Article Text
PDF
Abstract
Objectives Those in lower occupational classes have an increased risk of sickness absence due to musculoskeletal diseases (MSDs), but studies examining the associations simultaneously across specified diagnostic groups within MSDs are lacking. We examined occupational class differences in the occurrence and length of long-term sickness absence due to different musculoskeletal diagnoses.
Methods A 70% random sample of employed Finns aged 25–64 years old at the end of 2013 was linked to data on sickness absence of over 10 working days obtained from The Social Insurance Institution of Finland and occupational class from Statistics Finland. Sickness absences due to MSDs initiated in 2014 were followed until the end of each episode for female (n=675 636) and male (n=604 715) upper non-manuals, lower non-manuals and manual workers. Negative binomial hurdle models were used to analyse the associations.
Results Within the studied MSDs, the most common causes of absence were back disorders, particularly back pain, and shoulder disorders. Osteoarthritis, disc disorders and rheumatoid arthritis induced the longest episodes of absence. Clear hierarchical class differences were found throughout, but the magnitude of the differences varied across the diagnostic causes. The largest class differences in the occurrence were detected in shoulder disorders and back pain. The class differences in length were greatest in rheumatoid arthritis, disc disorders and, among men, also in hip osteoarthritis.
Conclusions Hierarchical occupational class differences were found across different MSDs, with large differences in back and shoulder disorders. Occupational class and diagnosis should be considered when attempting to reduce sickness absence due to MSDs.
- sickness absence
- musculoskeletal
- public health
Statistics from Altmetric.com
Key messages
What is already known about this subject?
Employees in lower occupational classes are at excess risk of sickness absence due to musculoskeletal diseases compared with those in higher occupational classes.
However, socioeconomic differences in various sickness absence measures among employed people across a wide spectrum of specified musculoskeletal diagnoses are not known.
What are the new findings?
Hierarchical occupational class differences were found in the occurrence and length of absence across different musculoskeletal diagnoses, but the magnitude of the class differences was diagnosis-specific.
The class differences were particularly large in back disorders and shoulder disorders, covering the most common causes of absence within musculoskeletal diseases.
How might this impact on policy or clinical practice in the foreseeable future?
Occupational class and diagnosis should be considered in planning of preventive measures aiming at reducing sickness absence due to musculoskeletal diseases.
Furthermore, measures should be targeted particularly at employees in lower occupational classes having sickness absence due to back disorders and shoulder disorders in order to narrow the impact of occupational class differences on sickness absence effectively.
Introduction
Musculoskeletal diseases are a major cause of long-term sickness absence,1 comprising various disorders with differentials in aetiology, onset and duration.2 These differences are reflected in sickness absence attributable to specified musculoskeletal diagnoses: low back disorders without radiation, and neck and shoulder disorders are the most common causes of sickness absence within musculoskeletal diseases,1 3 but rheumatoid arthritis, osteoarthritis and back disorders with radiation usually cause the longest episodes of absence.1 3 4 In general, sickness absence due to musculoskeletal diseases increases with age,1 4 and both incidence and duration of the absences tend to be higher in women than in men.1 4–6
Previous studies have shown that those in lower occupational classes have an increased risk of sickness absence compared with those in higher classes (see, for instance, refs 7 8). The socioeconomic gradient in sickness absence is particularly steep in musculoskeletal diseases,9–12 and this has remained over time.12 However, studies examining occupational class differences in sickness absence simultaneously across different musculoskeletal diagnoses are lacking. The socioeconomic gradient in sickness absence may, however, be diagnosis-specific, as previously shown in mental disorders,13 and depend on the measure of absence.14
It has been demonstrated that those in lower occupational class have a higher risk of sickness absence due to different musculoskeletal diseases, such as back pain,15–18 shoulder and neck pain,16 17 upper limb disorders,16 19 and osteoarthritis.18 20 In the case of individuals with rheumatoid arthritis, an increased risk of long-term sickness absence has been found compared with a general population21; as with the general population, the risk is associated with high physical demands of work.21 The focus in previous studies regarding the association between occupational class and sickness absence due to musculoskeletal diseases has been restricted to either specific anatomical sites15–17 19 or at most to a couple of different diagnostic groups.18 Diagnosis-specific evidence on occupational class differences in sickness absence examined simultaneously across multiple diagnostic groups might help to identify high-risk groups with regard to work disability in different musculoskeletal diseases. This would also help to target preventive measures effectively in order to narrow the impact that occupational class differences have on sickness absence and reduce sickness absence. Previous studies have demonstrated that prevention potential is particularly high in almost all diagnostic groups within the main chapter of musculoskeletal diseases.3
The aim of this study was to examine occupational class differences in long-term sickness absence due to various musculoskeletal diseases among 1.2 million Finnish women and men aged 25–64 years old. Sickness absence was assessed using both the occurrence and length of absence in order to give a broader picture of the class differences in sickness absence due to musculoskeletal diseases.
Methods
Data
A nationally representative 70% random sample of Finnish residents aged 25–64 years old at the end of 2013 was obtained from the register of The Social Insurance Institution of Finland (Kela). Information on long-term sickness absence was based on sickness allowance, nationally administered by Kela. The data on sickness allowance episodes retrieved from Kela’s register included the first and last day of work disability and diagnosis based on the International Classification of Diseases (ICD-10) for each episode. All sickness allowance episodes due to musculoskeletal diseases (ICD-10 codes M00–M99) initiated between 1 January 2014 and 31 December 2014 were linked to the sample data by personal identity code assigned to each Finnish resident.
Based on the Health Insurance Act, Finns 16–67 years old who are not on pensions are entitled to sickness allowance as compensation for work disability due to a disease or injury up to approximately 1 year. The allowance is granted normally after a waiting period including the first day of work disability and the following nine working days, that is, calendar days excluding mid-week holidays and Sundays. The waiting period continues for 55 calendar days for those who have not been engaged in any gainful activity 3 months prior to work disability or if annual earned income falls below the minimum level.22 In this study, long-term sickness absence is measured as receipt of sickness allowance and refers to sickness absence lasting longer than 10 working days.
We used two outcome variables. The first was the occurrence of at least one new long-term sickness absence episode attributable to different musculoskeletal diagnoses initiated in 2014. Individuals were considered to have at least one new episode if the number of sickness absence days in a diagnostic group exceeded zero. The second outcome was the total number of sickness absence days due to different musculoskeletal diagnoses among those having at least one new long-term sickness absence episode attributable to the diseases. Each sickness absence episode due to musculoskeletal diseases initiated in 2014 was followed up from the first day until the last day of work disability; under the Finnish system, sickness allowance is paid up to 1 year, hence the follow-up ended at the end of year 2015.
As previously suggested,4 selection of the diagnostic groups was made to reflect differences between specific disorders within musculoskeletal diseases, ranging from acute to chronic conditions, and the importance of work disability. The diagnostic groups chosen in this study constitute major causes of sickness allowance within musculoskeletal diseases in Finland.22 The separate diagnostic causes within musculoskeletal diseases (ICD-10 codes M00–M99) included in this study were back disorders (M40–M54), shoulder disorders (M75), osteoarthritis (M15–M19), rheumatoid arthritis (M05–M06) and other musculoskeletal diseases (all other ICD-10 codes in the main chapter of musculoskeletal diseases). Back disorders were further examined in the subgroups of back pain (M54) and disc disorders (M50–M51) due to their large proportions of the number of initiated sickness allowance episodes within back disorders22 and previously found differences in length of absences caused by the conditions.4 Knee osteoarthritis (M17) and hip osteoarthritis (M16) were included as separate subgroups in order to detect potential disparities in the magnitude of the occupational class differences between the diseases. In addition, musculoskeletal diseases of any cause (any ICD-10 code within M00–M99) were assessed.
Data on occupational class were based on year-end information derived from the register of Statistics Finland, consisting of seven different categories.23 In this study, the focus was on employees classified into three occupational classes: upper non-manual employees, lower non-manual employees and manual workers, excluding entrepreneurs and farmers (approximately 8% of employed women and 16% of employed men in the data). The study population comprised 675 636 women and 604 715 men.
Statistical analyses
The sickness absence data showed overdispersion and an excess of zeros, that is, a large number of individuals with no sickness absence days during the study period. Therefore, we adopted a two-part method, a negative binomial hurdle model, previously used in sickness absence research that encountered count data with abovementioned characteristics (see, for instance, ref 24). In this model the underlying assumption is that, initially, all individuals are at risk for an adverse event, and once the event has occurred, that is, the hurdle is passed, the second part models the number of adverse events among persons having at least one event.25 Furthermore, this model allows to examine both the occurrence of sickness absence and the amount of absent days in association with occupational class.
The analyses were performed separately in two steps.26 First, a log-binomial regression was used to produce age-adjusted relative risks and their 95% CIs of having at least one new sickness absence episode during the study period associated with occupational class. Second, a zero-truncated negative binomial model was used to produce age-adjusted incidence rate ratios (IRRs) and their 95% CIs for the number of sickness absence days associated with occupational class among those having at least one new sickness absence episode during the study period. The IRRs can be interpreted as how many times more (IRR above 1.0) or less (IRR below 1.0) sickness absence days those in the occupational class in question are expected to have compared with upper non-manual employees.27 All analyses were stratified by gender and adjusted for age using age as a continuous independent variable. Statistical analyses were conducted using SAS V.9.4 statistical software.
Results
The majority of women were lower non-manual employees, whereas among men the largest occupational class was manual workers (table 1). In contrast, manual workers formed the smallest group among women and lower non-manual employees the smallest group among men.
Characteristics of the study population
Back disorders, particularly back pain, were the most common diagnostic causes of women’s new sickness absence episode within musculoskeletal diseases during the study period, followed by shoulder disorders, osteoarthritis and rheumatoid arthritis, respectively (table 2). The longest average lengths of absences were, in turn, caused by osteoarthritis, particularly hip osteoarthritis, disc disorders and rheumatoid arthritis among women. In contrast, back pain induced the shortest episodes within the studied musculoskeletal diseases. Among men, similar patterns in the proportions and average lengths of absences were found as among women but, throughout the different diagnostic causes, the proportions were lower and the average lengths longer than among women. In disc disorders, however, the proportions were equal among women and men.
Number (n) and proportions (%) of persons having at least one new sickness absence episode and average length of absence due to different musculoskeletal diagnoses by occupational class
In both genders, those in lower occupational classes had higher occurrence of sickness absence and longer average length of absence in all studied diagnostic causes within musculoskeletal diseases than those in higher occupational classes (table 2). The only exception was rheumatoid arthritis among men, for which the proportions were equally high among lower non-manuals and manual workers, although percentages were low across all occupational classes.
Age-adjusted associations between occupational class and long-term sickness absence due to different musculoskeletal diagnoses among women
The age-adjusted relative risks of having at least one new sickness absence episode according to occupational class among women are presented in table 3. With regard to any musculoskeletal disease, the relative risk of a new sickness absence episode was higher for manual workers and for lower non-manual employees compared with upper non-manual employees. Clear hierarchical occupational class differences in the relative risks were also found in all specified diagnostic causes within musculoskeletal diseases. The class differences were particularly large in shoulder disorders, back pain and knee osteoarthritis. The class differences were smallest in hip osteoarthritis and disc disorders.
Relative risks (RR) of a new sickness absence episode and incidence rate ratios (IRR) for the number of sickness absence days by occupational class
The age-adjusted IRRs for sickness absence days according to occupational class among women having at least one new sickness absence episode due to different musculoskeletal diagnoses are shown in table 3. Compared with upper non-manual employees, manual workers were expected to have 29% more sickness absence days due to any musculoskeletal disease, while the corresponding percentage was 18 among lower non-manuals. For all diagnostic causes, clear hierarchical occupational class differences in sickness absence days were found. The class differences were large in rheumatoid arthritis, disc disorders and knee osteoarthritis. The class differences were smallest in back pain, shoulder disorders and hip osteoarthritis.
Age-adjusted associations between occupational class and long-term sickness absence due to different musculoskeletal diagnoses among men
Men also had clear relative age-adjusted occupational class differences in having at least one new sickness absence episode during the study period in all studied diagnostic causes (table 3). The relative risk of a new absence episode due to any musculoskeletal disease was higher for manual workers and for lower non-manual employees compared with upper non-manual employees. As among women, also among men the class differences were particularly large in sickness absence due to shoulder disorders and back pain. In contrast to the estimates for women, for male manual workers the risk of a new absence episode due to rheumatoid arthritis was slightly higher than the risk of a new episode due to knee osteoarthritis compared with upper non-manual employees. As with women, the class differences were smallest in hip osteoarthritis and disc disorders.
Also among men clear hierarchical class differences were found in all studied diagnostic causes of sickness absence days within musculoskeletal diseases (table 3). Manual workers were expected to have 36% and lower non-manuals 18% more sickness absence days due to any musculoskeletal disease than upper non-manuals. The class differences were largest in rheumatoid arthritis, in which manual workers were expected to have 129% more sickness absence days than upper non-manuals. The class differences were also large in hip osteoarthritis and disc disorders. Compared with upper non-manuals, manual workers were expected to have 79% and 77% more sickness absence days due to hip osteoarthritis and disc disorders, respectively. As with women, the smallest class differences in sickness absence days were found in back pain.
Discussion
This register-based study on 1.2 million Finnish employees examined occupational class differences in long-term sickness absence due to different musculoskeletal diagnoses among women and men. The study examined both the occurrence (ie, having at least one new sickness absence episode during the study period) and length (ie, the number of sickness absence days among those having at least one new sickness absence episode during the study period) of absence. In line with previous studies,1 3 back disorders, particularly back pain, and shoulder disorders were the most common causes of sickness absence among musculoskeletal diagnoses. In contrast, osteoarthritis, especially hip osteoarthritis, disc disorders and rheumatoid arthritis gave rise to the longest average lengths of absences in both genders, thus echoing earlier studies.1 3 4
The main findings of the study are summarised as follows. (1) Hierarchical occupational class differences were found in long-term sickness absence due to any musculoskeletal disease and due to different specified musculoskeletal diagnoses across the occupational classes. Lower occupational class was consistently associated with higher relative risk and IRR for occurrence and length of absence due to musculoskeletal diseases, respectively. (2) The largest class differences in the occurrence of absence were found in shoulder disorders and back pain in both genders. (3) The class differences in the length of absence were, in turn, largest in rheumatoid arthritis, disc disorders and, among men, also in hip osteoarthritis.
Our results corroborate previous studies showing hierarchical occupational class differences in sickness absence due to musculoskeletal diseases.9–12 Generally, the class differences in sickness absence have been explained mainly by adverse physical and psychosocial working conditions and unhealthy behaviours, and to some extent by health.7 8 10 Work-related factors, such as postural constraints, occupational hazards and low decision latitude, have been shown to explain a substantial part of occupational differences in sickness absence due to musculoskeletal diseases in both genders.10 In a French study,10 work-related factors explained nearly half of sickness absence attributable to musculoskeletal diseases among men, and almost one-third among women. Manual workers could also be exposed simultaneously to several adverse work-related factors more often than employees in higher occupational classes, which may increase the influence of work factors on the class differences in sickness absence.10 Unhealthy behaviours could also play a role; smoking, obesity and lack of exercise comprise major risk factors for several musculoskeletal diseases.2 In addition, musculoskeletal morbidity has been shown to follow a consistent hierarchical occupational class gradient in an employed population.28 Our results of both the occurrence and length of absence echo this phenomenon. Furthermore, health-related selection into occupational classes may also explain partly the observed class differences in sickness absence due to musculoskeletal diseases.10
Despite the occupational class differences being visible in sickness absence due to different musculoskeletal diseases, the magnitude of the relative class differences varied between different diagnoses and between the measures of absence. The class differences in the occurrence of absence were more pronounced in the common causes of absence under study, that is, back pain and shoulder disorders, than in the chronic, lifelong diseases. Previous investigations have shown that those in lower occupational classes have a higher risk of sickness absence due to back pain15–17 and shoulder disorders16 compared with those in higher classes. Work factors account for a major part of the class differences in sickness absence due to musculoskeletal diseases,10 and work-relatedness has been previously linked particularly to low back pain and upper extremity disorders.29 Heavy physical work, including lifting heavy weights and working in uncomfortable positions for example, constitutes a major risk factor for long-term sickness absence among employees with low back pain and shoulder pain.30 In fact, hierarchical occupational class disparities in the risk of rotator cuff syndrome, a common cause of absence due to shoulder disorders,19 are mostly explained by heavy physical workload, in particular repetitive movements among both female and male manual workers.31 Likewise, repetitive tasks at work have been shown to be associated with an increased risk of sickness absence due to back pain among manual workers.17 Of psychosocial work factors, low work satisfaction, in particular, has been shown to associate with the risk of sickness absence due to back disorders.32 The class differences in the occurrence of absence can also occur as a result of the occupational class differences in the ability to perform particular work tasks while having occupational constraints as a consequence of musculoskeletal diseases.
In contrast to results concerning the occurrence of absence, this study showed that the occupational class differences in the length of absence were large in rheumatoid arthritis, cervical and other intervertebral disc disorders and, among men, hip osteoarthritis. Previous studies have shown that these diseases constitute major determinants of impaired working capacity33 and prolonged sickness absence within musculoskeletal diseases.4 11 A Dutch study found that the impact of physically strenuous work on duration of sickness absence tends to increase in the course of time, suggesting a pronounced effect of high physical workload on duration of absence during the chronic phase of a disease.34 Furthermore, accumulation of adverse work-related factors may reinforce the effect.10 These findings may partly explain the observed large class differences in the length of absence due to chronic musculoskeletal diseases in this study. Furthermore, employees in higher occupational classes may have more opportunities to adapt their work tasks to better meet their current work ability,35 which could shorten sickness absence and promote return to work despite chronic diseases in higher occupational classes compared with manual workers. The observed large class differences in the length of absence in chronic musculoskeletal diseases may also reflect more serious health conditions in lower occupational classes compared with those in higher classes. For instance, intense pain has been shown to constitute an important risk factor for prolonged sickness absence due to musculoskeletal diseases.34 Lower occupational class may also hinder, for instance via poorer income, access to treatment.36
Our study showed that the class differences between lower non-manuals and manual workers tended to be smaller both in the length and the occurrence of absence due to chronic musculoskeletal diseases compared with the less disabling musculoskeletal diagnoses. In addition to manual workers, there are several physically demanding occupations classified as lower non-manuals with inflexible job tasks and low level of income (eg, nurses, practical nurses and childminders). Severe and usually permanent forms of disabilities combined with these factors may explain why rates of incidence and length of absence in lower non-manuals were generally similar to those in manual workers for chronic musculoskeletal diseases.
The prevalence of most of the musculoskeletal diseases tends to increase with age.2 We found that hierarchical occupational class differences in long-term sickness absence were uniform across the spectrum of musculoskeletal diagnoses also in different age groups (ie, 25–39 years, 40–54 years and 55–64 years; data not shown). The class differences in sickness absence due to musculoskeletal diseases were thus evident already in the early years of working careers, despite musculoskeletal diseases being generally more common among older employees than younger employees. The results were consistent with both men and women. In Finland, half of the employees are women, for whom the largest occupational class comprise lower non-manuals. This occupational class includes many typical female occupations, for instance in healthcare sector, with high physical and psychosocial work demands.
The present study has several strengths. All data were retrieved from comprehensive and reliable national register databases. A very large nationally representative sample of working-age Finns was linked through personal identity code to register data on occupational class and over 10-day-long sickness absence episodes attributable to musculoskeletal diseases initiated in 2014. Data on sickness absence were based on paid sickness allowances administered and registered by Kela, with practically no missing information. All sickness absence episodes were medically certified and assigned a diagnosis based on the classification of ICD-10; hence, self-report bias could be avoided. Sickness absence episodes for all employees in three occupational classes, in other words upper non-manuals, lower non-manuals and manual workers, were followed from the beginning to the end of each episode. We examined both the occurrence and length of sickness absence in order to give a broader picture of the class differences in sickness absence. In addition, through access to ICD-10 codes, several different diagnostic causes within musculoskeletal diseases could be examined in the present study. Concerning the occupational classes incorporated into the study, the results can be generalised to the Finnish labour force and with caution to other countries.
There are some limitations in the present study. Due to lack of data in national registers, we were unable to examine potential explanatory factors to the occupational class differences in sickness absence due to different musculoskeletal diagnoses. Sickness absence can be regarded as a step in a complex causal work disability process influenced by various, often overlapping factors at different phases.37 These factors include, for instance, health, health behaviours, working conditions, healthcare and social systems, and legislation.37 The observed class differences could have been attenuated, at least to some extent, if these factors were examined in models. Furthermore, this study focused on sickness absence episodes exceeding 10 working days since there are no nationwide register data on shorter sickness absence episodes in Finland. Moreover, musculoskeletal diagnoses within the ICD-10 chapter M00–M99 could be classified into the subgroups at the three-digit level due to registration practice during reimbursement of sickness allowance. We were unable to examine separately sickness absence episodes of other common musculoskeletal diseases in the working population, such as epicondylitis, in the present study. Some limitations arise also as a consequence of the healthcare system. For instance, discriminating between back pain and disc disorders is not always unambiguous for doctors when making a diagnosis. Furthermore, physicians’ sick leave prescribing practices may vary, and in a Finnish study, clinical specialists tended to prescribe shorter sickness absences than general practitioners.38
In the early 2010s, various amendments were made to Finnish legislation in order to prevent work disability and promote return to work despite constraints in work ability in cooperation with employees, employers and occupational health services.39 In the future, preventive measures should be continued and targeted particularly to sickness absence episodes due to back disorders, that is, back pain and disc disorders, and shoulder disorders in lower occupational classes in the attempts to reduce sickness absence and narrow the class differences effectively in the employed population. In Finland, the aforementioned diagnostic causes account for approximately half of total sickness allowance benefits paid due to musculoskeletal diseases.22 The actions should be focused particularly on manual workers’ heavy physical work tasks, such as lifting heavy weights, working in uncomfortable positions and performing repetitive movements in the attempts to prevent absence from work. Furthermore, feasible adaptations to working conditions among manual workers are recommended in order to shorten sickness absence episodes and prevent prolonged absence from work. Our findings emphasise also that preventive measures should be targeted to both younger and older employees working in lower occupational classes and also to lower non-manual employees with high work demands. In addition to work-related intervention, other preventive and curative measures, such as promoting exercise, preventing obesity and adequate medical treatment, should be paid attention.3
Conclusions
Hierarchical occupational class differences in long-term sickness absence were found across a wide spectrum of musculoskeletal diagnoses in the working population. Employees in lower occupational classes were at a higher risk of sickness absence and were expected to have more sickness absence days due to musculoskeletal diseases than those in higher occupational classes. The magnitude of the class differences varied, however, between the diagnostic causes, and particularly large differences occurred in the common causes of absence, that is, back disorders and shoulder disorders. Both occupational class and diagnosis should be taken into account in the attempts to prevent sickness absence and narrow the class differences in long-term sickness absence attributable to musculoskeletal diseases effectively in the future.
References
Footnotes
Contributors JP had the original idea for the present study. All the authors participated in planning the study. JP and OP conducted the statistical analyses. JP, OR, OP, EL and JB interpreted the results. JP wrote the first draft of the manuscript and all the later versions. OR, OP, EL and JB reviewed and revised the manuscript. All the authors approved the final manuscript for submission to the journal.
Funding This study was supported by the Doctoral Programme in Clinical Research at the University of Helsinki, The Social Insurance Institution of Finland (grant 67/26/2014), the University of Helsinki, the Strategic Research Council of the Academy of Finland (grant 293103/2015) and the Academy of Finland (grant 294514/2016).
Competing interests None declared.
Ethics approval This study used only secondary data obtained from registers. Permission to use the data in the present study has been obtained from The Social Insurance Institution of Finland (Kela) (permission number 59/522/2015). Kela and its Ethics Committee follow Finnish legislation and are committed to the guidelines provided by the Finnish Advisory Board of Research Integrity. According to Finnish law, ethics approval was not required since the present study is based solely on registers.40 Conventions of good scientific practice, data protection and information security were applied in analysing the data and in presenting the results.
Provenance and peer review Not commissioned; externally peer reviewed.
Presented at This study was presented at the 10th European Public Health Conference Sustaining resilient and healthy communities, and the abstract has been published (Pekkala J, Rahkonen O, Pietiläinen O, et al, Sickness absence due to musculoskeletal diagnoses by occupational class in Finland: a register study. Eur J Public Health 2017;27(suppl_3), ckx187.197.).