Reasons for presenteeism in different occupational branches in Sweden: a population based cross-sectional study

As pointed out by several scholars, sickness presenteeism, that is, going to work despite illness, is related to the employees working conditions, to their health and to a number of other individual and organizational factors (Grinyer and Singleton 2000; Johns 2010; Karanika-Murray et al. 2015). Presenteeism has been shown to cause large costs for companies and for the society (Kinman 2019; Strömberg et al. 2017; Vänni et al. 2017), but has also been shown to increase the risk of future ill health (Aboagye et al. 2019; Gustafsson et al. 2019; Gustafsson and Marklund 2011, 2014). One large longitudinal study found for example that obesity and long-term health conditions were strongly associated with presenteeism (Keramat et al. 2020). Thus, it is essential to better understand the different reasons employees give for going to work despite illness.

A number of studies have found significant differences in the prevalence of presenteeism in occupational groups and or occupational sectors. Higher prevalence of presenteeism has been found among employees in the welfare, health and educational sectors compared to other sectors (Aronsson et al. 2000), among managers when compared employees without a management role (Hansen and Andersen 2008), among public sector employees when compared to private employed (Bockerman and Laukkanen 2010; Johansen et al. 2014), among blue collar employees when compared to white-collar employees (Gustafsson and Marklund 2011), and among self-employed when compared to employees (Nordenmark et al. 2019). One study showed that manual workers reported higher rates of presenteeism than non-manual workers, and that temporary workers reported higher presenteeism than permanent workers (Navarro et al. 2018). A review of presenteeism in different occupations in the health care sector showed that physicians reported higher rates of presenteeism than nurses and nursing assistants (Webster et al. 2019). Similarly, a study of employed in the educational sector found that teachers had a lower prevalence of presenteeism than other school employed (Dudenhöffer et al. 2017).

Fewer studies have been concerned with occupational differences in the reasons lying behind sickness presenteeism, and large share of them concerns occupations in the health sector. However, three studies have compared different occupational groups (Johansen et al. 2014; McKevitt et al. 1997; Navarro et al. 2018). In a study of mixed occupational groups Johansen et al. (2014) found large deviations in reasons between managers and non-managers, where the managers rarely gave economic reasons but often said that nobody else could carry out their obligation. In contrast, non-managers often reported that they could not afford taking sick leave or that they assumed negative effects of presenteeism on colleagues (Johansen et al. 2014). This finding finds support in a study of salaried workers that found that the most common reason for presenteeism was that they ‘did not want to be a burden to colleagues’ (Navarro et al. 2018). In another study comparing reasons for presenteeism among physicians and accountants, a majority of the physicians reported that they did not want to be unfair to colleagues, while accountants to a high degree reported that ‘no-one else could do the work’ (McKevitt et al. 1997). The concern among physicians that presenteeism would burden the colleagues was also mentioned in several other studies (Al Nuhait et al. 2017; Bracewell et al. 2010; Gustafsson Senden et al. 2016; Jena et al. 2012; Webster et al. 2019).

However, most of the conducted studies have only compared a few occupations and most of the studies on reasons for presenteeism concern hierarchical differences between occupations. There is a shortage of studies on horizontal differences in working conditions between occupational branches and how these may affect reasons for presenteeism, focusing on variations in the work object such as personal services versus industrial production.

To study work and social stratification, Kohn and Schooler introduced in 1983 a horizontal classification of occupations in three categories “working with things”, “working with data” and “working with humans” (Kohn and Schooler 1983). Similar horizontal classification systems have later been developed, although based on different theoretical grounds and different empirical sources. One study primarily concerned labor market inequalities (Blackburn et al. 2001; Lippa et al. 2014), while other have focused on work psychology and occupational perceptions to understand general occupational differences and gender differences in working conditions (Cerdas et al. 2019; Leijon et al. 2002; Lips-Wiersma et al. 2016).

The present study has adapted a classification similar to the one by Kohn and Schooler, based on differences and similarities regarding the object employees are working with (Kohn and Schooler 1983). The three categories have in the present study been named: “handling people”, “handling things” and “handling symbols” and were based on the Swedish version of the Statistical classification of economic activities in the European Community 2008 (EU 2008). The selection of occupations in each of the three categories will be presented in detail in the methods section.

Although there is no comprehensive theoretical model for presenteeism, there are a few rather similar suggested frameworks based on empirical experiences. Johns’ dynamic model (2010) includes contextual factors, health and person factors, where the contextual factors include work environment and work organization and person factors include attitudes to work as well as personality (Johns 2010). Lohaus and Habermann (2019) agree with Johns but divide the contextual factors into environmental, work-related and organizational factors. Keramat et al. (2020) on the other hand, exclude the contextual factors and focus on job factors, individual factors and health. As the present study is studying differences between occupational categories, the main theoretical assumption is that employees’ obligations at work are essential in their reported reasons for presenteeism. However, a number of other factors that previous studies have found important for the choice of going to work despite illness have also been used in the present study (Gustafsson et al. 2020; Johns 2010; Keramat et al. 2020; Lohaus and Habermann 2019). These are, age, sex, level of education, sector of employment, disposable income, physical and psychosocial work environment exposures and health complaints.

The present study had three aims. The first aim was to examine the prevalence of presenteeism, that is, going to work despite illness, among three occupational categories with varying work objects: handling people, handling things, and handling symbols. The second aim was to describe self-reported reasons for presenteeism in the three occupational categories. The third aim was to examine to what degree the differences in reasons for presenteeism between the three occupational categories remained after control for potential confounders.

In Table 2, the distribution of socio-demographic factors, work environment factors and health symptoms for the three occupational categories are presented. There were notable differences between the three occupational categories concerning age, gender distribution, level of education and employment sector. The category of employees handling things was dominated by men, had fewer employees with university education, and was dominated by employees working in the private sector. Employees handling symbols had a more even age and sex distribution, a relatively large share of university educated and was evenly distributed between privately and publicly employed. Employees handling people were predominantly female, often aged over 50, well educated, had the lowest share with high income and a large majority was employed by county and local council or public organizations (Table 2).

Table 2 additionally shows some differences between the occupational categories regarding work environment indicators. As expected, employees handling things more often than the other two categories reported heavy physical work and strenuous postures, but less often high psychosocial job demands. In contrast, employees handling symbols reported lower physical demands but higher psychosocial demand and higher job control than the other two categories. Employees handling people reported relatively high physical strain and high psychosocial demands, but also the highest share of low job control among the three categories (Table 2). The proportion reporting back symptoms, being tired and having sleeping troubles was somewhat higher among employees handling people when compared to the other two occupational categories.

The prevalence of presenteeism differed between the three occupational categories. Those employed in occupations handling things reported a prevalence of 35% of never having practiced presenteeism in the last year, employees handling symbols reported 30% and the figure for handling people was 26% (Table 3). The prevalence for having practiced presenteeism 2–3 times and four times or more was highest among employees handling people (Table 3). The most common reason for presenteeism given by employees handling things and employees handling symbols was ‘nobody else could take my responsibilities’ (38% and 50%, respectively), while the most common reason given by employees handling people was ‘I do not want to burden my colleagues’ (54%) (Table 3).

To understand the role of individual factors when other relevant factors were adjusted for, multiple regressions were carried out to estimate the odds for reporting a specific reason for presenteeism. The results of conducted regression analyses showed that the socio-demographic factors remained important in relationship to differences in reasons for presenteeism even after control for occupational sector (see supplementary file). Thus, older employees less often than younger reported ‘I do not want to be considered lazy’ (OR 0.18–0.44) and younger less often reported ‘I do not want to burden my colleagues’ (OR 0.61–0.77). Women more often than men reported ‘I do not want to burden my colleagues’ (OR 1.27), while men more often said ‘Nobody else can carry out my responsibilities’, but this difference was not statistically significant (see supplementary file). As expected those with lower income more often than those with higher income gave the reason ‘I can’t afford to take sick leave’ (OR 2.40–3.25), and those with university degree more often than those with lower education claimed that ‘Nobody else can carry out my responsibilities’ (OR 0.62–0.70) (see supplementary file). The answer ‘I do not want to burden my colleagues’ was more common among employees whose work was purely physical (OR 1.19) and among those with low job control (OR 1.59), while the answer ‘Nobody else can carry out my responsibilities’ was more common among those who reported high control (OR 0.66 and high demands (OR 2.49) (see supplementary file).

However, the large and systematic differences in reasons between the three occupational categories remained even after control for all potential confounders, including sex, age, income, health symptoms and the physical work environment factors (Table 4). In the response reflecting solidarity with colleagues, ‘I do not want to burden my colleagues’, the rate for employees handling people was elevated when compared to employees handling symbols (OR 1.68) (Table 4). A contrast was that reporting ‘I am worried about being laid off’ was less common among employees handling people than among employees handling symbols (OR 0.51). Among employees handling things the two reasons ‘Nobody else can carry out my responsibilities’ (OR 0.72) and ‘I can’t afford taking sick leave’ (OR 0.71) were less common when compared to employees handling symbols (Table 4). It should be noted, however, that there were no significant differences in odds ratios between the occupational categories in reporting “I enjoy my work”, “My pride depends on not taking sick leave”, “Going to work was beneficial for my health” and, “I want to maintain my social network” (Table 4).

The present study showed differences between the three occupational categories both with respect to prevalence of presenteeism and with respect to the different reasons the employees reported. There were also variations in reasons related to socio-economic background factors, work environment factors and health, but the differences between occupational categories remained mostly significant even after controlling for these factors. The largest contrast was found between employees handling people and employees handling things. The most common reason among employees handling people was ‘I do not want to burden my colleagues’ indicating solidarity with workmates but also that work in this sector is often team-oriented. Among employees handling symbols the most common answer was ‘Nobody else can carry out my responsibilities’, which may indicate a high degree of specialization in the occupations involved as well as individualized organization of work tasks and obligations.

The shown differences in given reasons for presenteeism between women and men, as well as between young and old are in agreement with other studies (Al Nuhait et al. 2017; Dudenhöffer et al. 2017; Gustafsson Senden et al. 2016; Hansen and Andersen 2008; Jena et al. 2012; Johansen et al. 2014; Rosvold and Bjertness 2001). One recent research review on presenteeism related to infectious illness, however, concluded that results on gender and age were inconsistent in cases of infectious disease (Webster et al. 2019). The found differences in the present study that presenteeism was related to employment sector and educational level have also been reported in one other study (Johansen et al. 2014). Other studies have found that the reasons for presenteeism also varies in relation to health differences (Bockerman and Laukkanen 2010), and income (Navarro et al. 2018; Skerjanc and Dodic Fikfak 2020; Webster et al. 2019). High psychosocial job demands and low supervisor support among different groups of employees could explain the differences in prevalence and reasons for presenteeism in the present study as well as in previous studies (Brborovic et al. 2017; Dudenhöffer et al. 2017; Hansen and Andersen 2008; Johns 2010; Nordenmark et al. 2019; Webster et al. 2019). No previous study has been found that could verify our results on the role of low control at work in explaining reasons for presenteeism.

The fact that differences between occupational categories in the prevalence and reasons for presenteeism remained even after confounder control is in line with research that has shown differences related to occupational hierarchies as well as to horizontal divides (Al Nuhait et al. 2017; Bracewell et al. 2010; Dudenhöffer et al. 2017; Johansen et al. 2014; McKevitt et al. 1997; Nordenmark et al. 2019). Interestingly, the gender differences remained in our study even after control for work sector. Independent of all sector belonging women more often than men motivated their presenteeism by “not wanting to burden colleagues”, while men more often referred to irreplaceability. As the occupations ‘handling people’ are women dominated and the occupations ‘handling things’ are male dominated the combination of gender and occupational sector makes the occupational difference even stronger.

The idea about presenteeism as being at least partly health promoting for the individual (Karanika-Murray and Biron 2020), does not get support in the present study. Only few respondents in all three occupational categories reported that going to work would be beneficial for their health. A larger share, particularly in the occupational category ‘handling people’, reported that they could not afford taking sick leave, and this may indicate that presenteeism is involuntary in these cases as Holland and Collins has shown in a study of employees with rheumatoid arthritis (Holland and Collins 2018). However, our study relies on self-reported reasons and presenteeism may well be beneficial although the individual does not experience presenteeism in this way.

In general, the interpretations of presenteeism in previous studies are linked to differences in professional cultures, although few have empirically studied differences in such cultures related to employees’ self-reported reasons. As Johns (Johns 2010) has pointed out, the social dynamics or collective motives of occupational categories that are part of professional cultures should include ‘the reaction of colleagues and clients to the act of presenteeism, both as encouragers and discouragers’. The idea behind the present study was that occupations classified according to their work objects would take this aspect into consideration. The fact that in occupations handling people, where teamwork and vulnerable clients are common, the main reason for presenteeism was solidarity with colleagues and clients illustrates the role of the character of the work object in these occupations. Also, the responses in occupations handling symbols may be seen to reflect the social dynamic as individualistic and specialized, and the reason for presenteeism often given is that nobody else can carry out the individual’ responsibilities point at similar motives. However, more studies are warranted on the role of shortage of resources, degree of specialization and management strategies for absence and presence.

Differences between occupational categories with respect to the main work object seem to be important for the prevalence of presenteeism and the underlying reasons. As presenteeism has been shown to affect the future health of employees and the productivity of the work unit, attempts to reduce presenteeism may be important. In the case of infectious disease, presenteeism may also jeopardize the health of clients, patients or colleagues. Organizational measures including policies concerning job insecurity and sickness absence as well as improved work environments are warranted. As the reasons vary between occupations, specific measures should be applied in different occupational settings. Thus, among employees handling people it seems to be relevant to improve the system for covering up absence in often team-oriented work. In specialized functions among employees handling symbols and handling things the corresponding focus could be the introduction of shared responsibilities for specific tasks and obligations.