Article Text

Download PDFPDF

Original research
COVID-19 and healthcare workers: a rapid systematic review into risks and preventive measures
  1. J Valérie Gross,
  2. Judith Mohren,
  3. Thomas C Erren
  1. Institute and Policlinic for Occupational Medicine, Environmental Medicine and Prevention Research, University Hospital Cologne, Koln, Nordrhein-Westfalen, Germany
  1. Correspondence to Dr J Valérie Gross; juliane.gross{at}


Objective The COVID-19 pandemic is demanding for occupational medicine and for public health. As healthcare workers (HCWs) fight impacts of SARS-CoV-2 on front lines, we must create safe work environments through comprehensive risk assessments, evaluation and effective implementation of counter-measures. We ask: ‘What does current literature report on health risks at workplaces regarding COVID-19?’ and ‘What do current studies report on the effectiveness of enacted preventative recommendations?’

Methods As a snapshot of early HCW research, on 26 April 2020, we conducted a rapid systematic literature search in three databases (PubMed, Web of Science and PsycInfo) for COVID-19-related health outcomes and preventive measures in healthcare-associated workplaces.

Results 27 studies were identified as relevant for exploring the risk of infection, 11 studies evaluated preventive measures. The studies described that SARS-CoV-2 impacts significantly on HCW’s health and well-being, not only through infections (n=6), but also from a mental health perspective (n=16). 4 studies reported indirect risks such as skin injuries, one study described headaches to result from the use of personal protective equipment. Few studies provided information on the effectiveness of prevention strategies. Overall, most studies on health risks as well as on the effectiveness of preventive measures were of a moderate-to-low quality; this was mainly due to limitations in study design, imprecise exposure and outcome assessments.

Conclusions Due to widespread exposure of HCW to SARS-CoV-2, workplaces in healthcare must be as safe as possible. Information from HCW can provide valuable insights into how infections spread, into direct and indirect health effects and into how effectively counter-measures mitigate adverse health outcomes. However, available research disallows to judge which counter-measure(s) of a current ‘mix’ should be prioritised for HCW. To arrive at evidence-based cost-effective prevention strategies, more well-conceived studies on the effectiveness of counter-measures are needed.

  • public health
  • epidemiology
  • occupational & industrial medicine

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Strengths and limitations of this study

  • Empirical information for timely use and decision-making was assessed in an early phase of the COVID-19 pandemic and synthesised in a ‘real-time’ manner.

  • The review refers to workplaces of high relevance for individuals and public health.

  • We developed recommendations for further epidemiological studies that could aid understanding and controlling the pandemic.

  • The literature focuses on the first phase (until 26 April 2020) of the pandemic.


Challenges due to the SARS-CoV-2/COVID-19 pandemic are unprecedented in scope and scale. After massive ‘lockdowns’, numerous countries are choosing careful ways back to former life and work. To support this, occupational medicine can play a pivotal role:

Getting back to work is highly relevant for economic prosperity, individual well-being and wealth. Equally clearly, workplaces with countless human contacts bear the risk for case clusters. Therefore, expertise and insights from occupational medicine should be applied to create workplaces that are as safe as possible and to mitigate adverse health effects caused by a recession.1

Healthcare workers (HCWs)—as important actors of the critical infrastructure—are under considerable risk of infection and other diseases. Within the first months of the pandemic (until early March), 3300 HCWs in China were described as infected, in Italy about 20% of responding HCW.2 Because of the risks for HCW, in our review, we ask ‘What is reported regarding health risks at workplaces concerning COVID-19?’ and ‘What do current studies report on the effectiveness of enacted preventative measures?’. We further ask and discuss ‘What can we recommend from completed COVID-19 research on workplaces for ongoing and future epidemiology?’


On 26 April 2020, we conducted a rapid systematic literature search using three scientific databases (PubMed, Web of Science and PsycInfo). For PubMed we used the following search term: ((“Workforce”[Mesh] OR worker* OR employee* OR staff OR workforce OR labourer OR “Occupational Medicine”[Mesh] OR “Occupational Health” [Mesh] OR “Occupational Diseases”[Mesh] OR “Workplace”[Mesh] OR “occupational medicine“ OR “occupational health” OR workplace OR “occupational disease” OR “occupation“) AND (“COVID-19” OR “COVID-19” OR “severe acute respiratory syndrome coronavirus 2” OR “novel coronavirus” OR “new coronavirus” OR “COVID-19 pandemic” OR “wuhan coronavirus” OR SARS-CoV-2 OR “2019-nCoV disease”)). This search term was adapted for the other databases.

Inclusion and exclusion criteria

We screened identified relevant studies according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (see figure 1). We included publications focusing on COVID-19-related physical and psychological health outcomes in HCW or studies that investigated prevention measures regarding the risk through COVID-19. Studies focusing on other occupations, not presenting original data, focusing on coronaviruses other than SARS-CoV-2 or on diagnostics, therapies only were excluded. We further excluded articles exploring virus characteristics or transmissibility and those not published in English or German.

Figure 1

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flowchart of the included studies.

Quality appraisal

We critically appraised all original studies fulfilling the review’s criteria. To assess quality of quantitative studies, we used the Appraisal Tool for Cross-Sectional Studies (AXIS), the Critical Appraisal Skilss Programm checklist (CASP) for cohort studies, the Joana Briggs Institute (JBI) Critical Appraisal Checklist for case reports and case series and GRADE (Grading of Recommendations, Assessment, Development and Evaluations) for ecological studies. Qualitative studies were appraised by the CASP qualitative checklist. To identify the appropriate checklist, the study designs were evaluated in regards to our study questions. As several studies included various approaches to assess data in addition to investigating staff, our identification process may result in different study designs than stated by the authors. Where possible, overall scores were calculated or categorical scores were presented (see tables 1 and 2).

Table 1

Studies included for risk of infection, mental health, skin injuries and headaches

Table 2

Studies included for preventive measures

Patient and public involvement

No patient involved.


Our literature search identified 820 publications; after removing duplicates, 689 were screened for title and abstract. According to our research questions, JM reviewed studies on workplace-associated health outcomes associated with SARS-CoV-2. JVG reviewed studies on implemented prevention measures in healthcare settings and their implications for HCW’s health (see PRISMA flowchart, figure 1).

Health risks at healthcare workplaces

Of 27 included studies, 6 provided information on the risk of infection,3–8 16 on mental health,9–24 4 on skin injuries associated with PPE25–28 and 1 on resulting headaches29 (see tables 1 and 2).


Due to frequent contact with infected people, HCW’s own infection and the virus transmission to others is an important concern. Six3–8 out of twenty-seven included studies on health outcomes reported on the prevalence of SARS-CoV-2 infection as well as on COVID-19-related morbidity and mortality among HCW. In all six studies, exposure was assessed based on known contact with patients with SARS-CoV-2 positive and infection was diagnosed through laboratory tests (reverse transcription PCR).

According to Zhan et al who refers to statistics of the National Health Commission of People’s Republic of China, as of 24 February, 4.4% of the infected people in China (3387 of 77,262 patients with COVID-19) were HCW.8 A cross-sectional study conducted by the US Centers for Disease Control and Prevention (CDC) revealed that as of 9 April, HCW accounted for 11% of all reported cases in the USA.3

Two studies provided information on hospitalisation rates.3 6 For instance, in a long-term care facility in the USA, McMichael et al observed that 6% of 50 infected nursing staff were hospitalised,6 which is similar to a prevalence of 8% stated by the CDC.3 More specifically, in a case series Chu et al observed more cases of severe and critical illness among younger HCW than in older colleagues who were hospitalised due to COVID-19.4

Two studies provided information on HCW’s mortality.5 8 According to Zhan et al as of 3 April, 23 out of 3387 (~0.7%) COVID-19-positive HCW in China had died from the infection.8 Additionally, in their retrospective fatality analysis Li et al reported that as of 16 March, 13 out of 24 fatalities among HCW (54.2 %) were attributable to COVID-19 and most cases were recorded in Hubei province or Wuhan.5

Among the reasons for HCW’s infection, shortage of PPE and insufficient knowledge and training in infection control measures were reported to be most relevant.4 6 Chu et al pointed out that normal wards evinced a disproportionally high number of infected HCW compared with first-line departments (3.7% from first-line departments vs 72.2% from non-first-line departments) that were prioritised over second-line departments in terms of PPE allocation.4 Moreover, McMichael et al reported on disease spreading in a US-American long-term care facility, where staff worked while being symptomatic and moved between several facilities indicating lacking knowledge about infection control and containment measures.6

The high importance of protective measures was also emphasised in a retrospective cohort study by Ran et al who found that unqualified hand washing, suboptimal hand hygiene before and after patient contact and improper PPE increased the risk of infection by 2.64 (95% CI: 1.04 to 6.71), 3.10 (95% CI: 1.43 to 6.73), 2.43 (95% CI: 1.34 to 4.39) and 2.82 (95% CI: 1.11 to 7.18), respectively.7 The study also found that longer working hours, that is, longer exposure to SARS-CoV-2, was positively correlated with a higher proportion of infection.7

Mental health

Sixteen9–24 out of the twenty-seven included studies reported on mental health problems, four investigated PPE-related skin injuries25–28 and one reported on headaches associated with frequent PPE use.29

Most studies found that the exposure to SARS-CoV-2 was positively correlated with mental health problems.10 12–14 16 17 21 23 24 Depression, anxiety and stress were the most commonly assessed disorders showing a significantly higher prevalence in HCW in high-risk departments or high-risk regions (Wuhan, Hubei province) compared with medical staff in low-risk departments, non-clinical staff or staff in low-risk regions.10 11 13 14 16 17 19 23 24

Factors contributing to psychical exhaustion and development of mental illness were concerns over self-infection and infection of relatives due to insufficient protection, worry about patients, helplessness, unfamiliarity with infection control measures, overworking and traumatising events at work, to name but a few.9 15 17 19

More specifically, Kang et al indicated that the severity of psychological disturbances increased with the frequency of contacts to patients with COVID-19.13 A study by Huang et al compared different occupational groups during the pandemic and revealed that HCWs were more likely to develop sleep disturbances than other workers whereas no difference in prevalence of depression or anxiety was found.12

In contrast, Tan et al reported higher scores of anxiety and stress among non-medical workers (allied HCW, technicians etc) compared with medical workers.20 Moreover, a similar direction of effect was found by Li et al indicating that non-front-line nurses without any patient contact had significantly higher vicarious traumatisation scores than front-line nurses.15

Furthermore, a qualitative interview study by Sun et al showed that attitudes and feelings changed over time.19 They observed that front-line nurses experienced a high degree of helplessness and anxiety at the beginning of the pandemic, but could transform their negative emotions into feelings of higher professional identity and responsibility and personal growth. They further indicated that they felt grateful for the support from colleagues and family members.19

The importance of social support was demonstrated by Xiao et al.22 The authors highlighted a potentially mitigating effect on anxiety and stress and showed that social support improved HCW’s self-efficacy.22 Likewise, nurses in Sun et al.’s study emphasised the importance of support by their families and colleagues.19 Moreover, Mo et al identified that social characteristics such as being the only child could have a predisposing effect for the development of stress.18

Furthermore, underlying organic diseases were associated with developing mental health disorders including insomnia, anxiety, depression, obsessive–compulsive disorder, somatisation, stress and post-traumatic stress disorder (PTSD).11 24

Health effects through prevention measures

In a cross-sectional study, Ong et al explored the prevalence of de novo PPE-related headaches among medical workers from different high-risk areas of a tertiary health centre in Singapore.29 Data from a self-administered questionnaire revealed that 81% of the 158 respondents suffered from de-novo headaches when wearing N95 respirators or eye protection. The chance of developing PPE-related headaches increased significantly for those with pre-existing headaches (OR 4.20; 95% CI: 1.48 to 15.4) and for those who used PPE for more than 4 hours per day (OR 3.91; 95% CI: 1.35 to 11.31).29

Moreover, PPE-related skin damages were evaluated in four studies.25–28 Lin et al showed a positive correlation between self-perceived adverse skin reactions (such as dryness, erythema, maceration mostly on hands, cheeks and the nasal bridge) and the duration of wearing PPE (>6 hour).28 The prevalence of skin irritation was reported to be 74.5%.28 Furthermore, Gheisari et al pointed out the most common adverse skin effects (dermatoses, acne, skin irritation, frictional and pressure erosions) associated with goggles and respirator masks.25 Lan et al found that 97% of HCW suffered from some sort of skin damages, especially dryness and desquamation on the forehead and nasal bridge as a result of enhanced hygienic and infection control measures.27 Wearing N95 respirators for a period longer than 6 hour doubled the risk for facial skin lesions (OR 2.02; 95% CI: 1.35 to 3.01). The authors further observed that frequent hand hygiene (>10 times per day) may increase the risk of skin damage twofold, (OR 2.17; 95% CI: 1.38 to 3.43).27 Similarly, a large multicentre cross-sectional study performed by Jiang et al found an overall prevalence of skin injuries of 42.8%, involving pressure-related and moisture-related skin issues as well as skin tear. The prevalence of skin injuries was found to be higher when daily wearing time exceeded 4 hours and when high-grade PPE (ie, PPE3) was used.26

Effectiveness of enacted preventive measures

Only nine studies30–38 focused on evaluating the effectiveness of implemented preventive strategies regarding risk of infection, all of which were case reports or case series (see table 2). The studies observed operative procedures34 37 as well as general care and diagnostics.30–33 Infection was assessed by laboratory testing or clinical diagnosis of COVID-19.30 The number of investigated patients ranged from 132 33 37 to 308.30

Most of the studies reported zero staff infections,30–36 in one study four nurses who contacted a patient without protection (before quarantine) became infected while there was no transmission for staff members using PPE (after the patient’s confirmed infection).37 Malik et al suggested that surgical masks may be as effective as N95 masks, but further studies were needed to confirm this.33 However, in most studies the enacted preventive measures were complex and do not allow deducing the effectiveness of single measures.

Kabesch et al described how rigorous testing for SARS-CoV-2 and targeted self-isolating of HCWs in addition to improved hygienic measures and consequent use of surgical face masks could stop a COVID-19 outbreak in a large maternity and perinatal centre in Germany. While maintaining the clinical services, the virus was stopped within 3 weeks38.

Measures concerning mental health were evaluated in two studies.9 39 Concerning the effectiveness of supporting mental health, in qualitative interviews (n=13) in a tertiary general hospital in Beijing, communication with colleagues, family members or psychologists was described as being most helpful for most workers.9 Monitoring HCW in a Chinese hospital who received psychological support revealed that they would particularly appreciate designated resting areas and sufficient protection supply as well as the acquisition of necessary skills to manage patients’ anxiety.39

Quality assessment

Overall, most studies on health risks as well as on the effectiveness of preventive measures were of a moderate-to-low quality mainly due to limitations in study design, imprecise exposure and outcome assessments. Hence, selection, misclassification and information bias need to be taken into account when interpreting the findings (see tables 1 and 2 for the studies’ quality scores).


Currently, and presumably in the near future, we will have to cope with enormous challenges due to COVID-19. HCW at the front lines are under particular risk. In times when numerous preventative measures are taken for HCW, our review asks two questions, deliberately confining our evidence search to publications until 26 April 2020: ‘What is reported regarding health risks at workplaces concerning COVID-19?’ and ‘What do current studies report on the effectiveness of enacted preventative recommendations?’ We focus on a limited time period to identify effective evidence-based responses, on the one hand, or lack of studies into the effects of measures to protect HCW, on the other at an early stage of the pandemic. Answering our study question in ‘real time’ is important as we need to know—and apply—‘what’ does protect ‘whom’ effectively or otherwise instigate much-needed studies seeking such evidence.

Safety at healthcare workplaces

Among workplace-associated risks through SARS-CoV-2, virus transmission seems to be the biggest concern. However, robust data on incidence and prevalence are difficult to obtain due to uncertainty around infection pathways, inconsistent testing modalities within and between countries and varying disease manifestations including the unknown proportions of asymptomatic cases.

In most studies, exposure assessment referred to HCWs’ contact to patients with COVID-19 without specifying the exact contact time and duration. Yet, this is important to consider, as we know that viral shedding and infectivity differ over the course of disease.40 Moreover, a considerable proportion of included studies did not perform a thorough exposure assessment risking biased study findings on infection-related health consequences.

Another aspect to consider is virus transmission through HCW from outside healthcare facilities, which makes containment measures difficult and increases the risk for nosocomial infection.41 Of note, as protective equipment is usually prioritised to high-risk areas (fever clinics, emergency departments), undersupplied second-line departments could become breeding grounds for new infections.4

Beyond infections, the pandemic’s impact on mental health can be significant. Apart from the fear of self-infection and infecting others, high work pressure, changing relationships between HCW and patients, and possible stigmatisation in society42 43 may have detrimental effects on HCW’s psychological stability. While front-line HCW seem to be at highest risk, impacts on second line and non-clinical staff should not be underestimated. Since front-line HCW generally have better access to mental health support and protective and infection control measures, their psychological resilience is assumed to be higher than in second-line staff.15 20

To understand determinants of mental health disorders, qualitative studies can bring valuable insights. Since recent evidence is predominantly built on cross-sectional studies, it seems to be difficult to quantify the true effect of the ‘SARS-CoV-2 challenge’ on mental health, particularly with respect to pre-existing psychological and physical issues. For this review, we found great heterogeneity in self-administered outcome assessment (tables 1 and 2), which likely introduces information bias and complicates comparison of studies. Given both complexity and durability of mental health problems, longitudinal studies are warranted to explore HCWs’ mental and psychological health over time. Moreover, in future studies, mental health disorders such as PTSD may come on the study radar.

Our observations of COVID-19-related mental health outcomes were consistent with a recent systematic review and meta-analysis on the prevalence of depression, anxiety and insomnia by Pappa et al.44 Our work complements their findings as our broader scope of outcomes and study types adds valuable insight for occupational health research of HCWs in challenging times.

Overall, it becomes clear that healthcare workplaces pose multifaceted challenges to HCWs’ health, which require diligent investigations and a swift implementation of effective precautionary and supportive measures.

Effectiveness of enacted preventive measures

Since the COVID-19 pandemic could last for long(er) and effective medications, improving and developing appropriate protective measures are a must. While many publications describe preventive measures for general and specific healthcare settings, the imbalance regarding—or scarcity of 11—studies into the very effectiveness of a ‘mix’ of enacted measures is striking and needs to be corrected.

Indeed, the dynamic pandemic necessitates flexible adaptation of precautionary measures to take note of rapidly changing recommendations.45 With so many people under risk worldwide, one main objective must be to protect as many of those most under risk as possible. This does certainly not imply to use a maximum prevention in a ‘mix’ of available counter-measures. To the contrary, recommending highest standards of protection can be(come) overuse which has been described as a form of misuse.46 To exemplify, during the last months we all experienced that shortage of prevention measures, especially for PPE47 or laboratory tests,48 can occur. Therefore, when creating effective prevention measures, we also have to consider whether or how long they are available.

Given the obvious need of studies into which counter-measure(s)—alone or in combination—is/are effective, it is unfortunate that only very few studies focus on measuring effectiveness. Moreover, since most studies describe complex prevention concepts, effects of individual measures could hardly be investigated. Finally, the identified case and cross-sectional studies clearly complicate a generalisation of results. Moreover, as the virus is not specific for workplaces, studies should assess the risk of infection in private life, too. Unfortunately, non-workplace associated risks were not explicitly considered.

Cost–Benefit ratio

When discussing prevention measures, possible negative effects have to be considered and investigated. As described above, the PPE’s use can have negative impact on health and well-being. Additionally, the inappropriate use of preventive measures could be a danger49 resulting from a false sense of security due to unawareness which can have a negative impact on workers’ actions and the protective effect of instigated measures. Preferably, before implementing preventive strategies, the effectiveness of the specific measure (the benefit) has to be weighed up against possible risks and inconveniences (the costs).

Acceptance and appropriate use

When interpreting the epidemiological studies, it seems misleading to assume the effectivity of PPE tested under laboratory conditions to be equal to real workplace situations as the PPE’s effectivity depends on the acceptance and actual use by the individual worker. For instance, in a Chinese hospital the use of PPE was investigated, revealing that 100% of the staff used masks, with an accuracy rate of 73.79% only, but the compliance rate of hand hygiene of all staff was limited to 40.78%.50 Reasons for HCWs’ ability and willingness to follow guidelines regarding respiratory infections were investigated in a Cochrane review51: Insecurity on how to handle guidelines, frequently changing guidelines, the resulting additional workload and fatigue, little support, uncomfortable PPE and unwanted effects on the patients (for instance through covering parts of the face by a mask) were described as negative factors. On the other hand, a clear communication, training on how to use PPE and seeing the value of the recommended measures were deemed positive.

It is astonishing that even in hospital settings, PPE is often used inappropriately: In 2019, Herron et al found that only 18% of theatre staff (n=1034) used face masks according to the CDC standard.52 During the current pandemic, Wennmann et al described that HCW in an emergency department of a German University hospital required more PPE training than expected.53

Overall, ignoring the acceptance and appropriate use of preventive measures in epidemiological studies can result in severe information and misclassification bias. As inadequate knowledge is a risk for inappropriate PPE use,54 information and transparent communication is important to improve the acceptance and correct use of prevention measures. Moreover, exaggerated and untargeted measures might decrease workers’ acceptance and willingness to wear PPE as intended with far-reaching consequences for individual and public health.

Suggestions for ongoing research

What could be viewed as a limitation of this review, viz that we are still at an early stage of the pandemic and that our review covers a short period of time only, can be judged as a strength as well: Empirically, we identified lack of evidence for which counter-measure(s) of a current ‘mix’ provides the necessary protection of HCW and should be prioritised in what setting. Material regarding our review questions may now be considered to shape much-needed studies seeking such evidence. Clearly, (some) data for currently missing ‘effect studies’ may be collected now or never.55

Indeed, this review about HCW allows a first ‘real-time’ insight into current research and highlights challenges we face when investigating COVID-19. Due to high exposure to SARS-CoV-2, healthcare workplaces provide ideal conditions to study virus transmission, the effectiveness of prevention measures and findings may be transferred, at least partially, to other workplaces. Against the background of emerging disease clusters in other essential services, as recently observed in military service or in slaughterhouses, lessons learnt in the healthcare sector can be critical for occupational research to prevent further disruptions of everyday life.

However, as demonstrated in this review, high-quality studies from different parts of the world are lacking. Therefore, we do agree with Glasziou that the current phase of the pandemic has generated replicative studies of low methodological quality highlighting the need for well-defined study questions and robust study designs in order to reduce waste in COVID-19 research.56 Moreover, studies from different countries and economical settings are needed to avoid ‘geographical bias’. To achieve this end, multifaceted studies investigating the risk through COVID-19 at healthcare—and other—workplaces should include and consider:

  1. The risk of infection.

  2. The effectiveness of prevention measures.

  3. The availability of prevention measures.

  4. The risks of prevention measures.

  5. The cost–benefit ratio of prevention measures.

  6. The acceptance and appropriate use of prevention measures among workers.


COVID-19 poses severe risks for HCWs’ physical and mental health. While a plethora of prevention measures is currently enacted, we lack informative evaluations of their effectiveness and possible negative consequences. Taken together, available research disallows to judge which counter-measure(s) of a current ‘mix’ should be prioritised at healthcare workplaces. After reviewing hitherto scarce COVID-19-HCW-associated literature, recommendations for effect-based epidemiology to arrive at evidence-based cost-effective counter-measures should be considered when shaping ongoing and future research.



  • Contributors JVG was involved in creating and conducting the literature search and provided the first draft of this paper. JM was involved in creating and conducting the literature search and drafting the paper. TCE was involved in drafting the paper and supervision of this review.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement All data relevant to the study are included in the article. Studies included in this review are presented in tables 1 and 2.