*3.4. Quality Assessment of the Included Studies*

According to the "Standard quality assessment criteria for evaluating primary research papers from a variety of fields" [45], the mean article quality score was 0.82 ± 0.10 out of a total of 1 (Table 3). Eleven articles scored below the mean score with a minimum score of 0.61. Fourteen articles scored above the mean score with a maximum score of 0.95.


**3.**Qualityassessmentoftheincludedstudies.

#### *Nutrients* **2020** , *12*, 397



### **4. Discussion**

In this systematic review, an overview of studies investigating the association between energy balance-related behavior and burn-out is provided. In total, 25 studies were found, of which 21 assessed the relationship between physical activity and risk of burn-out, two studies investigated the association between physical activity, sedentary behavior and the risk of burn-out, and two assessed the link between physical activity, dietary behavior and risk of burn-out. No articles using a holistic approach—i.e., investigating the relationship between energy balance and energy balance-related behavior as a whole and risk of burn-out—were found. Nevertheless, it is important to use a combined approach to gain more insight regarding, e.g., which behavior may have a bigger impact on burn-out risk. Moreover, as previously mentioned, van der Ploeg et al. [27] indicated that sedentary behavior and physical (in)activity should be targeted at the same time in public health strategies. The present systematic review, for example, shows that the association between physical activity and risk of burn-out was stronger in workers with a sedentary job, showing a possible interaction between both physical activity and sedentary behavior in relation to burn-out risk [58]. A similar interaction between both behaviors, but with mortality as the outcome measure, was found in a large-scale meta-analysis (including over one million men and women), where high levels of moderate intensity physical activity (about 60–75 min per day) eliminated the increased risk of death associated with high sitting time [71]. In addition, previous studies showed that increased screen time was associated with an overall poor diet quality [72,73]. The above indicates possible triangular interactions, again highlighting the importance of combining all energy balance-related behaviors when investigating their association with burn-out risk.

Furthermore, despite the fact that no studies investigating the relationship between energy balance (i.e., energy intake vs. energy expenditure [74]) and burn-out were found, a positive or negative energy balance may also be associated with burn-out. The interaction and co-existence of energy balance-related behaviors determine whether or not a positive or negative energy balance is experienced [75]. As energy imbalances may lead to the development of overweight and obesity (= physical health) [74], one may hypothesize that a similar imbalance may lead to decreased mental health as well. Previous research, in fact, demonstrates that obesity is associated with higher levels of burn-out [76].

Despite the heterogeneity of populations, assessment methods for both risk of burn-out and physical activity, and physical activity interventions, the vast majority of the experimental studies, including 5 RCTs, concluded that physical activity is effective in reducing the risk of burn-out, suggesting a causal link between both. The experimental studies showed a decrease in burn-out risk ranging from 6.9% to 41.3% [49,52–55], a decrease in emotional exhaustion between 10% and 33.4% [47,48,51,56] and a decrease in cynicism of 33.3% and 4% [48,51]. One experimental study by Freitas et al. [50], showing a relatively low quality score of 0.63, did not find any significant effect of physical activity on risk of burn-out. This may be due to the small sample size (*n* = 21) and the duration of the physical activities performed in this study. The participants had to perform a 10-min workplace physical activity session on weekdays (no information on type or intensity was provided), while the World Health Organization (WHO) [16] recommends that adults should be physically active at a moderate intensity for at least 150 min per week or at a vigorous intensity for at least 75 min per week or an equivalent combination of both moderate and vigorous intensity activity. So, the total duration of the physical activities performed in the study by Freitas et al. [50] (ten minutes per day, for five days a week) was not meeting the above guidelines [77]. On the other hand, the study by Stenlund et al. [54], in which the performed physical activities (60 min, twice a week, at a moderate intensity) also failed to meet the WHO guidelines, did report a significant effect of physical activity on burn-out. It should be mentioned that, despite not meeting the recommendations, total physical activity duration in the latter study was still much higher compared to the study by Freitas et al. [50] (i.e., 120 min versus 50 min per week, respectively). In another experimental study [55], two groups of participants performed highand low-intensive physical activities, respectively, for 60 min twice a week (=120 min in total). The low

intensive group also failed to meet the WHO guidelines. Nevertheless, both groups (high vs. low intensity) had more or less the same effect on burn-out (reductions of 8.1% and 8.2%, respectively) [55]. These findings suggest that lower (than recommended) amounts of physical activity of 120 min per week (at a low to moderate intensity) may already be effective in reducing burn-out risk. Furthermore, the role of intensity may be questioned. It should be mentioned, however, that the physical activities performed in the other experimental studies (showing a positive effect of physical activity on burn-out risk) were in line with the WHO guidelines, as they lasted 20 to 60 min for two to five times per week at a moderate to vigorous intensity [47–49,51,53]. The remaining two experimental studies did not give clear information about the duration of the physical activity sessions [52,56]. Further, four experimental studies [47,52,53,56] also conducted follow-up measurements and showed a decrease in risk of burn-out three, six and 12 months after the interventions, indicating long-term effectiveness, even when the physical activity intervention did not remain in place.

Our results are in line with the systematic review by Naczenski et al. [1] showing strong evidence for the effect of physical activity on reducing (emotional) exhaustion, but limited evidence for the effect on professional efficacy and cynicism. Further, Naczenski et al. [1] concluded that being physically active one or two times per week for four to 18 weeks has promising effects on reducing burn-out symptoms. The present systematic review slightly deviates from this conclusion, showing that positive effects on burn-out risk were achieved when being physically active two to five times per week for 20 to 60 min, for six to 18 weeks, with 18 weeks showing the biggest reduction in burn-out risk (−47.8%) [53]. The latter suggests that the longer the duration of the physical activity intervention, the higher the reduction in burn-out risk. However, due to the large variety in type, intensity, duration and frequency of the performed physical activities in the included studies, comparison of the effectiveness of the individual interventions remains difficult. Further research to unravel the respective effects of type, intensity, duration and frequency is therefore highly recommended. Furthermore, to better understand the relationship between physical activity and burn-out risk, it is also important to investigate the underlying physiological mechanisms. For example, the role of BDNF might be interesting as BDNF-levels increase when physical activity is performed, while on the other hand, BDNF-levels were found to be decreased in people having burn-out [22]. Four experimental studies [49,52,54,56] compared physical activity interventions with other interventions—such as basic care, cognitive interventions and a multimodal rehabilitation program—and did not find physical activity to be more effective compared to the other treatment arms. It should be mentioned that, in one of these studies [52], the aforementioned conclusion was based on the results of six and 12 months after the intervention was completed, possibly causing differences between treatment effects to have been diminished. Another possible reason for these results may be a shared effect between the interventions, resulting in no effect of intervention type [78]. As suggested by Heiden et al. [52], the same (psychosocial) attention was given to all patients in both interventions, suggesting (psychosocial) attention to be such a shared effect. Furthermore, as human interaction was also present in the other experimental studies (e.g., interaction between the researcher or therapist and the participant during the intervention phase), part of the intervention effects might be explained by the same psychosocial component [49,52,54,56].

These results might suggest that physical activity may be equally effective compared to other types of interventions, as long as there is a psychosocial component involved. Future research should further unravel the relative importance of physical activity versus other intervention components when aiming at reducing burn-out risk.

Regarding sedentary behavior, only positive relationships with risk of burn-out were reported, indicating that higher levels of sedentary behavior are associated with higher burn-out risk. It should be mentioned, however, that only two observational studies [62,67] were found, making it difficult to draw any firm conclusions. Nevertheless, the quality score of these two studies was high (0.86 and 0.95). Besides, it is important to mention that these studies primarily aimed to assess the relationship between physical activity and risk of burn-out. In both studies, the assessment method of physical activity was the 4-level Saltin Grimby Physical Activity Level Scale [79], of which the lowest

level reflects sedentary behavior. This level means that participants were not participating in any leisure-time physical activities or sport activities, which is not in line with the definition of sedentary behavior [25]. Furthermore, no validated objective (e.g., inclinometers, accelerometers) nor subjective (e.g., more detailed context-specific questionnaires) assessment methods for sedentary behavior were used. Despite the fact that sedentary behavior may influence mental health and despite its impact on the risk of the occurrence of common mental disorders such as depression, decreased mood and anxiety, no experimental studies investigating the influence of sedentary behavior on burn-out risk were found. Furthermore, the study by Bernaards et al. [58] concluded that the association between physical activity and burn-out risk was stronger in workers with a sedentary job. This suggests a possible moderating role of sedentary behavior in the physical activity-burn-out risk relationship pathway. So, the interaction between physical activity and sedentary behavior should be further investigated while explaining burn-out risk.

Two observational studies [46,69], with quality scores of 0.80 and 0.91 respectively, found that a healthier diet is related to a lower risk of burn-out. Because only two studies were found and the fact that these two studies were investigating different aspects of dietary behavior, namely fast food consumption and the amount of healthy diets during work days, it is difficult to draw reliable conclusions regarding this relationship. Moreover, these two studies did not use valid and reliable assessment methods to measure both aspects. Alexandrova-Karamanova et al. [46] measured fast food consumption by one single self-constructed question, namely "How many times in a week do you eat fast food?". In the study by Gorter et al. [69] dietary behavior was measured by asking how many healthy diets participants consumed during workdays, while "healthy diets" was not defined. These methodological shortcomings make it even more difficult to draw firm conclusions.

It has been shown that diet may influence mental health. A systematic review and meta-analysis by Tolkien et al. [80] for example, concluded that an anti-inflammatory diet may play an important role in preventing or reducing depression risk and symptoms. Moreover, because of the link between burn-out and neurotransmission and the role diet may play herein, it is important to further investigate the link between dietary behavior and burn-out risk by conducting experimental research using valid and reliable assessment methods (e.g., food diaries, 24-hour recalls or food frequency questionnaires).

Because all studies regarding sedentary behavior, dietary behavior and burn-out risk had an observational design, no conclusions about the causal relationship can be made. As hypothesized above, being sedentary may increase the risk of burn-out through different physiological mechanisms, while a burn-out may also cause people to be more sedentary, and so reversed causality is possible. The same may be true for dietary behavior. This shows the need for more experimental studies investigating the causal relationship between sedentary and dietary behavior and burn-out risk.

There are several limitations to the included studies. A first limitation is the fact that some studies [57,58,69] measured physical activity with only one or two single self-constructed questions. Moreover, some measurement methods for physical activity had methodological shortcomings. The study by Liang et al. [61], for example, measured exercise behavior on a 5-point scale with anchors 1: 'never' to 5: 'many times in a week', while 'many times' was not defined. Future research should use validated questionnaires (e.g., International Physical Activity Questionnaire [81]), and preferably objective measures, such as accelerometers or pedometers. The same can be said for sedentary and dietary behavior. A second limitation is that less than half of the included studies used the Maslach Burnout Inventory (MBI), while the MBI is the gold standard assessment tool for burn-out [82]. All the other assessment tools for burn-out used in the included studies are based on other definitions of burn-out and are mostly measuring one dimension, namely (emotional) exhaustion. It is recommended that future research uses the gold standard assessment tool for measuring burn-out, in order to increase measurement homogeneity across studies. A third limitation is that some studies are mixing up the terms "physical inactivity" and "sedentary behavior". Two of the included studies [57,63] use a physical activity questionnaire to classify people as active or inactive, and considered inactive people to be sedentary, while literature clearly shows that physical inactivity and sedentary behavior are

two different concepts [27,28,83]. A fourth limitation is that the majority of the included studies did not take physiological, psychological and sociological confounders into account, and so conclusions may have to be interpreted with caution. A fifth limitation is that most studies consisted mostly of female participants, which may have influenced the results. Research shows that women generally eat healthier but are less physically active compared to men [84–89]. Moreover, a meta-analysis showed that women are more likely to report burn-out [90]. More specifically, women reported to be more emotionally exhausted than men, while men were more depersonalized [90]. Hence, future studies investigating the association between energy balance-related behavior and burn-out should take sex into account. Lastly, some articles were missing some relevant information, such as how physical inactivity was measured, so the authors had to be contacted. Unfortunately, we did not always get a response leaving some queries unanswered.

There are also a few limitations to the present systematic review. Since non-English written publications were excluded, we may have missed out on important scientific articles in other languages. Additionally, the used quality assessment tool does not distinguish between experimental and observational studies. Experimental studies can be considered higher in quality and so they should receive a higher score in the quality assessment. However, when calculating the mean quality score per study design, a higher mean quality score of 0.87 ± 0.08 for the observational studies was found, compared to a mean quality score of 0.74 ± 0.08 among the experimental studies.

A strength of this systematic review is the fact that this is the first systematic review aiming to include articles of all study designs investigating the relationship between energy balance-related behavior as a whole (i.e., the combination of physical activity, sedentary behavior and dietary behavior) and burn-out risk. As—in the present review—no studies using this holistic approach were found, and because of the hypothesized role these three components may play in reducing or preventing burn-out, further research on this topic is needed.
