Low-quality employment trajectories and risk of common mental disorders, substance use disorders and suicide attempt: a longitudinal study of the Swedish workforce

Objective High-quality longitudinal evidence exploring the mental health risk associated with low-quality employment trajectories is scarce. We therefore aimed to investigate the risk of being diagnosed with common mental disorders, substance use disorders, or suicide attempt according to low-quality employment trajectories. Methods A longitudinal register-study based on the working population of Sweden (N=2 743 764). Employment trajectories (2005–2009) characterized by employment quality and pattern (constancy, fluctuation, mobility) were created. Hazard ratios (HR) were estimated using Cox proportional hazards regression models for first incidence (2010–2017) diagnosis of common mental disorders, substance use disorders and suicide attempt as dependent on employment trajectories. Results We identified 21 employment trajectories, 10 of which were low quality (21%). With the exception of constant solo self-employment, there was an increased risk of common mental disorders (HR 1.07–1.62) and substance use disorders (HR 1.05–2.19) for all low-quality trajectories. Constant solo self-employment increased the risk for substance use disorders among women, while it reduced the risk of both disorders for men. Half of the low-quality trajectories were associated with a risk increase of suicide attempt (HR 1.08–1.76). Conclusions Low-quality employment trajectories represent risk factors for mental disorders and suicide attempt in Sweden, and there might be differential effects according to sex – especially in terms of self-employment. Policies ensuring and maintaining high-quality employment characteristics over time are imperative. Similar prospective studies are needed, also in other contexts, which cover the effects of the Covid-19 pandemic as well as the mechanisms linking employment trajectories with mental health.


INDEX
• Detailed description of the creation of employment trajectories.
• Table S1. Complete and exploratory fit indices for the total population, women and men.
• Table S2. Characteristics of the final employment typology with six classes.
• Table S5. Sociodemographic characteristics of the total population (2005) according to trajectories.
• Table S6a. Sociodemographic characteristics of women according to trajectories.
• Table S6b. Sociodemographic characteristics of men according to trajectories.
• Table S7a. Hazard ratios (crude and adjusted) for common mental disorders for the total population, women and men. • Table S7b. Hazard ratios (crude and adjusted) for substance use disorders for the total population, women and men. • Table S7c. Hazard ratios (crude and adjusted) for suicide attempts for the total population, women and men. • Table S8a. Adjusted hazard ratios for specific outcomes (depression, anxiety, stress related disorders, alcohol and non-alcohol use disorders) for the total population. • Table S8b. Adjusted hazard ratios for specific outcomes (depression, anxiety, stress related disorders, alcohol and non-alcohol use disorders) for women. • Table S8c. Adjusted hazard ratios for specific outcomes (depression, anxiety, stress related disorders, alcohol and non-alcohol use disorders) for men. • Figure S1. Flow chart of the (total) study population.
• Figure S2. Directed Acyclic Graph for the effect of employment trajectories on mental health.

Detailed description of the creation of employment trajectories
Employment trajectories spanning 2005 to 2009, for the total population and for women and men separately, were created in two steps.

First step
Individuals were classified according to employment type (or unemployment) for each year of the exposure assessment window. Individuals that had been unemployed for >180 days during the year were categorized as unemployed. In order to distinguish employment types of precarious nature (or lack thereof) for the non-unemployed population, three dimensions and five items of precarious employment (PE) (as described by Kreshpaj et al 2020), were operationalized for each year by the means of available register data. This process has been thoroughly described in a previous publication (Jonsson et al 2021). Briefly, dimensions, items and their operationalization (within parenthesis) were: 1. Employment insecurity with three items, including (1a) contractual relationship insecurity (Directly employed by the employer; Employed by an agency; Combination of selfemployed and directly employed; Self-employed with >1 employee; Solo self-employed), (1b) contractual temporariness (Stable employment. i.e., having the same employer for three consecutive years; Unstable employment. i.e., not having the same employer for three consecutive years), and (1c) multiple jobs/sectors (1-2 jobs (employers) during the year; >3 jobs during the year; >3 jobs in >1 economic sector) 2. Income inadequacy containing (2a) yearly income level (>200% of the median income of the study population in the specific year; 120-199% of the median; 80-119% of the median; 60-79% of the median; <60% of the median) 3. Lack of rights and protection including (3a) lack of unionization (operationalized as likelihood of being covered by collective bargaining agreement in the organization of employment; >90%; 71-90%; 0-70%).
For the sake of the operationalization of contractual temporariness, employer data from LISA was also retrieved for years 2003    Directly employed or employed by an agency with unstable employment and one job. Mainly median or around median income, high to moderate collective bargaining agreement coverage.

Solo Self-employment
Solo self-employed with stable employment and one job. Poor income and poor collective bargaining agreement coverage.

Hybrid Multiple Job-Holding
In combined employment (directly employed and self-employed). Unstable employment and the largest proportions of multiple jobs and multiple jobs in multiple sectors. Income and collective bargaining agreement coverage distributed across all levels.

Business Ownership
Self-employed with stable employment and one job. Median to high income and poor to moderate collective bargaining agreement coverage.

Trajectories characterized by mobility
Classification: Spending a minimum of one year in one employment state at the beginning of the trajectory, followed by a minimum of two years in another employment state at the end of the trajectory. The last two years could also be characterized by combinations of LQ-or HQ-employment states.   Within EU-28 7.0 6.9 6.6 5.0 8.3 5.2 11.9 6.9 8.0 6.7 9.3 4.7 6.9 6.9 6.9 5.0 5.7 8.9 6.6 6.9 5.6 6.8 Outside