Sickness absence is a major health concern and considered as a global measure of health status [1]. In Sweden, sickness absence has increased by 70% since 2010 and governmental spending on health insurance has increased by 11 billion Swedish crowns (around 1.2 billion Euro by December, 31 2014) between 2010 and 2014 [2].

Some important contributing factors to sickness absence are found in the work environment; such have low job control and low decision latitude at work been shown to increase the odds for sick leave [3–5]. A relatively newly established determinant of employee health is employees’ perceptions of fairness in the organization, i.e., organizational justice. Organizational justice can be divided into several sub-dimensions: the fairness of formal decision-making in the organization (procedural justice), the perceived fairness of decision outcomes (distributive justice), and the treatment of employees by supervisors (interactional justice) [6]. The interactional justice component of organizational justice is the focus of this study, and can be further divided into informational justice, i.e., receiving truthful and candid information with adequate justifications, and interpersonal justice concerning the respectful and dignified treatment from the supervisor [6].

Relatively little is known about how organizational justice relates to health and only since around the turn of the millennium research on organizational justice has been expanded to health outcomes [7]. Studies have investigated the relationship between justice perceptions and different health indicators, such as well-being, burnout, health problems, or sickness absence [8, 9]. Regarding sickness absence, the few existing studies generally suggest that perceived low organizational justice increases the risk of sickness absence. However, there is little consistency in the results and between-study differences are immense [9]. This may be due to the heterogeneity in sickness absence measures and an insufficient differentiation between long-term, short-term and frequent sickness absence. A sufficient differentiation between sickness absence measures is of importance; different mechanisms have been suggested to lie behind long and short term sick leave. While short-term sickness absence might be a means for the individual to get relief from high strain and therefore could be considered as a coping mechanism (the ‘withdrawal’ explanation), long-term sickness absence might be a sign of more serious health problems due to a prolonged stress reaction to poor work conditions (the ‘stress’ explanation) [5, 6]. However, to be able to look at different indicators of sickness absence, and to differentiate between long term and short term sickness absence, truly longitudinal studies with repeated measures of both predictor and outcome are needed. Until today, relatively few studies have examined the organizational justice–sickness absence relationship using longitudinal analysis techniques which take repeated measurements of organizational justice into account [10–12].

Consequently, the first aim of our study is to investigate the association between informational and interpersonal justice and long and frequent sickness absence, respectively, employing repeated measured covering a time span of six years.

The Uncertainty Management Model states that justice perceptions are of greater importance for people in insecure situations than for those in secure situations [13]. It is assumed that in uncertain situations, individuals use fairness as a heuristic to judge whether they can trust their employer and supervisor. If individuals perceive unfairness in uncertain times, they may perceive their situation as even more unpredictable, which may just increase their feelings of stress and their tendency to call in sick. Indeed, previous experimental studies indicated that affective and behavioural reactions to perceived justice were stronger among those in uncertain situations [13]. A cross-sectional study reported a stronger association between procedural injustice and stress symptoms among employees with high compared with low job insecurity [14]. Another study found that job insecurity accentuated unfairness perceptions, which related negatively to well-being, but also here evidence is cross-sectional [15]. The Uncertainty Management Model assumptions were also supported by a study which found that associations between procedural and interpersonal justice and sickness absence were dependent on experienced work-time control and perceived changes at work [16]. However, while low interpersonal justice may increase the risks for long and frequent sickness absence due to high stress levels, low informational justice may lead to decreases in frequent but short absence when there is high job insecurity, since employees may go to work and seek information to reduce their uncertainty. Still, to the best of our knowledge, very little is known about the role of informational and interpersonal justice for sickness absence in times of high uncertainty at work, such as high job insecurity.

Thus, our second aim to investigate the association between interpersonal and informational justice at work and long and frequent sickness absence under conditions of job insecurity.

In line with our expectations, perceptions of lower levels of interpersonal and informational justice increased the risk of long as well as frequent sickness absence, using standard GEE. These findings suggest that lower levels of justice at work might relate both to increased withdrawal behaviour in form of shorter, but more frequent sickness absence periods and to more serious health effects as indicated by an increased risk of longer sickness absence episodes. Thus, our results are in line with earlier research showing that low levels of organizational justice are associated with an increased risk for ill-health and sickness absence [16, 32, 33]. However, observed differences in risk ratios regarding the two different sickness absence measures were small and future studies should further investigate possible differentiating effects of organizational justice on long and frequent sickness absence.

Based on the results from the standard GEE, no conclusions can be drawn regarding the temporal relationship between low justice perceptions and sickness absence, i.e. we do not know if effects of organizational justice are due to differences in perceived organizational justice between participants (cross-sectional) or within participants (longitudinal). Thus, using autoregressive GEE we removed the between-subjects part of the relationships by taking previous periods of sickness absence into account and investigated possible longitudinal associations between organizational justice and sickness absence. Regarding interpersonal justice, autoregressive GEE revealed that the increased risks of long and frequent sickness absence, respectively, can be explained by within-effects, i.e., having experienced a decrease of interpersonal justice at a personal level related to an increase in sickness absence in that person. Regarding informational justice, we found no significant increased risk of long sickness absence using autoregressive GEE, i.e., our results indicate that our findings from standard GEE rely mainly on between-effects. In other words, people with lower justice perceptions have an increased risk of long sickness absence in comparison to those with higher justice perceptions. With regard to frequent sickness absence, we found a small statistically significant effect of informational justice, indicating that changes in informational justice perceptions were related to increased withdrawal behaviour in form of frequent sickness absence episodes.

In a next step, we investigated the relations justice and sickness absence in interaction with job insecurity. In contrast to our expectations, we could not find any statistically significant interaction between justice perceptions and job insecurity; i.e., the association between lower levels of justice perceptions and any type of sickness absence did not differ under conditions of higher levels of job insecurity. It is quite possible that this null-finding can be explained by the general low levels of perceived job insecurity in our study population; seventy percent of the study population did not report any job insecurity at all.

Our results contradict previous findings showing that the relationship between low justice and strain was moderated by uncertainty-related aspects of the workplace [14, 16]. However, while we measured job insecurity in terms of being afraid of losing one’s job, Elovainio et al. operationalised insecurity in terms of control over working hours and negative changes [16]. Also, the authors of the uncertainty management model are quite ambiguous in their definitions of uncertainty and it might make a difference whether we investigate environmental uncertainty (e.g., job insecurity) or personal uncertainty (e.g., lack of standing/respect/belonging to a work group, self-esteem issues). Also, job insecurity might not be bad for everyone, e.g., a temporary work contract might be accepted as a step in career or as a possibility which offers flexibility.

Our findings indicate that lower levels of informational and interpersonal justice are associated with an increased risk for long and frequent sickness absence. Also, higher levels of job insecurity turned out as an important predictor of both long and frequent sickness absence. Perceived fairness at work is a modifiable aspect of the work environment, as is job insecurity. Organizations have significant control over both aspects and our results suggest that organizations may gain by investing or improving their policies and rules for fair treatment of their workforce and by improving job security. Further, organization might gain by selection of supervisors for their qualities associated with fair practices, training supervisors in justice principles, and implement performance management practices for them that consider their use of organizational justice [45]. Indeed, training in justice principles has been shown to be successful in different organizational contexts and is connected with relative lower costs as compared to other justice interventions (e.g. raising pay) [46].