*3.2. Analysis of Variance*

The analysis of variance fixed effects omnibus test (Type III analysis of variance with Satterthwaite's method), conducted on the LMMs, revealed no significant differences between frontline and non-frontline HCWs concerning personality traits, F(4, 2720) = 1.664, *p* = 0.155 (see Figure 2A), intolerance of uncertainty, F(1, 680) = 0.131, *p* = 0.718 (see Figure 2B) and coping strategies, F(1, 1360) = 2.253, *p* = 0.106 (see Figure 3A).

**Figure 2.** Effect plot of: (**A**) personality traits; (**B**) intolerance of uncertainly. The bars represent the 95% CI.


**Table 1.** Descriptive statistics of the sample characteristics.

**Figure 3.** Effect plot of: (**A**) coping strategies; (**B**) perceived stress of groups. The bars represent the 95% CI.

In particular, in both groups of HCWs


On the contrary, the results of the analysis of variance omnibus test (Type III analysis of variance) conducted on the linear model (LM) revealed significant differences between the two groups about the perceived stress, F(1, 680) = 9.394, *p* = 0.002, Cohen's d = 0.240, i.e., the quantitative size of the estimated effect is closer to the small than to the medium value. Specifically, the levels of perceived stress were higher among the frontline Italian HCWs (M = 22.032, SD = 8.649, 95% CI [21.447, 22.617]) rather than among the non-frontline (M = 20.042, SD = 8.119, 95% CI [19.584, 20.501]) (see Figure 3B).

In Figure 4, we report the correlations and the descriptive statistics for personality traits, intolerance of uncertainly, coping strategies, and perceived stress.

**Figure 4.** Pearson's Correlations and descriptive statistics of the experimental variables (\* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001).

#### *3.3. Multi-Group Path-Analysis*

We ran the multi-group path-analysis (see path-diagram in Figure 1) using the "Diagonally Weighted Least Squares" (DWLS) estimator. For a proper analysis, the minimum ratio between the number of observations and the number of model parameters should be greater than 5:1 [85]. In our case, we had 110 estimated parameters with 682 participants, therefore the ratio was 6.2:1, and the sample size was adequate. We obtained adequate fit indices: CFI = 0.959, TLI = 0.917, RMSEA = 0.077, and SRMR = 0.060.

Using RMSEA as effect size and alpha = 0.05, the results of the post-hoc power analysis show that a sample size of N = 682 is associated with a power larger than 99.99%.

The Chi-squared difference test between the multi-group unconstrained and constrained models revealed significant differences (see Table 2) between Italian frontline and non-frontline HCWs concerning the associations (structure of the relations) among personality traits, intolerance of uncertainty, coping strategies, and perceived stress.



The arrows in the following Figure 5 show respectively the significant association identified among the Italian frontline (see Figure 5A) and non-frontline HCWs (see Figure 5B). Asterisks indicate the level of significance of the estimated effects. Just by looking at the two figures, it is possible to notice how the structure of the relations among the variables differs noticeably in the two groups of HCWs and is more complex in the non-frontline one.

**Figure 5.** Structure of relations in the: (**A**) Frontline group of Italian HCWs; (**B**) non-frontline group of Italian HCWs. Signif. codes: 0 "\*\*\*" 0.001, "\*\*" 0.01, "\*" 0.05, "." 0.1.

The analysis revealed significant differences in the structure of the associations between the two groups of HCWs, concerning specifically the relations between:


Regarding the relation between coping strategies and stress no difference was identified between the two groups. In both of them, emotionally focused coping was negatively

related with perceived stress (i.e., the more emotion focused strategies, the less stress), whereas the dysfunctional one was positively correlated with stress (i.e., the more dysfunctional strategies, the more stress).

#### **4. Discussion and Conclusions**

COVID-19 was (and is) an arduous challenge for HCWs all over the world, but especially for those working in the areas characterized by a rapid spread of the virus, which caused, specifically during the first waves, high numbers of hospitalizations (in intensive care units) and deaths, such as in Italy [22].

Many research studies, focused on the impacts of COVID-19 on mental health [98], have revealed that HCWs, involved in fronting this pandemic as those engaged during the past ones [14–16], were at particular risk of developing severe mental symptoms due to the very demanding [24], uncertain [44], and stressful situation. Nevertheless, it is reasonable to assume that such psychological outcomes are also due to more specific contextual conditions (first of all, as argued above, having worked on the frontline or not), as well as to individual differences in personality traits, ability to tolerate uncertainty and to cope with it. Thus, the main aim of the present paper was to investigate, in a sample of 682 Italian doctors and nurses, whether specific job conditions, i.e., having worked in a dedicated COVID ward (280) or not (402) during the first wave of the pandemic in Italy, would have determined differences in the associations (i.e., in the structure of relations) between personality traits, intolerance to uncertainty, coping strategies, and perceived stress.

In line with our expectations, the analysis (LMMs) did not reveal significant differences between the two groups of Italian HCWs with regard to personality traits, intolerance of uncertainty, and coping strategies. In other words, the two groups of HCWs appear to be homogeneous not only in terms of dispositional traits, but also in terms of coping strategies adopted to face the situation, at least in the first phases of the pandemic. Conscientiousness and high levels of prospective intolerance of uncertainty seem to characterize our sample of HCWs. Both the ability to be organized, responsible, and productive (typical of the conscientious personality) and to not be paralyzed by uncertainty—but on the contrary to be engaged in information seeking, which is a typical trait of the prospective intolerance of uncertainty—seem to be focal in HCWs' work. Furthermore, resorting to coping strategies mainly focused on emotions (acceptance, seeking emotional support, humor, positive reframing, religion) seems understandable in a situation such as a pandemic, which, especially in its early stages, was characterized by high levels of uncertainty and unpredictability, and which confronted health workers every day with suffering and death.

Nonetheless, the analysis (LM) revealed higher levels of perceived stress among the frontline HCWs rather than in the non-frontline. This finding is consistent with the results of previous studies on both past epidemics [99] and the COVID-19 pandemic [17]. Specifically, during the outbreak in 2020, both Italian [100,101] and Chinese [21] frontline HCWs reported high levels of perceived stress and were more exposed to psychological burden than second line HCWs in terms of anxiety, depression, insomnia, and distress [25]. Moreover, this data seem to be understandable in the light of the increased risks faced by the frontline HCWs.

Multi-group path-analysis based on the lavaan R-software package was used and the results mainly confirmed our hypotheses, revealing for the two groups of Italian HCWs different models of associations among the variables taken into account (see Figure 5A,B). Specifically, the analysis revealed more complex associations in the non-frontline HCWs' group. This could be due to the greater heterogeneity of this group of healthcare professionals who, unlike the group who worked in dedicated COVID wards, continued to work in different types of wards (which are characterized per se by an intrinsic diversity).

Specifically, significant differences were found between frontline and non-frontline Italian HCWs concerning the associations between: personality traits and intolerance of uncertainty; intolerance of uncertainty and coping strategies.

As for the associations between personality traits and intolerance of uncertainty, among the non-frontline HCWs, high levels of conscientiousness are negatively related to prospective intolerance of uncertainty. In other words, the more they are conscientious (i.e., able from an organizational, productive, and responsible point of view), the less they show need and desire for predictability and active engagement in increasing their certainty. High levels of agreeableness and open-mindedness are, instead, in the same group, negatively related with inhibitory intolerance of uncertainty. In other words, the more individuals are agreeable (i.e., confident and compassionate) and open-minded (i.e., curious and open to the unexpected events), the less they seem to be paralyzed by inhibitory uncertainty. Vice versa, among the frontline HCWs, personality traits seem to have had a poor influence on intolerance of uncertainty, except for neuroticism, that seems to act similarly in both groups of HCWs by increasing both the prospective and inhibitory intolerance of uncertainty.

As for the associations between intolerance of uncertainty and coping strategies, no significant relationship has been identified among the frontline HCWs. It is as if knowing with certainty dealing with infected people reduces the effect of the intolerance of uncertainty on coping strategies. On the contrary, among the non-frontline HCWs, prospective intolerance of uncertainty is positively linked with problem and emotionfocused coping strategies. In other words, HCWs with high levels of prospective intolerance of uncertainty resort to problem and emotion-focused strategies in facing the situation, i.e., acting in order to reduce uncertainty. In this sense, prospective uncertainty seems to be predictive of greater use of functional coping strategies among health professionals. Vice versa, inhibitory intolerance of uncertainty is negatively related to problem and emotion-focused coping strategies (in line with its paralyzing traits) and positively related with dysfunctional coping strategies (in line with its avoidance characteristics). In other words, higher levels of inhibitory intolerance of uncertainty seem to be predictive of greater recourse to dysfunctional coping strategies. In this sense, inhibitory intolerance of uncertainty seems to have a more negative impact on HCWs' ability to cope with stressful situations rather than the prospective one.

Interesting similarities were found instead between the two groups of HCWs regarding the role of negative emotionality (neuroticism) in affecting intolerance of uncertainty, and concerning the association between coping strategies and perceived stress.

The finding according to which negative emotionality (neuroticism) affects intolerance of uncertainty is in line with the results of much research, mainly concerning general population samples [49,59,102,103], also during the pandemic [104], according to which "poor emotional regulation skills contribute to intolerance to uncertainty" [105] (p. 4). Irrespective of having worked on the frontline or not, Italian HCWs with high levels of neuroticism also showed high levels of intolerance of uncertainty (both prospective and inhibitory). In other words, neuroticism seems to be predictive of high levels of intolerance of uncertainty also among healthcare professionals irrespective of their being frontline or not.

As for the association between coping strategies and perceived stress, in both groups, resorting to emotion-focused coping strategies (acceptance, seeking emotional support, humor, positive reframing, religion) was negatively related to perceived stress, whereas the dysfunctional one was positively linked to stress. The first association we identified, according to which emotion-focused coping strategies are linked to lower level of perceived stress, thus functioning as a protective factor against negative psychological outcomes, is consistent with the results of many other studies on the general population during COVID-19 [106], as well as on HCWs before COVID-19 [107–109] and during it [110,111]. Similarly, also the second association we identified, according to which, on the contrary, dysfunctional coping strategies (behavioral disengagement, denial, self-blame, self-distraction, substance use) are linked to higher levels of perceived stress, is consistent with the results of many other studies [79,100,110], as well as in line with our expectations. Furthermore, consistently with the results of other works on Italian HCWs employed in facing COVID-19 during the first months of its spread [101,112–114], we did not find positive associations between problem-focused coping strategies and stress reduction in both

groups of HCWs. Analogously to the results of these studies [101,112–114], our findings revealed that, problem-focused coping strategies were not effective in the reduction of HCWs' perceived stress during the first wave of the pandemic (that was the period the respondents of our questionnaire referred to) probably due to the "lack of scientific knowledge about the therapeutic and treatment procedures effective for COVID-19" [114] (p. 3). In other words, the insufficient knowledge and the wide-spread uncertainty about the effective procedures to apply in order to prevent the spread of the virus and to treat infected people, seem to have made it difficult for health professionals to resort to problem-focused coping as strategies for stress reduction.

There are some limitations to the current research that should be considered. The first one concerns methodology: the research used self-report measures (although exclusively validated scales has been utilized, they can lead to potential bias related to social desirability), involved a non-probabilistic sample and it was a cross-sectional study. Moreover, although information has been collected on doctors and nurses, the level of severity of the patients with whom the participants had been in contact has not been specified. Furthermore, other significant variables such as years of service (seniority), age, medical specializations, gender, etc., have not been taken into account. Finally, doctors and nurses were not randomly assigned to workplaces. In addition, we did not evaluate the role played by reasoning processes, i.e., by cognitive strategies used by HCWs to reach decisions. While non-frontline HCWs (who generally did not have to deal with virus-related emergencies) probably have had more time to process information, to evaluate possible alternatives, and to make decisions, frontline HCWs more likely have had less time to think, to consider alternatives, and to assume decisions. This may have led the frontline HCWs to resort more frequently to shortcuts in thinking (heuristics) [115,116], which may have had some influence on the associations between the variables we examined, perhaps even inhibiting or reducing the strength of dispositional traits and the associations between them. Nonetheless, since we did not take into consideration reasoning processes, their possible impact remains a supposition, which deserves to be explored in future research.

Future studies might also take into account how the socio-demographic and workrelated variables impact stress, as well as other psychological outcomes, among HCWs. Furthermore, it would also be interesting to investigate HCWs' point of views, i.e., to analyze not only their responses to the items of validated scales (using quantitative methods), but also to analyze (using qualitative methods), their open-ended responses and/or interviews. Regarding this last point, our research team is qualitatively analyzing a sample of responses given to the last open-ended question of our questionnaire, which aimed to know if and how the experience of working during the pandemic had an emotional impact on HCWs (see Section 2.1). It would have been interesting also to repeat the survey after the second wave of the virus outbreak to test if new knowledge concerning the virus spread and its cure had influenced the use of coping strategies and the levels of perceived stress among Italian HCWs, and if so, how.

Despite the limitations, the results of the current study might be useful for planning and adopting preventing approaches to reduce HCWs' stress burden during a health emergency. The inability to tolerate uncertainty or the use of dysfunctional coping strategies, in fact, not only lead to negative outcomes for HCWs, but may also have an impact on patients and healthcare systems. The planning of training courses aimed to provide HCWs with skillsets they can use to cope with uncertain and stressful situations (such as that related to the COVID-19 pandemic) might be effective not only in reducing and controlling perceived stress (thus improving their mental wellbeing), but also in improving HCWs' effectiveness, and, thus have positive impacts on patients' health and on reducing costs for healthcare systems. Effective interventions should be designed to fit the specific traits of HCWs at the forefront. Health professionals who are better equipped (in psychological terms) to cope with uncertain and stressful situations would undoubtedly lead to improve the quality of care.

**Author Contributions:** Conceptualization, R.B. (Ramona Bongelli), C.C., A.F., M.M., I.R., A.B. and R.B. (Roberto Burro); methodology, R.B. (Roberto Burro); formal analysis, R.B. (Ramona Bongelli), A.F. and R.B. (Roberto Burro); data curation, R.B. (Ramona Bongelli), C.C., A.F., M.M., I.R., A.B. and R.B. (Roberto Burro); writing—original draft preparation, R.B. (Ramona Bongelli); writing—review and editing, R.B. (Ramona Bongelli), C.C., A.F., M.M., I.R., A.B. and R.B. (Roberto Burro); visualization, R.B. (Roberto Burro); supervision, R.B. (Ramona Bongelli). All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by PhD meeting curriculum in Psychology, Communication, and Social Sciences, (University of Macerata. Protocol code n. 19435, 3 August 2020).

**Informed Consent Statement:** Informed consent was obtained from all the participants involved in the study.

**Data Availability Statement:** All relevant data presented in the study are included in the article. The datasets analyzed are available from the corresponding author on reasonable request.

**Acknowledgments:** We would like to thank professional orders and associations, as well as each doctor and nurse who compiled the questionnaire.

**Conflicts of Interest:** The authors declare no conflict of interest.
