionnaire on individual and health-related factors was derived from the standardised Nordic Questionnaires for the analysis of musculoskeletal symptoms (Kuorinka et al. 1987). For the assessment of psychological concepts, we used the Modified Tampa Scale of Kinesiophobia (Vlaeyen and Crombez 1998), the Pain Catastrophising Scale (Sullivan et al. 1995), the Positive Affectivity Negative Affectivity Scales (Watson et al. 1988) and an adapted version (29 items) of the Psychosomatic Symptom Checklist (Van Dixhoorn and Duivenvoorden 1985), respectively. All items were scored on four- or five-point Likert scales, and for each concept, a total score was calculated. For the analyses, these scores were split up into tertiles. Body mass index (BMI) was categorised as BMI <20, BMI 20–<25 (normal), BMI 25–<30 (overweight) and BMI ≥30 (obese). Language was collected since a previous study in Belgium has shown that back injuries with a longer sick leave were more prominent in the French-speaking part of the country compared to the Flemish-speaking communities. As such, language may represent subtle cultural language-linked factors and/or regional differences in economic climate that play a role in the sick leaving process (Mazina et al. 2012).

Questionnaires at 1 year of follow-up

One year later, participants completed another questionnaire. At that moment, workers were asked, among others, whether they had stayed at home because of low back complaints since the start of the study about 12 months ago (yes/no). The occurrence of sick leave due to LBP after 1 year of follow-up, that is, the outcome for this article, was thus registered as the proportion of workers who stayed home because of LBP in the first year of follow-up. Sick leave was defined as any absence from work, however, short. If ‘yes’, participants were asked to further detail the number of sick leave spells (‘how many times’) and the total duration of sick leave (‘how many days in total’) (Kuorinka et al. 1987; Von Korff et al. 1992). As described, all sick leaves from work in Belgium have to be medically certified and are then compensated.

Analytic methods

Univariate analyses were performed by means of chi-square or Fisher exact tests (categorical variables) and Mann–Whitney U or unpaired T tests (continuous variables). For the multivariate analyses, we opted for a Cox regression with a constant period of risk for all subjects in order to obtain relative risk estimates rather than odds ratios, which would result from logistic regression (Thompson et al. 1998). Age and gender were included as confounders, irrespective of their relation with LBP. Variables that met the 10 % level of significance in the univariate analyses were considered for inclusion in the multivariate analyses. We calculated correlation coefficients among these variables as an approximate manner to detect possible multicollinearity. In the final multivariate models, backward selection was used retaining variables with a P value less than 0.05. Analyses were conducted with the SPSS package (version 18). In a last step, we calculated the attributable fraction among the exposed workers (AFexposed) as well as the population attributable fraction among the entire study population (PAFtotal) for those variables that were found to constitute risk factors. The AFexposed estimates the fraction of exposed cases that would not have occurred if exposure had not occurred. The PAFtotal estimates the fraction of all cases that would not have occurred if exposure had not occurred (Rothman and Greenland 1998).

Results

Descriptive statistics

Risk factors

The study characteristics for the 716 participants are given in Tables 1 and 2. The Table is structured as follows: (1) physical load at work and during leisure time, (2) psychosocial work characteristics and (3) individual variables. This last group of variables includes demographic, health-related and psychological factors. Psychosocial work characteristics and psychological factors have been categorised into tertiles. Categorical variables are given in Table 1 and continuous variables in Table 2.
Table 1

Categorical characteristics of the study population (n = 716)

Variable

n

%

Current physical load

  

Professional

  

 Bent and twisted position

  

  No

448

63.5

  ≤2 h/day

181

25.7

  >2 h/day

76

10.8

 Driving vehicles or machines

  

  No

408

57.8

  ≤6 h/day

141

19.9

  >6 h/day

158

22.3

 Pushing or pulling of loads

  

  No

324

45.7

  <1 time/h

201

28.3

  ≥1 time/h

184

26.0

 Lifting or carrying of loads

  

  No

123

17.5

  ≤10 kg

84

11.9

  >10 kg, ≤25 kg, ≤12 times an hour

122

17.4

  >10 kg, ≤25 kg, >12 times an hour

50

7.1

  >25 kg, ≤12 times an hour

311

44.3

  >25 kg, >12 times an hour

13

1.8

 Sitting for long periods

  

  Yes

138

19.4

 Standing for long periods

  

  Yes

182

25.9

 Ability to change posture regularly

  

  No

90

12.7

 Working schedule

  

  Night duty

185

26.2

  Day duty

522

73.8

 Percentage of employment

  

  More than 75 %

626

88.3

Extra-professional

  

 Sporting activities at least weekly

  

  No

326

45.7

 Embellishment works at home

  

  Yes

418

59.7

 Construction works at home

  

  Yes

145

21.9

Psychosocial work load

  

Possibilities to develop skills

  

 Low (≤32)

197

28.2

 Moderate (>32–≤36)

195

27.9

 High (>36)

307

43.9

Decision authority

  

 Low (≤28)

137

19.5

 Moderate (>28–≤36)

301

43.0

 High (>36)

263

37.5

Psychological job demands

  

 Low (≤30)

269

40.0

 Moderate (>30–≤34)

200

29.7

 High (>34)

204

30.3

Supervisor support

  

 Low (≤11)

138

20.1

 Moderate (>11–≤12)

207

30.2

 High (>12)

340

49.7

Co-worker support

  

 Low (≤12)

156

22.9

 Moderate (>12–≤13)

76

11.2

 High (>13)

448

65.9

Job insecurity

  

 Low (≤8)

246

35.9

 Moderate (>8–≤10)

250

36.5

 High (>10)

189

27.6

Job dissatisfaction

  

 Low (≤9)

301

44.2

 Moderate (>9–≤11)

221

32.5

 High (>11)

159

23.3

Individual variables

  

Gender

  

 Women

433

60.5

Language

  

 Dutch-speaking

504

70.4

 French-speaking

212

29.6

Education

  

 Higher university

41

5.8

 Higher non-university

263

37.0

 Higher secondary/professional

281

39.5

 No diploma–primary school–lower secondary/professional

126

17.7

Smoking

  

 Never smoked

430

61.1

 Ex-smoker

88

12.5

 Current smoker

186

26.4

Body mass index BMI (kg/m2)

  

 Normal (≥20–<25)

379

56.4

 Underweight (<20)

110

16.4

 Overweight (≥25–<30)

132

19.6

 Obese (≥30)

51

7.6

Perceived general health

  

 Very good

339

47.4

 Moderate to fair

376

52.6

In the 12 months before inclusion

  

 Complaints of the neck

  

  Yes

194

27.1

 Complaints of the back

  

  Yes

345

48.2

 Complaints of the upper limbs

  

  Yes

134

18.7

 Complaints of the lower limbs

  

  Yes

180

25.1

Pain-related fear

  

 Low (≤35)

240

33.6

 Moderate (>35–≤41)

238

33.4

 High (>41)

235

33.0

Catastrophising of pain

  

 Low (≤10)

237

33.2

 Moderate (>10–≤17)

249

35.0

 High (>18)

227

31.8

Somatisation

  

 Low (≤16)

240

33.6

 Moderate (>16–≤20)

240

33.6

 High (>20)

235

32.8

Negative affectivity

  

 Low (≤47)

288

40.2

 Moderate (>47–≤58)

202

28.3

 High (>58)

225

31.5

Table 2

Continuous characteristics of the study population (n = 716)

Variable

Mean

Median

SD

IQR

Current physical load

    

 Seniority in the current function (years)

3.6

3.0

2.7

1.0–5.0

 Motor vehicle driving outside the work (km/year)

18,674.0

15,000.0

22,624.0

10,000.0–25,000.0

Individual variables

    

Age (years)

26.2

26.0

2.7

24–29

SD standard deviation of the mean, IQR interquartile range

Out of the 716 workers, 64 % were employing in the health care sector and 36 % in the distribution sector and 61 % were women. We registered a median age of 26 years (interquartile range of 5 years) and a median seniority in the current function of 3 years (interquartile range of 4 years); 88 % had a full-time employment.

The majority of the participants reported either higher education of non-university level (37 %) or higher secondary/professional education (40 %). Although we had tried to limit previous antecedents of back pain by including only workers free of episodes of LBP lasting seven consecutive days or more in the year before inclusion, 48 % of the participants reported pain of the lower or upper back in the 12 months before inclusion. More specifically, 10 % reported some pain in the upper back and 43 % reported some LBP (but not lasting seven consecutive days or more) in the 12 months before inclusion. Thirty-eight workers (5 %) interrupted their normal activities at home, outside the home or at work because of LBP in the year before inclusion; the median number of days of these interruptions was 4.5 days with an interquartile range between 2 and 7 days. In this young population aged 30 years or younger, 47 % rated their health as ‘very good’; the remaining 53 % gave a rating of ‘moderate to fair’.

Study outcome

After 1 year of follow-up, 42 (6 %, 95 % CI 4–8) of the 716 workers reported sick leave from work due to LBP. There was no significant difference between men and women (Table 3). Although only 42 workers reported back-related sick leave, sick leave due to LBP was substantial with a median duration of 10 days (interquartile range between 5 and 21 days). Notably, 35 (84 %) of the 42 workers took sick leave as a single episode. Collectively, LBP-related sick leave accounted for 696 days or 12 % of all the sick-leave days in the population of 716 workers.
Table 3

Occurrence of sick leave due to LBP after 1 year of follow-up

 

n

Sick leave due to LBP at follow-up

n

%

95 % CI

Study population

716

42

5.9

(4.1–7.6)

Men

283

21

7.4

(4.4–10.5)

Women

433

21

4.9

(2.8–6.9)

Risk factors for the occurrence of sick leave due to LBP

Univariate analyses

Results of univariate analyses are presented in Tables 4 (categorical variables) and 5 (continuous variables). With regard to the physical load, two factors were significantly related to back-related sick leave in the following year, that is, motor vehicle driving outside work (Mann–Whitney U test, P = 0.018) and driving vehicles or machines at work. For the latter, a moderately elevated risk was observed for driving up to 6 h a day (RR 2.54, 95 % CI 1.27–5.08), but there was no evidence for a clear dose–response relationship. None of the variables concerning awkward working postures, manual materials handling or static work postures were predictive. As to the psychosocial work load, one dimension of the Karasek model turned out to be predictive for sick leave in the following year, that is, a lack of ‘possibilities to develop skills’. Three- to four-fold elevated risks were observed for workers who reported low (RR 3.86, 95 % CI 1.36–10.99) or moderate (RR 3.06, 95 % CI 1.01–9.35) ‘possibilities to develop skills’ at work in comparison with their colleagues who indicated high possibilities at baseline. Three individual factors were related to the occurrence of sick leave in the following year: (1) general health perceived as moderate to fair (RR 2.82, 95 % CI 1.40–5.56), (2) obesity defined as a BMI of 30 or higher (RR 2.94, 95 % CI 1.29–6.67) and (3) complaints of the upper limbs in the year before inclusion (RR 2.20, 95 % CI 1.19–4.07). None of the psychological variables predicted sick leave (neither as categorical nor as continuous variables).
Table 4

Categorical risk factors for the occurrence of sick leave due to low back pain after 1 year of follow-up (SLLBP at t1) in univariate analyses

Variable at baseline

SLLBP at t1

n

n

%

*P value

RR

95 % CI

Current physical load

     

Professional

      

 Bent and twisted position

      

  No

448

23

5.1

0.173

1.00

 

  ≤2 h/day

179

11

6.1

 

1.20

(0.60; 2.40)

  >2 h/day

75

8

10.7

 

2.08

(0.97; 4.46)

 Driving vehicles or machines at work

      

  No

406

16

3.9

0.024

1.00

 

  ≤6 h/day

140

14

10.0

 

2.54

(1.27–5.08)

  >6 h/day

158

11

7.0

 

1.77

(0.84–3.72)

 Pushing or pulling of loads

      

  No

324

15

4.6

0.389

1.00

 

  <1 time/hour

200

15

7.5

 

1.62

(0.81; 3.25)

  ≥1 time/hour

182

11

6.0

 

1.31

(0.61; 2.79)

 Lifting or carrying of loads

      

  No

123

5

4.1

0.127

1.00

 

  ≤10 kg

84

7

8.3

 

2.05

(0.67; 6.25)

  >10 kg, ≤25 kg, ≤12 times an hour

121

10

8.3

 

2.03

(0.72; 5.78)

  >10 kg, ≤25 kg, >12 times an hour

50

4

8.0

 

1.97

(0.55; 7.04)

  >25 kg, ≤12 times an hour

310

13

4.2

 

1.03

(0.38; 2.83)

  >25 kg, >12 times an hour

12

2

16.7

 

4.10

(0.89; 18.87)

 Sitting for long periods

      

  No

570

36

6.3

0.224

1.00

 

  Yes

138

5

3.6

 

0.57

(0.23–1.43)

 Standing for long periods

      

  No

521

31

6.0

0.858

1.00

 

  Yes

179

10

5.6

 

0.94

(0.47–1.88)

 Ability to change posture regularly

      

  Yes

617

33

5.3

0.076

1.00

 

  No

89

9

10.1

 

1.89

(0.94–3.82)

 Working schedule

      

  Night duty

185

9

4.9

0.576

1.00

 

  Day duty

519

31

6.0

 

1.23

(0.60–2.53)

 Percentage of employment

      

  More than 75 %

624

37

5.9

1.000

1.00

 

  75 % or less

82

5

6.1

 

0.97

(0.39–2.40)

Extra-professional

      

 Sporting activities at least weekly

      

  Yes

386

20

5.2

0.371

1.00

 

  No

325

22

6.8

 

1.31

(0.73–2.35)

 Embellishment works at home

      

  No

281

12

4.3

0.171

1.00

 

  Yes

416

28

6.7

 

1.58

(0.82–3.05)

 Construction works at home

      

  No

514

31

6.0

0.583

1.00

 

  Yes

145

7

4.8

 

0.80

(0.36–1.78)

Psychosocial work load

     

Possibilities to develop skills

      

 High

197

4

2.0

0.023

1.00

 

 Moderate

193

12

6.2

 

3.06

(1.01; 9.35)

 Low

306

24

7.8

 

3.86

(1.36; 10.99)

Decision authority

      

 Low

137

9

6.6

0.888

1.00

 

 Moderate

301

18

6.0

 

0.91

(0.42; 1.98)

 High

260

14

5.4

 

0.82

(0.36; 1.85)

Psychological job demands

      

 Low

268

21

7.8

0.101

1.00

 

 Moderate

200

8

4.0

 

0.51

(0.23; 1.13)

 High

202

8

4.0

 

0.51

(0.23; 1.12)

Supervisor support

      

 Low

137

4

2.9

0.227

1.00

 

 Moderate

206

15

7.3

 

2.49

(0.85; 7.35)

 High

339

21

6.2

 

2.12

(0.74; 6.06)

Co-worker support

      

 Low

156

7

4.5

0.510

1.00

 

 Moderate

75

3

4.0

 

0.89

(0.24; 3.36)

 High

446

29

6.5

 

1.45

(0.65; 3.24)

Job insecurity

      

 Low

245

10

4.1

0.378

1.00

 

 Moderate

249

16

6.4

 

1.57

(0.73; 3.40)

 High

188

13

6.9

 

1.69

(0.76; 3.77)

Job dissatisfaction

      

 Low

300

18

6.0

0.230

1.00

 

 Moderate

221

9

4.1

 

0.68

(0.31; 1.48)

 High

157

13

8.3

 

1.38

(0.69; 2.74)

Individual variables

      

Gender

      

 Men

282

21

7.4

0.153

1.00

 

 Women

431

21

4.9

 

0.65

(0.36; 1.18)

Language

      

 Dutch-speaking

502

28

5.6

0.584

1.00

 

 French-speaking

211

14

6.6

 

1.19

(0.64; 2.21)

Smoking

      

 Never smoked

430

21

4.9

0.296

1.00

 

 Ex-smoker

88

7

8.0

 

1.63

(0.71; 3.72)

 Current smoker

183

14

7.7

 

1.57

(0.81; 3.01)

Education

      

 Higher university

41

3

7.3

0.061

1.00

 

 Higher non-university

263

8

3.0

 

0.42

(0.11; 1.50)

 Higher secondary/professional

279

18

6.5

 

0.88

(0.27; 2.87)

 No diploma–primary school–lower secondary/professional

125

12

9.6

 

1.31

(0.39; 4.42)

Body mass index (kg/m2)

      

 Normal (≥20–<25)

378

18

4.8

0.021

1.00

 

 Underweight (<20)

109

3

2.8

 

0.58

(0.17; 1.93)

 Overweight (≥25–<30)

132

10

7.6

 

1.59

(0.75; 3.36)

 Obese (≥30)

50

7

14.0

 

2.94

(1.29; 6.67)

Perceived general health

      

 Very good

339

10

2.9

0.002

1.00

 

 Moderate to fair

373

31

8.3

 

2.82

(1.40; 5.65)

In the 12 months before inclusion

      

 Complaints of the neck

      

  No

520

33

6.3

0.396

1.00

 

  Yes

193

9

4.7

 

0.73

(0.36; 1.51)

 Complaints of the back

      

  No

369

19

5.1

0.384

1.00

 

  Yes

344

23

6.7

 

1.30

(0.72; 2.34)

 Complaints of the upper limbs

      

  No

581

28

4.8

0.011

1.00

 

  Yes

132

14

10.6

 

2.20

(1.19; 4.07)

 Complaints of the lower limbs

      

  No

535

32

6.0

0.858

1.00

 

  Yes

178

10

5.6

 

0.94

(0.47; 1.87)

Pain-related fear

      

 Low

240

12

5.0

0.538

1.00

 

 Moderate

237

13

5.5

 

1.10

(0.51; 2.35)

 High

233

17

7.3

 

1.46

(0.71; 2.99)

Catastrophising of pain

      

 Low

235

12

5.1

0.463

1.00

 

 Moderate

249

13

5.2

 

1.02

(0.48; 2.19)

 High

226

17

7.5

 

1.47

(0.72; 3.01)

Somatisation

      

 Low

239

17

7.1

0.616

1.00

 

 Moderate

240

13

5.4

 

0.76

(0.38; 1.53)

 High

233

12

5.2

 

0.72

(0.35; 1.48)

Negative affectivity

      

 Low

287

22

7.7

0.213

1.00

 

 Moderate

202

8

4.0

 

0.52

(0.23; 1.14)

 High

223

12

5.4

 

0.70

(0.36; 1.39)

RR relative risk, 95 % CI 95 % confidence interval

P value calculated with chi-square tests or Fisher exact tests,

Bold values indicate P < 0.05

Table 5

Continuous risk factors for the occurrence of sick leave due to low back pain after one year of follow-up (SLLBP at t1) in univariate analyses

Variable at baseline

SLLBPat t1

No SLLBP at t1

P value*

n

Median (Q1–Q3)

n

Median (Q1–Q3)

Current physical load

     

Professional

     

 Seniority in the current function (years)

41

3 (2–6)

657

3 (1–5)

0.188

Extra-professional

     

 Motor vehicle driving outside the work (km/year)

39

20,000 (12,000–30,000)

560

15,000 (10,000–25,000)

0.018

Individual variables

     

Age (years)

42

27 (25–29)

661

26 (24–29)

0.307

P value calculated with Mann–Whitney U tests

Bold values indicate P < 0.05

Multivariate analyses

The following variables, associated with a P value of ≤0.10 in univariate analyses, were considered for inclusion in multivariate analyses: the physical factors: (a1) motor vehicle driving outside the work (Mann–Whitney U test, P = 0.018), (a2) driving vehicles or machines at work (chi-square test, P = 0.024) and (a3) inability to change posture regularly (chi-square test, P = 0.076); the psychosocial work characteristic: (b1) possibilities to develop skills (chi-square test, P = 0.023); and the individual variables (c1) perceived general health (chi-square test, P = 0.002), (c2) complaints of the upper limbs in the year before inclusion (chi-square test, P = 0.011), (c3) body mass index (chi-square test, P = 0.021) and (c4) education (chi-square test, P = 0.061). Age and gender were included as epidemiological confounders, although they were not significantly related to back-related sick leave (Mann–Whitney U test, P = 0.307 and chi-square test, P = 0.153, respectively).

Multivariate analyses were based on backward selection. We constructed a model for each comprehensive combination of unrelated variables significant at P ≤ 0.10 in univariate analyses. In Table 6, we show the results of a model with unrelated variables. In case of interrelationships, the most significant variable was included in this model. This model showed a strong and significant relationship between a lack of ‘possibilities to develop skills’ at work and subsequent sick leave for LBP in the following year. In contrast with colleagues who reported high ‘possibilities to develop skills’ at baseline, workers with moderate and especially low ‘possibilities to develop skills’ showed, respectively, a five (RR 5.01, 95 % CI 1.10–22.88) and six (RR 6.08, 95 % CI 1.42–26.07) times higher risk on sick leave due to LBP 1 year later. Furthermore, the risk ratio for obese workers as opposed to workers with a normal BMI was 3.41 (95 % CI 1.37–8.48).
Table 6

Risk factors for the occurrence of sick leave due to low back pain after 1 year of follow-up (SLLBP at t1)

Variable at baseline

n

SLLBP at t1

Univariate analysesa

Multivariate analysesb

AF exposed

PAF total

n

%

P value

RR

95 %CI

P value

RR

95 %CI

AF

95 %CI

PAF

95 %CI

Psychosocial workload

             

Possibilities to develop skills

   

0.005

  

0.008

      

 High

154

2

1.3

 

1.00

  

1.00

     

 Moderate

151

10

6.6

 

5.10

(1.12–23.27)

 

5.01

(1.10–22.88)

0.80

(0.10–0.96)

0.22

(0.07–0.60)

 Low

234

20

8.5

 

6.58

(1.54–28.16)

 

6.08

(1.42–26.07)

0.84

(0.30–0.96)

0.31

(0.04–0.69)

Individual variables

             

Body mass index

   

0.019

  

0.030

      

 Normal

307

14

4.6

 

1.00

  

1.00

     

 Underweight

82

2

2.4

 

0.54

(0.12–2.35)

 

0.52

(0.12–2.27)

    

 Overweight

110

9

8.2

 

1.79

(0.77–4.15)

 

1.74

(0.75–4.04)

    

 Obese

40

7

17.5

 

3.84

(1.55–9.51)

 

3.41

(1.37–8.48)

0.71

(0.27–0.88)

0.30

(0.06–0.58)

RR relative risk, 95 % CI 95 % confidence interval, AFexposed attributable fraction among the exposed workers, PAFtotal total attributable fraction for the entire study population

Results from the model with no missing values [n = 539] for gender, age, body mass index, complaints of the upper limbs in the year before inclusion, perceived general health, inability to change posture regularly, motor vehicle driving outside the work, and possibilities to develop skills

aCox regression

bCox regression, backward selection, Pin = 0.10, Pout = 0.05

Bold values indicate P < 0.05

Attributable fractions

Table 6 presents the proportion of LBP-related sick leave that can be attributed to the risk factors identified among exposed workers [AFexposed] and among the entire study population [PAFtotal]. From these data, it follows that a lack of ‘possibilities to develop skills’ is responsible for a proportion of LBP-related sick leave comparable to that of ‘obesity’, which constitutes a factor that cannot be changed by workplace interventions.

Discussion

Principal findings of the study

This study aimed at investigating the development of sick leave due to LBP in a population of young workers with no or a modest history of back complaints. We have shown that only a small proportion of workers listed sick because of LBP and that the decision to report sick was influenced to a great extent by a poor psychosocial work environment. Once workers were at home, they failed to return to work for quite a long time.

More specifically, it was low ‘possibilities to develop skills’ that turned out to be the most important occupational risk factor for back-related sick leave. This finding may suggest that back-related sick leave is mainly taken by workers who lack professional challenges. The perceived lack of professional challenges may have led to a lower threshold to take sick leave.

Moreover, low ‘possibilities to develop skills’ was responsible for a proportion of LBP-related sick leave comparable to that of ‘obesity’, which constitute a factor that cannot be changed by workplace interventions. Thus, the influence of the environment is important and exists in job organisation and design to create a challenging professional environment.

Strengths and weaknesses

The main strength of the BelCoBack Study is its prospective design, respecting the temporal relationship between cause and consequence. Furthermore, the study took a different scope by focusing on young workers with no or a modest history of back pain. This choice allowed studying more accurately the development of back pain (Van Nieuwenhuyse et al. 2004, 2006) and sick leave. With respect to the multi-factorial nature of LBP, several categories of variables were included as follows: physical, psychosocial and individual.

Sick leave was registered by self-reported questionnaires because systematic company records on (cause-specific) absenteeism are not available to the occupational health physician in Belgium. The majority of studies have found acceptable validity for self-reports (Stapelfeldt et al. 2012). However, a Dutch study in blue collar workers reported low sensitivity of self-reports (55 %) to detect frequency of spells (Van Poppel et al. 2002). In our study, taking of sick leave (yes/no) and, if 'yes', the number of spells and the number of sickness absence days were questioned 1 year after the intake. The reason was practical: examinations in the BelCoBack Study were organised in relation with the annual medical examination by the occupational health physician to limit the extra load of the study for the companies. No diary methods in between were used. The recall period of 1 year is long and may have led to some underreporting. More specifically, for duration of sickness absence, there is evidence that shorter recall periods could increase the precision of self-reporting with an optimum recall period of no longer than 2–3 months (Severens et al. 2000).

Furthermore, both the potential predictors and the outcome were reported by the workers themselves. Therefore, we cannot exclude common-source bias, that is, bias due to the collection of information from only one source that may lead to correlated reports of predictors and effects and thus false-positive results (Dionne et al. 2002). It may be advisable in future studies to evaluate whether the self-reported lack of possibilities to develop skills, as important determinant, can be validated by assessments from supervisors or human resources personnel.

Due to the composition of the study population and the short follow-up, only 42 workers took sick leave because of LBP. As the number of sick days in these workers was substantial, it would have been interesting to analyse risk factors for the duration of sick leave. However, this was not possible due to lack of power.

Comparison with the literature

Only a limited number of studies have investigated physical and psychosocial characteristics as risk factors for the occurrence of sick leave due to LBP in a prospective way (Smedley et al. 1997; Hemingway et al. 1997; Wickstrom and Pentti 1998; Hoogendoorn et al. 2002; Tubach et al. 2002; Elders et al. 2003; van den Heuvel et al. 2004; Ijzelenberg and Burdorf 2005; Bergstrom et al. 2007; Alexopoulos et al. 2008).

Our results are in line with those of two studies that have also focused on workers with limited back antecedents. In a Swedish follow-up study, Bergstrom et al. (2007) showed only one occupational factor to be predictive for sick listing due to back or neck pain at the 18-month follow-up, that is, ‘few positive challenges at work’. One-third of the initial cohort reported no back or neck pain in the year previous to baseline. Workers who indicated that the work was not meaningful or challenging and that their skills and knowledge were not useful at work had twice as high sick listing as workers with the highest positive challenges (15 vs. 7 %). Similarly, in a cohort study among British nurses, Smedley and colleagues (Smedley et al. 1997) found psychosocial factors, that is, ‘low mood’, and not physical factors, to be predictive for sick leave due to LBP during follow-up. Lack of control, which reflects both skill discretion and decision authority, was also related to sick leave in a British follow-up study of office workers (Hemingway et al. 1997). However, it has to be noted that the physical workload in office workers is negligible and, therefore, quite different from that of our population. Other studies have shown the important role of a good psychosocial working environment also for sick leaves in general (Duijts et al. 2007), which may suggest that it is not specific for workers with LBP.

The majority of cohort studies have included workers regardless of LBP history. In populations with LBP complaints, both physical and psychosocial workloads play a primary role in back-related sick leave. On the whole, the influence of the physical work environment seems more important than that of the psychosocial work environment, and in extreme occupations, that is, populations with high disability rates at baseline and enormous physical loads (15,000 kg/day) such as scaffolders (Elders et al. 2003), the effect of physical load has been shown to dominate all other effects. The most likely explanation is that, in workers with a history of back pain, the high physical workload constitutes a hindrance to continue work.

In our population of young workers with no or limited antecedents of back pain, it was a poor psychosocial work environment that showed to be associated with future sick leave due to back pain and may thus have led to a lower threshold to take sick leave.

Conclusion to health professionals and policy makers

Physical work factors have been recognised for a long time as risk factors for LBP itself. Although literature about work-related risk factors for back-related sick leave is scarce, it is intuitively accepted that the same factors are also important for back-related sick leave. We reveal a more nuanced reality. In populations with LBP complaints, physical workload indeed plays a primary role in back-related sick leave. However, in workers with no or only limited LBP complaints at baseline, the physical workload does not appear to be the predominant reason for taking sick leave. Nevertheless, sick leave was substantial in this population and turned out to be associated mainly with a non-stimulating psychosocial work environment. Because the psychosocial work environment can be modified, this finding represents a potentially reversible cause of sick leave.

Acknowledgments

The BelCoBack Study was supported by the Belgian Federal Office for Scientific, Technical and Cultural Affairs (OSTC), projects PS/93/25, PS/12/26, PS/01/27.

The authors thank the other co-workers from the BelCoBack Study group for their contribution to data collection and data processing: D. Pirenne, E. Persijn, A. Leys, and L. Moors.

Conflict of interest

The authors declare that they have no conflict of interest.

Ethical standard

The study protocol was approved by the Commission for Medical Ethics appointed by the College of Physicians n°117 at the External Service for Prevention and Protection at Work IDEWE, Leuven, Belgium. A written informed consent was given by all included employees prior to their participation in the study.

Copyright information

© Springer-Verlag Berlin Heidelberg 2012