The prevalences of high need for recovery and burnout symptoms in the current study population were 38.4% and 20.3% respectively.

The prevalence of high need for recovery found in the current study was similar to the percentage found in an earlier study in workers in the Flemish public sector (37.7%) [8]. However, in the current study younger workers tended to have a higher need for recovery than older worker, which is in contrast with earlier reports [8, 9]. This inconsistency could be explained by the fact that in the current population the older employees are in a more “comfortable” occupational and family related position than the younger ones. It could be argued that the older employees are more likely to occupy a permanent position as a member of the higher academic staff, while the younger are more likely to have a more precarious employment status (such as temporary PhD scholarship) and to have a less “consolidated” family life situation.

Compared to a random sample from the Flemish workforce representative of age, gender and industry [7] our results showed a lower prevalence of burnout symptoms (20.3% vs. 25.6%). The difference could be due to a different constitution of the study population. The current population was a specific university population, comprising a substantially higher proportion of female employees (64.7% vs. 45.7%) with a lower mean age (38.8 yrs. vs. 41.3 yrs.) and, because this concerns a population of university workers, one could expect a higher proportion of higher educated employees.

Having a poor work-private life balance was by far the most important factor for the two considered mental health outcome parameters, both in univariable and multivariable analyses. In the multivariable models the odds ratios reached 5.14 for having a high need for recovery and 2.80 for having burnout symptoms. These findings are in accordance with the findings of earlier studies. In a representative sample of the working population, employees with high work life conflicts showed a high risk of anxiety and depression, lack of energy and optimism, headaches, sleep disorders and fatigue [13]. Similarly, in a sample of academics, poorer work-life balance was associated with perceived job stress [14]. Our findings are also in line with the theoretical framework of need for recovery: if a poor work-life balance does not enable to recover sufficiently after work, the need for recovery will increase, eventually leading to a high need for recovery.

Having no quiet place to work was significantly associated with the presence of burnout symptoms and highly significant associated with a high need for recovery, with odds ratios of 2.00 and 3.23 respectively. To our knowledge these associations have not been reported before. Earlier studies have shown that an “open plan” working place, also a situation where employees have less privacy to concentrate on their work, can induce increased stress and other adverse health effects [22]. A possible explanation could be that the absence of a quiet workplace induces a situation where it is difficult to concentrate, which is an additional stressor, consuming extra energy and resulting in a higher need for recovery and subsequently burnout.

Being worried about short-term work situation and being worried about long-term work situation were both significantly associated with need for recovery and burnout in univariable analyses (p < 0.001). However, in the multivariable analyses only being worried about long-term work situation was significantly associated with the presence of a high need for recovery (odds ratios ranging from 2.18 to 4.98) and burnout symptoms (odds ratios ranging from 1.49 to 6.30). For both outcome parameters the OR’s showed a logical gradient: the more worried about long-term work situation, the higher the risk for the presence of a high need for recovery and burnout symptoms.

No literature was found on the relationships of being worried about long-term work situation with need for recovery and burnout. However, it could be hypothesized that job insecurity can be considered as a worry of the employee about his current employment. Therefore, we considered job insecurity as an alternative for being worried about long-term work situation in our search for literature. Job insecurity has been found to be related to high need for recovery in earlier studies. In a study in employees from various sectors, job insecurity was found to be positively related to need for recovery [23]. This was confirmed in subsequent research, where in a population of public sector employees job insecurity was significantly associated with a high need for recovery [18]. Job insecurity was also shown to be significantly associated with exhaustion, a core component of burnout [23, 24] and positively associated with burnout [25, 26]. In the specific case of university personnel the worries relate to not meeting the research deadlines due to the closure of research labs, and hence jeopardizing their scholarship, and subsequently resulting in a postponing or even a cancelling of their future academic career. In that regard our results seem to corroborate earlier findings. A possible explanation could be that being worried about long-term work situation is a stressor that consumes extra energy, resulting in a higher need for recovery, and subsequently to burnout.

No significant association between worries being discussable with the supervisor and need for recovery was found. This in concordance with earlier research where social support from supervisor was also not associated with need for recovery [18].

Our results showed only a significant association between worries being discussable with the supervisor and burnout if the worries were discussable to a very small extent. Earlier research showed that fewer sources of support go along with a significantly increased risk of burnout symptoms [27] and that supervisor support was negatively associated with burnout [28]. It was also suggested that supervisors can help university employees lower emotional exhaustion by reducing the degree of the perceived uncertainties [29]. Our results indicated a less pronounced negative relationship (only if the worries were discussable to a very small extent). This could be due to the very specific situation of the current study population. It could be argued that in mandatory home work implemented without any preparation, contacts with the supervisors are not institutionalized properly from the beginning and are less a factor of concern.

No significant associations were found with professional status. Apparently, the mechanisms for having a high need for recovery or burnout symptoms are not related to professional status, but rather to factors that interfere with energy levels.

Factors related to the family situation (taking care of ill or old people, number of children < 12yrs. at home and having a family member at risk for Covid-19) were not associated with both mental health outcomes, neither in univariable and multivariable analyses. This is in accordance with earlier research, where no significant associations were found between need for recovery and taking care of ill or old people, number of children < 12yrs. at home [8, 18]. This could be an indication that people are more likely to cope with the (normal) familial situation, than with external stressors. Possibly, taking care of family members is more close to our ‘natural’ behaviour (our genetic program is designed to maintain the species, which involves the ability to taking care of family members) than dealing with the artificial institution of work (for which our genetic program has not been designed).

The strongest associations were found for the presence of a high need for recovery. This could be explained by the fact that a high need for recovery is a sensitive indicator of reduced well-being in its very early stage [1], while the presence of burnout symptoms reflects already the presence of a long term health effect, making this measure less sensitive than the need for recovery. It could be expected that the higher sensitivity of the need for recovery will make it more ‘vulnerable’ for certain risk factors, resulting in higher effect measures (higher odds ratios).

This study is subject to some limitations. First of all, this study has a cross-sectional design, not allowing to draw reliable conclusions on causal relationships between the considered factors and health outcomes. The direction of the causal relationship cannot be determined with the current study design; a longitudinal study is needed to allow causal interpretations. Furthermore, it cannot be excluded that people with a poor mental health (high need for recovery or burnout symptoms) might perceive certain factors as worse than people with a good mental health. The associations found could be inflated by the state of mind of the participants.

Secondly, although a few home environment factors have been asked for, it is clear that many factors in the (psychosocial) environment of the home setting (e.g. interpersonal relationship issues, …) could be of influence on the mental health of the home working employee. Mandatory home work could force the employee to work in a “toxic” familial setting and hence influence his/her mental health and productivity.

Thirdly, job related factors (quantitative demands, emotional demands, …) were not asked for and could possibly have an influence as well [18]. It is not sure if they would overrule the significant associations found in this study or only be additional significant factors.

Fourthly, the rather low response rate (45.2%) could be subject to selection bias. This is relevant when there is a difference between responders and non-responders in mental health outcomes and influencing factors and when the reason for nonresponse is correlated with the variables investigated. However, low response rates need not necessarily lead to biased results. Bias is more likely to be present when examining a simple univariable distribution than when examining the relationship between variables in a multivariable model [30].

The study design of the current study did not allow to monitor changes. Future (longitudinal) research should focus on changes in mental well-being when a mandatory shift from workplace office to home office is implemented. Simultaneously, the influence of the changing environmental factors when moving from the workplace environment to the home environment should be explored. Impact of “non-traditional” occupational factors linked to the home environment should be studied as well, with special focus on the psychosocial home environment.