In this study, we found that European hospital HCWs reported low levels of emotional wellbeing while providing hospital care and following local IPC recommendations during the COVID-19 pandemic, with particular concerns for those providing direct COVID-19 patient care, junior staff and women.

There is more evidence emerging on the psychological strain on HCWs working during the COVID-19 pandemic [7,24]. A recent systematic review and meta-analysis investigating the psychological impact of COVID-19 predominantly in Asia identified high levels of depression among medical staff [3]. Across Europe, we similarly identified psychological strain in hospital HCWs. Not only were there high levels of concern about HCWs’ own risk of COVID-19 infection, almost all respondents were worried about the risk to their families due to their job role. One-third of HCWs reported fear of looking after COVID-19 patients. Further, when quantifying emotional wellbeing using the WHO-5 Well-Being index, we found that more than one-third of HCWs had a score which was indicative of depression [23].

It is known that health emergencies can magnify already existing structural inequities [10]. Our study highlights impacts for women in particular, who comprise the majority of all HCWs worldwide. In general, women perform the majority of care within the home, are more likely to care for family members ill with COVID-19, and may face gender bias at work [10,25]. In our study, female HCWs reported higher levels of emotional strain compared to males, echoing findings from previous COVID-19 studies [9,26,27]. These results highlight the need for awareness of the differential psychological impacts on HCWs during the COVID-19 pandemic and future pandemics.

IPC recommendations aim to protect patients, HCWs, and the health system, and are therefore tied to risk perception as recommendations should ensure that HCWs feel protected during their work. Perceived lack of control is known to be an important stressor of mental health, and is associated with occupational burnout [28–30]. Research during the 2002-2003 SARS epidemic suggested that the combined role of three specific institutional measures impacted HCWs’ risk perception (including avoidance of patients and acceptance of risk) and belief in the effectiveness of protective measures at work [31]: clear policies and protocols, available specialists, and adequate training. In our study, junior HCWs in particular felt that getting infected with COVID-19 was outside of their control. This feeling itself was associated with a WHO-5 score indicative of depression. Risk perception is complex and multifactorial. A qualitative study among frontline HCWs who chose to respond to the West Africa Ebola virus disease outbreak (2014-2016), showed that next to individual and social-level factors, institutional trust was a key risk attenuator [32]. Our multivariable analysis identified organisational and personal factors that were positively associated with feelings of having control over becoming infected, including COVID-19 IPC training, confidence in the use of PPE as well as institutional trust. These aspects may therefore be important to consider in IPC implementation.

Overall, our survey suggests that institutions should adopt a multifaceted approach in IPC preparedness and training in order to best support hospital HCWs at work during an infectious diseases pandemic. Such an approach begins with recognition of the importance of preventing and mitigating adverse impacts on wellbeing, together with the explicit assurance to protect the clinical workforce. These foundations for protecting physical and mental health protection are made through ensuring safe working conditions and accessible support.

We echo recommendations that local IPC guidelines should be clearly communicated, and HCWs should receive specific IPC training, have access to appropriate PPE, and feel confident in its use [33]. Crucially, we add that health facilities must be aware of, and address differential impacts and experiences, such as those of female and junior HCWs. Given the already-heavy burden and time pressures on HCWs, effective support must therefore be readily accessible and responsive to the different needs, pressures, and barriers experienced by different groups. Together, these measures should work to increase HCWs’ feelings of self-efficacy to undertake IPC behaviors and, subsequently, their perceived safety.

Previous studies have also shown the importance of managers’ support and a safe workplace culture [33]. To identify where most improvements could be made, we further suggest that health facilities rapidly assess local perceptions of their HCWs on IPC recommendations and their wellbeing using the WHO data collection tool (freely available online) [16]. This would serve to identify specific key local needs that are/are not being met, enabling wellbeing support services to be better targeted and tailored, as research has shown that these are often poorly utilised by HCWs [9].

There are important limitations of the current study that should be acknowledged. First, in round 1 we chose to rapidly recruit HCWs from existing research networks where we could accurately describe our study base population, limiting the number of HCWs we could invite for participation. Along with data collection taking place during the peak of the pandemic in Europe and restricting to a single reminder because of high clinical workload of HCWs, this ultimately led to small sample size. Therefore, round 2 was performed using convenience sampling, leading to different sampling methods used in the current study. Still, we excluded HCWs that possibly participated twice, and therefore do not expect that this influenced our results. Second, inherent to our survey design there is a risk of response bias, both on the level of demographic characteristics as well as on personal views. Third, this was a cross-sectional observational study using quantitative data, we therefore cannot draw conclusions on the direction of effects. Lastly, we were unable to triangulate responses with actual local PPE supplies and hospital infection rates.