Frequent short sickness absence, occupational health service utilisation and long-term sickness absence due to mental disorders among young employees

Mental disorders are increasingly contributing to sickness absence (SA) from work in the economically developed countries (Lidwall et al. 2018; Nicholson 2018; Mauramo et al. 2018; Stansfeld et al. 2011; Bultmann et al. 2005). In Finland, mental disorders are the leading cause of medically-certified long-term sickness absence (LTSA) and disability retirement alongside with musculoskeletal disorders (Kela 2017). Overall LTSA causes large financial burden to employees, employers and societies due to loss of productivity, health care costs, and financial compensation for absence (Nicholson 2018, van Hoffen et al. 2019). At the individual level mental disorders affect the quality of life (Saarni et al. 2007). In addition, LTSA may cause social stigma and isolate the individual from the normal day-to-day life (Nicholson 2018; Baumann 2007). This effect might be more substantial among younger employees (Knapstad et al. 2014). Therefore, the identification and treatment of mental disorders at early stages is very important.

The association between prior SA and future SA has been well-established in previous studies. The length and frequency of prior SA are associated with the risk for subsequent SA (Notenbomer et al 2019; Koopmans et al 2008; Laaksonen et al. 2013). Even short-term SA spells predict future longer SA spells (Laaksonen et al. 2013), also among younger employees and SA due to mental disorders (Sumanen et al. 2017). In addition, previous SA has been found to be a risk factor in studies aiming to develop effective screening tools for all-cause (Airaksinen et al. 2018; Schouten et al. 2016) and mental disorder-related LTSA (van Hoffen et al. 2019). Thus, SA records might be used to identify workers vulnerable to future LTSA. It is also known that frequent attendance (FA) at primary healthcare units is associated with mental disorders (Gili et al. 2011), and that especially persistent FA may indicate underlying chronic psychiatric conditions (Smits et al. 2014). However, it is yet to be determined, whether short-term SA spells and frequent primary care utilisation identify the same or different individuals at risk of LTSA due to mental disorders. From the perspective of targeting preventive measures, the question is whether the identification of at-risk individuals could be increased using linked data obtainable from different electronic health records, such as SA records and medical records.

In Finland, occupational health service (OHS) is a major provider of primary care for working-age population. OHS typically provides primary care services alongside with preventive measures. OHS primary care services are accessible to 90% of the whole of working population, three out of four workers name OHS as their main and preferred type of primary care provider, and the patients of OHS physicians only seldom visit public health centre GPs (Ikonen et al. 2013; Virtanen and Mattila 2011). Hence OHS is the most suitable location to investigate the use of primary care services of the working population in Finland. Recent studies from Finland have found that frequent occupational health service (OHS) utilisation may be associated with higher levels of subsequent SA (Reho et al. 2018; Reho et al. 2020; Harkko et al. 2020). Our study extends the previous research on the associations between frequent short-term sickness absence (FSTSA) spells for any cause and LTSA due to mental disorders by supplementing these analyses with data on primary care utilisation (i.e. OHS utilisation). To our best knowledge, no prior studies have assessed whether and to which extent these indicators are complementary from the perspective of identification of persons at risk of LTSA due to mental disorders.

Several studies from different European countries show that LTSA due to mental disorders is more common among women than men (Lidwall et al. 2018; van Hoffen et al. 2019; Halonen et al. 2018). A Swedish register-based study (Lidwall et al. 2018) with the entire workforce aged 16–64 (n = 6 192,397) showed that the risk of SA due to mental disorders exceeding 14 days is highest among those aged 30–39. A Norwegian study with survey and register data (Foss et al. 2010) among 30–45-year-olds (n = 8333) with outcome of > 8 weeks SA due to mental disorders showed that the age-related differences exist only among men. Other demographic risk factors were lower socioeconomic position (Foss et al. 2010), and working in health care, education and social services (Lidwall et al. 2018). Moreover, it has been shown that, in Finland, a high use of OHS mostly adheres to these sociodemographic and socioeconomic patterns (Harkko et al. 2020). For these reasons, in this study, we include gender and occupational class in the analyses.

The aim of the present study was to assess whether data on primary care utilisation may enhance the identification of the risk of LTSA due to mental disorders, compared with examining the relationship between FSTSA and subsequent occurrence of LTSA due to mental disorders only. To do so, we examined (1) whether FSTSA associates with subsequent LTSA due to mental disorders, (2) whether OHS visits before and after FSTSA associate with LTSA due to mental disorders, and (3) whether gender, occupational class and disease burden at baseline period affect the examined associations.

In this study, we found that frequent short-term sickness absence (FSTSA) due to all causes was associated with subsequent long-term sickness absence (LTSA) (over 11 calendar days) due to mental disorders. In addition, we found that increased primary care use in occupational health services (OHS), measured as the number of visits at one-year baseline period, was associated with LTSA due to mental disorders independently of FSTSA. These associations remained statistically significant, after controlling for the studied covariates, i.e. occupational class and disease burden at baseline period. The results point to the following interpretations. First, both FSTSA and increased OHS utilisation may be interpreted as indicators of unobserved processes of health degradation leading to LTSA due to mental disorders, and that the use of primary care utilisation data may be used to identify a larger proportion of individuals at risk of LTSA due to mental disorders compared with monitoring FSTSA data solely. Second, we found that occupational class had modest-to-none influence to the estimates, i.e. despite differences in the overall distribution of FSTSA and LTSA due to mental disorders, as seen in the descriptive results, there were no substantive differences in how FSTSA or OHS utilisation associated with LTSA due to mental disorders across occupational classes. While the processes leading to LTSA due to mental disorder may be dependent on different sociodemographic and occupational factors, such as gender or the differences of job-related characteristics of different occupations, FSTSA and OHS use provide indication of these processes consistently across these population groups. Third, the changes in estimates after controlling for the disease burden at baseline period suggest that FSTSA and OHS identify both individuals with and without accumulated sickness absence days. Fourth, in the stratified analysis of OHS use after short-term SA among those with and without FSTSA, we did not find association between the OHS use and LTSA due to mental disorders among those with FSTSA, whereas the association was found among those with no FSTSA. The result may be explained by the partial correlation between FSTSA and OHS use but, also, we find it likely that the use of OHS for obtaining a medical certificate for SA lasting more than 3 days is more prevalent among those with FSTSA. We acknowledge that as OHS use during the follow-up is measured during the same period as the outcome, further studies are required to confirm these associations. Finally, the marginal effects estimates that were presented in sensitivity analyses indicate that the potential for reduction in LTSA rates among FSTSA and frequent OHS user populations is discernible.

Short-term SA spells are relatively common among young employees. Only few had none of them at all, and slightly less than one quarter of young employees had six or more short-term SA spells in a year. This led us to use six spells threshold for FSTSA, which is higher than the usually used threshold (Notenbomer et al. 2019; Koopmans et al 2008; Sumanen et al. 2017). In line with previous studies, we found that accumulation of short-term SA predicted subsequent LTSA due to mental disorders (Notenbomer et al 2019; Koopmans et al 2008; Laaksonen et al. 2013; Sumanen et al. 2017). Our results complement these results by finding that employees with FSTSA were also more likely to have a higher level of use of OHS both before and after a short-term SA spell, which in turn was associated with an increased risk for LTSA due to mental disorders. Our results provide grounds for further investigation of accumulation of short-term SA together with primary care utilisation data.

Our results point out the gendered distribution of both LTSA due to mental disorders and the preceding risk factors. We found that FSTSA spells were more common among women than among men, as was subsequent long-term SA due to mental disorders. Moreover, the highest prevalence of SA due to mental disorders was found in the female dominated occupational group of lower non-manual workers, i.e. among occupations such as child and elderly care workers. The result is in line with earlier findings on the gendered occupational statuses and their relation to the prevalence of mental disorders (Halonen et al. 2018; Lidwall et al. 2018). Our results complement these findings by showing that FSTSA and OHS may be used as risk indicators consistently across these occupational classes. However, due to the limitations of the register-based data we could not differentiate between gender-related biological, individual or work-related factors, such as experienced mental distress, poor support from superiors (Foss et al. 2010), adverse psychosocial work environments (Lidwall et al. 2018) or emotional demands (Framke et al. 2019) that are risk factors more prevalent in female dominated health and social care occupations. Despite this limitation, a balanced approach between health care-based and workplace-based interventions in prevention of LTSA due to mental disorders is likely to be required.

We found that the employee’s disease burden at baseline period, measured as accumulation of past SA days and OHS utilisation at baseline, turned out to be a significant factor explaining the link between repeated short-term SA and LTSA. Individuals who, in addition to FSTSA had longer spells of previous SA and high levels of OHS use at baseline were more likely to experience LTSA due to mental disorders. Individuals who, in addition to FSTSA spells, also had preceding longer spells of LTSA due to other reasons than mental disorders or more accumulated days during short-term SA or high levels of OHS use were in the highest risk of experiencing subsequent LTSA due to mental disorders. Therefore, our analysis suggests that FSTSA spells, alongside with other indicators that are retrievable from electronic records, may accurately indicate altered levels of mental burden and processes leading to LTSA due to mental disorders. Our results thus highlight the importance of sensitivity to accumulation of short-term SA together with other indicators of cumulative disease burden at baseline period. As the symptoms at early stage could be more easily treated, further investigation is needed to explore whether the measures applied in this study could be used as proxies for the symptoms of common mental disorders (Framke et al. 2019; Knudsen et al. 2013) or mental well-being and work capacity (Bertilsson et al. 2015) that have also been proven to operate as risk-factors to the onset, duration and recurrence of subsequent SA.

As far as we are aware, this is the first study to explore the role of OHS primary care use in the progress from FSTSA to LTSA due to mental disorders. The registers of the largest employer of Finland were used in this study. The registers are comprehensive and reliable, cover all employees, information of OHS visits and all SA spells. Thus, this study avoids a common limitation of OHS utilisation studies that are based on cross-sectional data, or limited samples with data on only those who have used the services. However, limitations of the data include the lack of information of diagnoses regarding short-term SA. Despite this limitation we found that the routine OHS utilisation and SA data could be used to identify those who will have subsequent LTSA due to mental disorders. The study also indicates that OHS visits alone are not sufficient in preventing LTSA and resources might be used in potentially supportive actions, such as cognitive behavioural therapy and problem-solving therapy (Lexis et al. 2011).

Previous evidence underlines the importance of monitoring the accumulating short-term SA in order to utilise preventive measures at early stages, already before more severe work disability leading to longer SA occurs (Kivimäki et al. 2004; Koopmans et al. 2008). The Finnish OHS and workplaces have widely reacted to this, and the FSTSA monitoring is generally a part of their early support models. Early support models are one of the key elements of statutory work ability management performed in co-operation between OHS and workplaces. Monitoring FSTSA is possibly a very feasible preventive measure and also an equal way of implementing early support when and where needed, as the operating models are written policies that are applied to all employees.

A cautionary remark on the generalisability of the results is warranted. For example, we studied a population consisting of municipal employees in a particular OHS context. However, the OHS system is the main primary care service provider for the working population in Finland (Ikonen et al. 2013; Virtanen and Mattila 2011), and the principle of primary care use is similar as in the general practice in most other Western countries. Our findings obtained from a reliable longitudinal register data encourage integrating indicators from primary health care services in sickness absence studies to inform the programs aiming at preventing and mitigating the risks for LTSA due to mental disorders.

Both frequent short-term sickness absences and occupational health service utilisation were associated with subsequent long-term sickness absence due to mental disorders. Routinely collected data on health service utilisation could be used as a source of information for identification of individuals at risk of long-term sickness absence due to mental disorders and targeting preventive measures for them.