The search strategy for this pre-registered rapid review involved searching Medline, PsycInfo, Web of Science, and the first 10 pages of Google Scholar, as well as hand searching references. Search terms included a combination of terms related to health-care workers (e.g., “physicians,” “nurses”), and distress (e.g., “stress,” “anxiety”). The full search term list is available on PROSPERO (CRD42020178185). We conducted searches in a rolling manner, starting on April 6, 2020, then with updates on June 7, July 2, July 10, July 30, 2020, and November 15, 2020 to capture and integrate the most up-to-date evidence given the ongoing COVID-19 pandemic and the associated rapid release of research.

We used a predefined search strategy (see full details on PROSPERO, https://www.crd.york.ac.uk/PROSPERO/; registration ID: CRD42020178185). Studies were included in this Review if they were empirical research; published or accepted for publication in peer-reviewed journals; written in English; included participants who were HCW who worked in a hospital environment during a major infectious outbreak (COVID19, SARS, MERS, H1N1, H7N9, Ebola); had a sample size of >80, and included data on factors associated with psychological distress during an outbreak. One investigator screened citations for potential full-text review, and a second investigator conducted the full-text review of each study for inclusion. Exclusions were verified by the other investigator, and disagreements resolved through discussion. Data was extracted by one investigator, entered into a table, and verified by a second investigator. For studies that included tests for multiple measures of psychological distress, we included the study as reporting a significant association with a particular factor if at least one of the measures of distress were significant.

Although rapid reviews do not always include a formal assessment of study quality and risk for bias (18), a lack of a quality assessment can have important implications for the utility of the results (17). Accordingly, we evaluated the methodological quality of the studies in the review using a tool adapted for the current study. The assessment tool included eleven questions chosen from the Appraisal tool for Cross Sectional Studies, AXIS (19) as being most relevant for the current study, an approach advocated by Quintana (20). Two authors independently rated the quality of the studies using the 11 questions to assess the quality of the study procedures, sampling, and the measures. The assessment yielded a total score that categorized studies as having low (<5), moderate (5–7), or high (8–10) quality. Inter-rater agreement was calculated and assessed using Cohen's Kappa coefficient (21). Discrepancies were resolved through discussion. In addition to the formal quality assessment, we only included studies that reported findings for a sample size of >80, which allows enough power to detect a medium effect size with an alpha of 0.05 (21, 22).

We conceptually organized the factors in this Review identified as contributing to or mitigating psychological distress into three broad categories: (1) fixed or unchangeable factors (sociodemographic and occupational factors), (2) potentially modifiable factors (social and psychological factors), and (3) factors related to infection exposure. Evidence was synthesized according to these conceptual categories, with non-significant and contrary findings noted in addition to significant findings to provide a more complete picture of the weight of the evidence for each factor. The balance of evidence for each factor was further presented graphically. We assigned factors within each conceptual category as reflecting either risk or resilience for psychological distress according to logic and theory (e.g., maladaptive coping as risk, adaptive coping as resilience). Factors that could be interpreted as either risk or resilience (e.g., sex, age) were assigned according to how they had been framed in the majority of the research that examined these factors.

The quality of the studies ranged from moderate to high, with no studies rated as having low quality. The majority of the 139 studies were rated as having high quality (118; 84.9%), and 21 studies were rated as having a moderate quality (15.1%). Inter-rater agreement was high, 90.65% agreement, Cohen's Kappa = 0.642 (see Supplementary Table 2).

Seventy-two studies examined age as a predictor of psychological distress among HCW during an epidemic (see Table 2). Of these, 39 found that age was a significant risk factor for distress. In two studies of HCW during the SARS outbreak, staff who were younger than 33 experienced greater stress, but not greater psychiatric morbidity, compared to older staff (134), and staff under 35 were more likely to report severe depressive symptoms 3 years after the outbreak (92). In another study, medical staff who were between 20 and 30 years old and exposed to patients with H7N9 had elevated post-traumatic stress disorder scores compared to older staff (157). Similarly, general practitioners in working during the SARS outbreak who met psychiatric caseness for PTSD were more likely to be younger (144). In a study during the H1N1 outbreak, hospital staff who were in their 20's had greater anxiety about becoming infected than did older staff (103). During COVID-19, HCW who were younger were more likely to experience higher levels of post-traumatic stress symptoms, depression, anxiety, and acute stress compared to older HCW (23, 26–28, 30, 32, 42, 49, 54, 55, 57, 61, 65, 71, 75, 78, 89, 90, 117, 118, 121, 123, 127, 131, 149, 153, 154). In contrast, eight studies conducted during COVID found that HCW who were older were at greater risk of experiencing higher levels of psychological distress (40, 66, 86, 95, 102, 114, 122, 132). Lastly, 33 studies found that age was not a significant predictor of distress in HCW during the SARS, MERS or during the COVID-19 outbreaks (Table 2).

Ninety studies tested sex as a possible risk factor for distress among HCW during an outbreak (Table 2), with all but 33 finding that being female was associated with higher risk for psychological distress. Notably, the 57 studies that found that female sex was a significant risk factor spanned six different infectious diseases (MERS, SARS, COVID-19, H1N1, H7N9, and SARS), suggesting that being a female HCW increases vulnerability for distress more generally when working during an infectious outbreak. Notably, among the studies 30 studies that did not find that being female created significant risk for distress, eleven (36.6%) were conducted with nurses and included predominantly female participants (24, 43, 44, 59, 70, 79, 80, 89, 108, 140, 153).

Of the 69 studies that examined marital status as a risk or resilience factor for psychological distress, 19 found evidence to suggest this as a risk factor (Table 2). For example, two studies of HCW during the SARS outbreak found that HCW who were single were 1.4 times more likely to experience psychological distress than married HCW (41), and more likely to have sever depressive symptoms 3 years later (92). Similarly, HCW during the COVID-19 outbreak who were single experienced higher levels of distress than those who were married (54, 57, 66, 69, 111, 122, 126). Conversely, four studies conducted during COVID-19 found that being married was a risk factor for greater distress (66, 75, 89, 94), and two studies found that married HCW with children reported greater stress than single HCW or those who were married without children (72, 83). Forty-seven other studies conducted during the SARS, MERS, and COVID-19 outbreaks found no associations between HCW marital status and distress (Table 2).

Thirty-three studies examined education levels in association with distress. Only eight studies, six conducted during the COVID-19 pandemic (27, 66, 70, 89, 94, 149, 151), along with studies conducted during the Ebola outbreak (76), and the MERS outbreak (81) found that HCW with higher educational levels reported significantly lower psychological distress. Twenty-two studies found that education level was not predictive of psychological distress among HCW working during the MERS or the COVID-19 outbreaks (Table 2).

Thirty-four studies examined and found evidence that the HCW occupational role created risk for psychological distress while working during the SARS, H1N1, MERS, and COVID-19 outbreaks (Table 2). In all but 16 studies, being a nurse was associated with a range of mental health issues, including higher stress, burnout, anxiety, depression, PTSD symptoms, psychiatric morbidity, and psychological distress compared to being a physician or other HCW (see Tables 1, 2). The extent to which nurses experienced greater psychological distress whilst working during an outbreak was estimated in several studies. For example, nurses were 1.2 (83), 1.4 (124), 2.2 (63), and 2.8 (107) times more likely to be at risk for poor mental health. In contrast, five studies found that physicians (13, 97, 119, 128) and technicians (41) were more likely to experience distress while working during the COVID-19 pandemic and the SARS outbreak. Sixteen studies conducted during the COVID-19 outbreak did not find that occupational role was a risk factor for distress (Table 2).

Other occupational factors examined included years of work experience, and full-time vs. part-time status. Twelve of the 35 studies found evidence to suggest that less work experience may create risk (Table 2). HCW who had worked for <2 years experienced significantly greater stress than those with more work experience in a large sample of HCW during the SARS pandemic (13). In HCW during the SARS outbreak, those with <10 years of experience reported higher levels of psychological distress, but not burnout or post-traumatic stress, 13–26 months after the outbreak (14). HCW who had less clinical experience were also more likely to experience stress during the COVID-19 outbreak (23, 28, 55, 65, 69, 154). Years of clinical experience was not associated with PTSD symptoms, acute stress or anxiety, depression, mental health status, or burnout in 21 other studies (Table 2). Two studies found that less work experience was protective against distress for HCW during COVID-19 (121, 122). Lastly, in one study, part-time worker status was a significant predictor of greater emotional distress in HCW during the SARS outbreak (107), whereas another study found no evidence of part-time work status creating risk for distress in HCW during COVID-19 (119).

A number of social and interpersonal factors mitigated or contributed to psychological distress. Receiving direct social support from friends, family, colleagues and supervisors was a key protective factor in all of the 19 studies that examined its association with psychological distress (Table 2). For example, in HCW during the COVID-19 outbreak, higher levels of social support were associated with significantly lower levels of stress, depression, anxiety, depression and PTSD (28, 31, 38, 62, 70, 78, 88, 90, 94, 102, 139, 156). These findings were consistent with that of a study of frontline medical staff during the COVID-19 outbreak who reported that a positive attitude from co-workers was important for reducing their distress (39). Analogously, emergency nurses working during MERS outbreak who reported poor support from family and friends experienced higher levels of burnout (81). Similarly, studies of HCW during the SARS outbreak found that higher levels of family support were associated with lower depression and anxiety whereas inadequate support from relatives, lack of gratitude from patients and relatives, and perceiving less of a team spirit at work was associated with higher levels of psychological distress (44, 134).

Organizational support was an important factor in buffering psychological distress of HCW during an outbreak in all 11 studies that examined this factor. In nurses working during the SARS outbreak in Canada, higher perceived organizational support in the form of receiving positive performance feedback from doctors and co-workers, was associated with lower perceptions of SARS-related threat and reduced feelings of emotional exhaustion (59). Similarly, nurses, physicians, and HCW working during the MERS, COVID-19, and SARS outbreaks who perceived support from their supervisors and colleagues, experienced better mental health in the form of lower PTSD symptoms, lower distress, and being less likely to develop psychiatric symptoms, respectively (24, 28, 41, 54, 59, 70, 79, 80, 88, 99, 116).

Seven studies examined receiving useful information from others (a common form of social support). In one study, HCW who received adequate communication and information about the H1N1 outbreak from their organization were less likely to experience psychiatric symptoms because it helped them cope better, and worry less about the pandemic (63). Similarly, HCW during the SARS outbreak who had confidence in the information they received from their organization (130), and who received clear communication about directives and how to take precautionary measures (41), experienced reduced psychological distress. HCW working during the COVID-19 outbreak who felt that they did not receive sufficient information, scored significantly higher on anxiety and acute stress than those who were satisfied with the information provided (60, 99, 116, 142).

Negative social perceptions created risk for poor mental health for HCW in all 12 studies that examined this factor. In nurses during the MERS outbreak, perceived social stigma was associated with higher stress and poorer mental health (108). Similarly, during the COVID-19 pandemic, HCW who felt stigmatized, perceived stigma concerning negative public attitudes and disclosing about one's work, experienced higher levels of depression, anxiety, and psychological distress (55, 78, 102, 108, 109, 116, 138). During the SARS outbreak, HCW who felt people avoided their family because of their job were twice as likely to have elevated levels of post-traumatic stress symptoms (83). Importantly, experiencing stigma and avoidance from others was significantly associated with higher levels of post-traumatic stress symptoms during the SARS outbreak (104), and 13–26 months later (14).

The psychological factors examined in the studies included adaptive and maladaptive coping responses, beliefs and attitudes, and personality traits. Fourteen studies examined how perceptions of control were associated with distress among HCW (Table 2). In eight studies, higher self-efficacy was associated with lower anxiety, depression, distress, and lower levels of fear about SARS and post-traumatic stress symptoms during the COVID-19 and SARS outbreaks, respectively (10, 35, 68, 72, 90, 105, 123, 150). Conversely, feeling a loss of control was associated with greater distress (148) during the SARS outbreak in Hong Kong. Analogously, appraisals of personal risk were linked to higher levels of PTSD symptoms in HCW during the MERS (126) and SARS (130) outbreaks. Only one study conducted with nurses during COVID-19 did not find evidence that risk appraisals were linked to greater distress (70).

Positive attitudes toward one's work were protective against distress in all six studies that examined this factor. Higher work satisfaction was associated with less psychological distress among hospital staff during the H1N1 outbreak (63), lower PTSD among nurses (145), and lower rates of burnout among HCW during the COVID-19 outbreak. Similarly, HCW during the SARS outbreak who felt their work had become more important were less likely to develop psychiatric symptoms (41), and those who viewed their work altruistically were less likely to have severe symptoms of depression 3 years later (92). HCW who held a positive attitude toward their work reported less stress during the peak of the COVID-19 outbreak (31).

Seventeen studies examined whether coping styles were associated with HCW distress during an outbreak (Table 2). Emergency physicians and nurses working during the SARS outbreak who used denial, mental disengagement, or venting of emotions to cope were more likely to score higher on psychiatric morbidity (110). Similar results were found in frontline nurses during COVID-19, with use of negative coping associated with higher PTSD and psychological distress (102), and positive coping linked to lower PTSD (145). In HCW during the SARS outbreak, those who used maladaptive coping strategies, such as escape-avoidance, and self-blame coping, reported higher levels of burnout, psychological distress, and post-traumatic stress when surveyed 13–26 months after the outbreak (14). However, the use of adaptive strategies, such as problem-solving and positive reappraisal, were not associated with any of the distress outcomes. This finding was consistent with those from studies in which coping ability was not significantly associated with PTSD symptoms during the MERS outbreak (126), and problem-solving and turning to religion to cope were not associated with reduced distress during COVID-19 (31).

Twelve studies investigated the role of personality in HCW's psychological distress (Table 2). During the SARS outbreak, neuroticism was linked to poorer mental health (96), and HCW who had an anxious attachment style reported experiencing higher burnout, psychological distress, and post-traumatic stress 13–26 months after the outbreak (14). Those with an avoidant attachment style reported greater distress, but not burnout or post-traumatic stress. Eight studies examined the role of dispositional resilience. Among nurses working during the MERS outbreak, higher levels of hardiness were associated with lower stress and better mental health (108), and resilience was associated with lower anxiety, depression, post-traumatic stress symptoms, and burnout among frontline nurses and HCW during COVID-19 (38, 45, 72, 74, 89, 91, 153).

Fifty-three studies examined the impact of direct contact with infected patients on HCW's psychological distress. Of these, the majority (65) found that being in direct contact with and/or treating patients infected with COVID-19, SARS, MERS, or H7N9 was a risk factor for psychological distress (Table 2). Only eight studies did not find that contact with infected patients increased risk for distress in HCW during the COVID-19, SARS, and MERS outbreaks. Similarly, 24 studies found that risk of contact with infected patients due to working in high-risk areas (e.g., ICU, isolation areas and infection units) was associated with higher levels of anxiety, stress, and post-traumatic stress symptoms than not working in such areas (Table 2). Notably, one study found that HCW in a high-risk unit during SARS reported higher and sustained perceived stress 1 year after the outbreak compared to those in low-risk units, with those in low-risk units reporting a decrease in stress over time, but those in high-risk units experiencing an increase in stress post-outbreak (16). Three studies conducted during COVID-19 found that risk of contact was not associated with greater distress (53, 94, 132). Spending time in quarantine due to risk of being infected was associated with higher levels of burnout, depression, and psychological distress in HCW during SARS and COVID-19 (14, 59, 92, 132), but was unrelated to post-traumatic stress symptoms and psychological distress in HCW during the MERS outbreak or the COVID-19 outbreak (78, 85, 109, 125, 156). Lastly, one study found that HCW who had colleagues who became infected, had deceased due to infection, or had been quarantined, also experienced higher levels of post-traumatic stress symptoms and acute stress during the COVID-19 outbreak (118).

Provision of adequate training, protection, and other resources to manage and reduce risk of infection was associated with less psychological distress in all 19 studies that examined this factor (Table 2). Receiving clear infection control guidelines predicted lower psychological morbidity in frontline HCW during SARS (134), and having sufficient hospital resources for the treatment of MERS was associated with lower MERS-related burnout (81). After the implementation of a SARS protection training program, HCW experienced significant decreases in anxiety and depression 2 weeks and 1 month after the starting the program (44). Similarly, medical staff receiving inadequate training related to managing H7N9 had higher PTSD symptoms than those who received appropriate training (81). During COVID-19, HCW who felt HCW who felt that they did not have adequate information, training, personal protective equipment (PPE), felt unsafe, and perceived lower logistic support, reported higher levels of depression, anxiety, and acute stress symptoms (51, 54, 60, 65, 72, 74, 99, 100, 102, 106, 120, 142).

There are several limitations of this rapid systematic review. Conducting the review during the ongoing outbreak of COVID-19 imposed time constraints. This meant that we only included published peer-reviewed literature and did not search more thoroughly through gray literature or online pre-print repositories. Most study samples were quite large, increasing confidence in the generalisability of the findings.

In terms of the evidence base, the majority of the studies were cross-sectional, providing only a snapshot of the factors associated with HCW psychological distress. This limits conclusions about the direction of causality between the factors and distress, especially for those factors that are modifiable. Only three studies examined the potential long-term effects of the risk and resilience factors on HCW's mental health by using follow-up and time-lagged designs (14, 16, 92), providing some support for the assumed contribution of the factors to distress. More research needs to track the associations of risk/resilience factors over time with distress and the extent to which certain factors link to sustained or transient distress.

The majority of the studies were conducted during COVID-19, with relatively fewer studies reporting results from other infectious outbreaks such as SARS, MERS, H1N1, H7N9, and Ebola. On the one hand, this could be viewed as a limitation on the generalisability of the findings from the predominant outbreak, COVID-19, to other infectious outbreaks. On the other hand, we would argue that the consistency of the findings for a number of factors including participant sex, being a nurse, all 10 of the social and psychological factors, four of the five infection exposure factors, demonstrate that findings are likely to be generalizable across infectious outbreaks for these factors.

Although a number of studies investigated fixed factors and infection-related factors, relatively fewer studies examined how modifiable factors linked to distress (Figures 2, 3). There is a need for more research focusing on these factors to provide a more solid evidence base about potential targets for clinical intervention and treatment. A handful of studies used unvalidated measures of psychological distress, raising concerns about whether the findings would be the same had validated measures been used. For those studies that used validated measures, the ways in which cut-off scores for caseness were calculated, and/or the ways in classification of symptoms met thresholds for psychological distress, undoubtedly varied between measurement instruments. This likely introduced some variance into the results.

Few studies considered potential confounders in the associations with distress, compared found associations in matched non-HCW samples, or the extent to which the factors were predictive of distress outside of an outbreak. As well, the results extracted from the studies reflect a mix of bivariate and multivariate associations, as not all studies reported the bivariate only findings, which would be more comparable for making comparisons. Studies that examined the factors in multivariate analyses often used different covariates making it difficult to draw equitable conclusions from the studies. It is therefore difficult to assess the degree to which certain factors may independently predict psychological distress over and above other factors. Collectively, these limitations may have contributed to the equivocal findings noted for several of the factors reviewed.

Several strengths of the Review balance these limitations. Conceptually organizing the factors according to risk or resilience and whether they were fixed or modifiable, provided a theoretical framework for identifying who might be at most risk for psychological distress. This facilitates appropriate clinical intervention, and for noting which factors would be suitable targets for potential interventions. We also reported non-significant and contrary findings alongside significant findings to provide a more balanced and critical overview of the evidence. The Review included evidence from across six infectious disease outbreaks, with the majority of the research reporting findings from coronavirus outbreaks—Severe Acute Respiratory Syndrome Coronavirus-2 (COVID-19), Severe Acute Respiratory Syndrome-related coronavirus (SARS), and Middle East Respiratory Syndrome-related coronavirus (MERS)—that share similarities in their symptom and contagion profiles. Consistent evidence for risk and resilience factors was found across these various infectious diseases, suggesting that the findings from this review may be applicable across different outbreaks. This is relevant for understanding the mental health of HCW in future outbreaks. Lastly, conducting a series of search updates ensured integration of the most recent evidence from the ongoing COVID-19 outbreak into the review at the time of submission.

Whereas, other reviews have documented the extent of distress experienced by HCW during an outbreak (2), the current Review highlights the profiles of HCW most at risk for psychological distress and psychiatric morbidity during an outbreak. This identified modifiable factors that warrant further investigation as possible points of intervention to mitigate distress. Viewing risk and resilience factors from the lens of fixed and modifiable factors provides an efficient and useful approach for understanding who is most at risk and how to address that risk during and after an outbreak. Further research focusing on possible interactions among these factors would be useful to gain a better understanding of both the risk profiles and key modifiable factors, as the evidence reviewed did not consistently examine this area.

There is evidence that the psychological distress from working during an outbreak can persist for 2–3 years after the outbreak (14–16). Therefore, monitoring and providing appropriate support should continue beyond the outbreak period to ensure mental health recovery, especially among HCW who are most at risk. Our findings suggest that particular attention should be paid to female HCW and nurses (regardless of sex), and those who come into contact with infected patients or their environments to ensure that they receive necessary resources and provision of support to manage psychological distress. Proactive approaches at the organizational level can be effective (164) and may be necessary to help reduce the psychological distress of HCW. For example, a study of HCW during the COVID-19 outbreak in China found that mental health resources and services were mainly used by those experiencing mild and subthreshold levels of psychological distress rather than those who experienced more severe distress (11). Addressing the mental health needs of HCW with more severe distress will likely require more proactive means from health-care organizations.

There are a number of delivery methods to provide support and help HCW modify risk factors and foster resilience factors. These include telehealth, mobile apps, online toolkits, and peer-support, either in person or virtual (175). Combining different approaches may also be effective. For example, social support and perceived control can have an additive effect for reducing stress related to job demands (176). There is also evidence for the effectiveness of interventions for reducing HCW distress when delivered at the person level and organizational level (164), as well as those that target lifestyle practices (177, 178).

Evidence from randomized controlled trials suggests that third-wave cognitive behavioral therapeutic approaches, such as mindfulness (178), gratitude (177), and self-compassion (179), are effective for reducing stress and burnout among healthcare professionals, and could be beneficial. In low-resource settings, peer-support is one option that has been shown to be effective for reducing occupational distress in HCW (164). Raising awareness of the impact of an infectious outbreak on HCW mental health, providing appropriate treatment and therapy, and fostering proactive approaches such as an organizational culture of support (180), are recommended as possible approaches that can help prepare HCW for future outbreaks and address any persistent, long-term distress following the outbreak.