Beginning in late February 2020 Italy has been the first European country to face the COVID-19 pandemic. Despite evidence of a relevant impact of the lockdown measures on mental health in the general population (GP) (Rossi et al., 2020b), healthcare workers (HCW) were exposed to a number of additional stressful events while working under extreme pressure with COVID-19 patients, and thus represent a highly at-risk population (Rossi et al., 2020a).

Challenges for staff include the increased workload and physical exhaustion due to the severe condition of the patients, witnessing a higher-than-usual death’s rate among their patients, fears of contagion for themselves and their families and seeing colleagues falling ill or dying (Walton et al., 2020; Zhang Y. et al., 2020).

Indeed, in the very early stages of the pandemic, the Italian national healthcare service and its regional articulations were subject to a never seen before pressure, with most intensive care units (ICU) running short of beds in a few days. Furthermore, lack of preparation for such a pandemic resulted in lack of security protocols and protection devices for HCW, which resulted in a tremendously high number of infected and deceased personnel.

Italian media stressed the war-like scenario in which ICU were working, allegedly performing triage with physician having to cherry-pick which patient to provide care to.

The psychological impact of this unprecedented health emergency might have significant long-term reverberations. Also, addressing the exact consequences of the COVID-19 pandemic on mental health of HCW is additionally critical (Firew et al., 2020), as mental health issues may hinder working ability of medical staff. For this reason, supportive interventions for HCW are necessary at this stage.

Despite the huge number of publications on the mental health burden in HCW, very few data have been published so far.

Recent reviews and original investigations confirm a high rate of anxious and depressive symptoms, as well as poor sleep quality and post-traumatic symptoms (Johnson et al., 2020), among HCW (Chew et al., 2020; Pappa et al., 2020; Talevi et al., 2020; Vindegaard and Benros, 2020).

Preliminary data from China during the COVID-19 pandemic showed a depression rate 50.3%, anxiety 44.6%, and insomnia 34.0% (Lai et al., 2020), although another works from China report lower rates of anxiety and depression in medical HCW (Huang and Zhao, 2020; Zhang W. et al., 2020). Kang et al. (2020) found that as much as 36% of medical staff reported subthreshold psychological symptoms.

In this scenario, we reported preliminary data on the very immediate burden of the COVID-19 outbreak on mental health on 1300 Italian HCW, finding that frontline young women, regardless of the working position (i.e., nurse, physician, healthcare assistant (HCA), etc.), had higher odds of several mental health outcomes, including PTSD symptoms, anxiety, depression and insomnia (Rossi et al., 2020a). We identified a number of job-related risk factors, including having a colleague infected, hospitalized, or deceased by COVID-19. Working directly with COVID-19 patients, i.e., being a Frontline HCW, was a major risk factor for all of the selected outcomes.

However, our preliminary data left some unsolved questions. Firstly, the odds of negative mental health outcomes in HCW compared to GP remains to be elucidated. Secondly, considering potential differences in the degree of exposure to a number of COVID-19 related traumatic events in HCW and GP, a more in-depth analysis of post-traumatic symptoms warrants further consideration.

In this article, we report on the mental health outcomes of a sample of Italian HCW and a GP sample during the peak of the critical infection of the COVID-19 outbreak. Preliminary data from a part of this sample were previously published elsewhere (Rossi et al., 2020a,b). Results confirmed high rates of depression symptoms, anxiety symptoms, insomnia, and PTSS both in the GP and HCW group during the acute phase of the COVID-19 emergency. Further, we aimed at addressing, for the first time, mental health outcomes in front and second-line HCW and the GP together, allowing a comparison among the three groups.

In this respect, we firstly compared mental health outcomes among FHCW, SHCW, and the GP groups, finding that being an SHCW was associated with lower odds of endorsing anxious or depressive symptoms compared to the GP, while being a front and second-line HCW was associated with higher PTSS compared to the general population. GPS-Negative Affect symptoms were associated with being FHCW compared to the GP in the adjusted model.

Our data suggest a complex pattern of distribution of mental health outcomes among the three groups.

SHCW have a more positive outcome profile compared to both GP and FHCW, except for PTSS. This result was kept after controlling for education level, age, and gender. In order to explain this result, based on previous evidence (Gan et al., 2020; Rossi et al., 2020b), we assume that much of the impact on depressive and anxious symptoms in the GP could be due to lockdown measures, social distancing, and economic instability in addition to traumatic experiences. SHCW could have been less exposed to lockdown measures, but not to social distancing, compared to the GP because they continued to work without being put under excessive pressure, and this could be a reason for which they were somewhat protected against depressive symptoms compared to the GP. On the other hand, SHCW may have been exposed to traumatic events at their workplace (especially indirect traumatic events, such as knowing of colleagues infected or deceased), hence showing an increase in PTSS symptoms compared to the GP.

However, this interpretation should be taken with caution, as the pattern of exposure to traumatic events, lockdown measures, and social distancing was not actually captured by our or others’ data, rather it relies on lay evidence of how HCW and GP working pattern changed during the assessed time period.

FHCW didn’t show an increase in depressive or anxious symptoms compared to the GP, while they showed a relevant increase in trauma-related symptoms. Furthermore, compared to SHCW, they showed higher levels of negative affective post-traumatic symptoms, suggesting that working as a front-line HCW is associated with a complex pattern of traumatic exposure, that could include physical and mental exhaustion, witnessing a high number of deaths of patients and colleagues and fear of contagion.

Secondly, we explored COVID-related risk factors for mental health outcomes in the HCW sub-sample only, finding that being an FHCW was associated with higher odds of endorsing depressive and post-traumatic symptoms compared to SHCW. The COVID-related risk factors explored were specifically associated with trauma-related symptoms such as PTSS and negative affect symptoms. In particular, colleagues’ negative events, i.e., being infected, hospitalized, or deceased, were all associated with PTSS and trauma-related negative affective symptoms.

Taken together, these results suggest that, although the COVID-19 pandemic has had a relevant impact on the general population’s mental health as a whole (Rossi et al., 2020b), HCWs are a population at heightened risk specifically for trauma-related symptoms.

Regarding putative risk factors in the HCW group, contrary to early data from Chinese HCW (Lai et al., 2020), in our sample, no specific working position was associated with higher odds of mental health outcomes, except for nurses and healthcare assistants having higher odds of insomnia. However, in line with previous data on both the GP and HCW (Liu et al., 2020; Qiu et al., 2020; Wang et al., 2020), being female was associated with all the mental health outcomes considered, suggesting that female gender represents a risk factor for mental health issues such as PTSS, Depression, Anxiety symptoms, and Insomnia in the context of the current emergency. Also, similarly to previous reports (Qiu et al., 2020; Wang et al., 2020), younger age was associated with all the selected mental health variables, except for Insomnia. These findings therefore further encourage the implementation of targeted interventions for different at-risk populations.

This study has a number of limitations, mainly due to the on-line sampling strategy and cross-sectional design. Firstly, a self-selection bias, which is frequent in web-based surveys, could have led to an overestimation of effect sizes. Moreover, it is possible that this effect was different in the HCW and GP subsamples, leading to a biased estimate of the group effect on the selected outcomes. Secondly, it was not possible to assess how many subjects were reached by the questionnaire, so a response rate could not be estimated. A different sampling strategy, based on mailing lists of medical associations could have yielded a more accurate sample, however, getting access to mailing lists owned by Local Health Authorities could have introduced a relevant delay in sampling, eventually causing us to miss the relevant timeframe for this study.

Thirdly, this study is based on self-report measures that inherently convey a systematic bias in estimated the target construct.

Lastly, this study is based on a cross-sectional design. Although follow-up data will be collected in the future, no baseline data on the same participants were available at the time of the recruitment, and the only epidemiological study available in Italy so far (Girolamo et al., 2006) dates back to 2006 and is based on very different data collection instruments, hampering any possible comparison with our data.

However, this study has several strengths as well, consisting in its large sample size, and the prompt data collection, that was conducted during the highest peak of contagions of COVID-19 and burden on the national health service.

The raw data supporting the conclusions of this article will be made available by the authors.

The studies involving human participants were reviewed and approved by the University of L’Aquila IRB. The patients/participants provided their written informed consent to participate in this study.

RR, VS, FP, and GDL: conceptualization. RR: methodology and formal analysis. RR, SM, and GDL: data curation. RR and VS: writing–original draft. RR, VS, AD, FP, SM, AS, and GDL: writing–review and editing. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.