What are effective strategies to respond to the psychological impacts of working on the frontlines of a public health emergency?

Background The COVID-19 pandemic has disrupted the healthcare and public health sectors. The impact of working on the frontlines as a healthcare or public health professional has been well documented. Healthcare organizations must support the psychological and mental health of those responding to future public health emergencies. Objective This systematic review aims to identify effective interventions to support healthcare workers’ mental health and wellbeing during and following a public health emergency. Methods Eight scientific databases were searched from inception to 1 November 2022. Studies that described strategies to address the psychological impacts experienced by those responding to a public health emergency (i.e., a pandemic, epidemic, natural disaster, or mass casualty event) were eligible for inclusion. No limitations were placed based on study design, language, publication status, or publication date. Two reviewers independently screened studies, extracted data, and assessed methodological quality using the Joanna Briggs Institute critical appraisal tools. Discrepancies were resolved through discussion and a third reviewer when needed. Results were synthesized narratively due to the heterogeneity of populations and interventions. Outcomes were displayed graphically using harvest plots. Results A total of 20,018 records were screened, with 36 unique studies included in the review, 15 randomized controlled trials, and 21 quasi-experimental studies. Results indicate that psychotherapy, psychoeducation, and mind–body interventions may reduce symptoms of anxiety, burnout, depression, and Post Traumatic Stress Disorder, with the lowest risk of bias found among psychotherapy interventions. Psychoeducation appears most promising to increase resilience, with mind–body interventions having the most substantial evidence for increases in quality of life. Few organizational interventions were identified, with highly heterogeneous components. Conclusion Promoting healthcare workers’ mental health is essential at an individual and health system level. This review identifies several promising practices that could be used to support healthcare workers at risk of adverse mental health outcomes as they respond to future public health emergencies. Systematic review registration: https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=203810, identifier #CRD42020203810 (PROSPERO).


Introduction
The COVID-19 pandemic has been a significant disruption to health systems worldwide.Throughout the lifting and reinstatement of public health restrictions and waves of heightened pandemic activity, there was a growing recognition of a "shadow pandemic" of adverse mental health effects due to the pandemic (1).Healthcare workers (HCWs) in various settings are particularly at risk for significant and prolonged impacts given their frontline role in pandemic response.Previous research has demonstrated longstanding psychological consequences experienced by HCWs after SARS (2,3), and it is expected that there will be large-scale and longterm psychological impacts due to responding to COVID-19 among several occupational groups (4)(5)(6)(7)(8).Several systematic reviews of studies exploring the effects of the COVID-19 pandemic on the mental health of HCWs have reported a high prevalence of anxiety, depressive symptoms, insomnia, burnout, emotional exhaustion, and somatic symptoms (9, 10).Notably, factors such as the incidence of infection in the local setting, peer infections or deaths, and a shortage of personal protective equipment all increased the likelihood of experiencing adverse mental health outcomes (9, 10).At the same time, the availability of psychological support resources had a protective effect (9).With this knowledge, an action plan, support, and resources for individuals and organizations are needed to adequately support those experiencing or at risk for psychological distress or adverse mental health outcomes.
There is currently limited guidance on how best to support HCWs in response to and following public health emergencies, such as pandemics.Recent syntheses have gathered very low-to moderatecertainty evidence about interventions that may positively affect healthcare workers' mental health during a disease outbreak.However, these reviews have focused only on narrow categories of intervention, for example, online mindfulness programs (11), or limited populations, for example, those experiencing burnout (12).There is a need for a high-quality systematic review that captures a broad range of mental health-promoting interventions delivered during various public health emergencies to further inform strategies to improve mental health in HCWs.This review aims to answer the question: What are effective strategies to address the psychological impacts [i.e., anxiety, depression, and Post Traumatic Stress Disorder (PTSD)] experienced by HCWs responding to a public health emergency?

Methods
The protocol for this systematic review was registered (PROSPERO #CRD42020203810) and published a priori.There were no deviations from this protocol.The Cochrane Handbook for Systematic Reviews of Interventions guided our review protocol (13).

Search strategy
In collaboration with a health sciences librarian, a comprehensive search strategy was developed and executed in MEDLINE, PsycINFO, Embase, EMCARE, CINAHL, Sociological Abstracts, Business Source Premier, and the Cochrane Central Register of Controlled Trials from inception to 1 November 2022.Search terms related to (1) healthcare personnel; (2) pandemic, epidemic, COVID-19, natural disaster, or public health emergency; (3) psychological supports, interventions, and programs; and (4) mental health outcomes were used (see Supplementary Files, Figure 1 for complete MEDLINE search strategy).Reference lists of included studies and relevant systematic reviews were hand-searched, and relevant experts were contacted to identify any studies that were not captured via our search.Gray literature sources were searched until 22 January 2022, following the Canadian Agency for Drugs and Technologies in Health Grey Matters tool for searching health-related gray literature to identify unpublished studies (14).

Inclusion and exclusion criteria
Eligible studies described interventions to prevent, minimize, or treat adverse psychological or mental health or wellbeing at an individual, organizational, or systemic level resulting from a public health emergency, defined as emergent situations as those in which the scale, timing, and unpredictability of health consequences threaten to overwhelm routine capabilities (15); for our purposes, we defined these as an epidemic, pandemic, natural disaster, or mass casualty event.The population was limited to frontline HCWs who were involved in the response to the public health emergency.Ongoing public health crises without distinct starting points are excluded from this contextual requirement (e.g., the opioid epidemic, poverty, homelessness as epidemics, or the prolonged HIV/AIDS epidemic).HCWs included medically trained healthcare professionals typically considered at the frontlines (e.g., nurses and physicians in acute care, outpatient, public health, and long-term care) and allied health professions (e.g., personal support workers, infection control professionals, and social workers).Emergency and law enforcement workers whose work is not primarily associated with health (e.g., police and firefighters; non-health-related essential workers such as grocery store employees and cleaning staff; and civilian volunteer emergency responders) were excluded.Any comparison group, such as placebo control, standard care, or pre-intervention data, was included.Studies that reported occupational, psychological, or mental health outcomes were all included.For this review, mental health is defined per the World Health Organization's definition as "a state of mental wellbeing that enables people to cope with the stresses of life, realize their abilities, learn well and work well, and contribute to their community" (16).These outcomes were defined a priori by the research team as anxiety, depression, burnout, exhaustion, resilience, altered mood, or clinical diagnoses of mental disorders such as depression or PTSD.Descriptive-only or in-progress studies that described interventions without reporting on mental health or wellbeing outcomes were excluded.No restrictions were placed on study design, language, publication status, publication date, or study location.

Study selection and data extraction
Once databases were searched and duplicates were removed, the title and abstract of each reference were screened independently by two trained reviewers.Full texts of all potentially relevant articles were retrieved and screened independently by two experienced reviewers.Disagreements were resolved through discussion or a third team member when needed.
Two trained team members independently completed data extraction and critical appraisal, with discrepancies resolved through discussion or a third team member when needed.A data extraction form was created and trialed with three studies and contained 42 items related to study characteristics (study aim, study design, and country); study sample (participant age, gender, professional role, organization, and sample size); intervention and comparator groups; and outcomes (primary and secondary outcomes, measurement methods including anxiety; depression; PTSD; burnout; quality of life; resilience and coping; and mood-related outcomes).Mean differences between groups were extracted to compare outcomes across studies with expected results.Where within-group baseline and end-of-study data were reported, they were transformed into the mean difference and standard deviation of the difference using guidance from the Cochrane Handbook for Systematic Reviews (17).The intervention-related information was extracted using the template for intervention description and replication (TIDieR) checklist (18).Critical appraisal was completed using the appropriate Joanna Briggs Institute critical appraisal tool for the study design extracted (19).Studies rated as unclear or high risk of bias on more than half of domains were defined as having an increased risk of bias; those rated as unclear or high for three or more domains were defined as having some concerns.

Data synthesis
A visual examination of the data extraction table organized by contextual information (i.e., disaster type, setting, and intervention) revealed significant heterogeneity in the data; thus, a meta-analysis was deemed inappropriate.Interventions were categorized post-hoc as falling within four broad categories to support the narrative synthesis: psychotherapy, psychoeducation, mind-body, and organizational interventions; studies that did not fit within one of these categories were classified as "other." Harvest plots were created  for each intervention type and mode of delivery to display study outcomes graphically.Each row of the plot showed effect sizes for each category of results extracted (e.g., anxiety, depression, PTSD, burnout, quality of life, resilience or coping, and mood), with the size of the bar indicating the effect size.Effect size estimates were extracted from the published articles when reported as Cohen's d or Hedge's g or were calculated using the Carlson-Schmidt method for paired pre-treatment and post-treatment data (20,21) or the Campbell Collaboration's effect size calculator created by Wilson (22,23).Bars were color-coded by study design (RCT vs. quasi-experimental) and risk of bias (low, some concerns, and high), and studies with a statistically significant change were labeled.

Characteristics of included studies
Across databases, 30,492 records were identified, along with 47 from reference lists and gray literature (Figure 1).Following de-duplication, the remaining 20,018 records were screened by title and abstract.Of the 1,856 articles assessed at the full-text level, 39 publications describing 36 unique interventions met eligibility criteria (Figure 1) .
Overall positive effects were reported across mental health outcomes (Figure 4, Table 2).However, most studies had some concerns or a high risk of bias.Many were not statistically significant: group CBT during Ebola and e-CBT during COVID-19 both significantly decreased anxiety scores with small-moderate effect sizes (24,34), but a CBT-and positive-psychology-focused app during COVID-19 and group-based Acceptance and Commitment Therapy (ACT) did not (25,27).Group-based CBT during Ebola also significantly reduced depression with small to moderate effect sizes, as did group-based ACT (24, 27); however, this was not seen in interventions of a CBT and positive psychology-based app or e-CBT during COVID-19 or in-person CBT in response to the September 11th attacks (25,26,34).Symptoms of PTSD generally improved across four studies with moderate to large effect sizes (24,26,28,30); however, in one study that compared the effect of in-person group CBT following a natural disaster, those with and without diagnosed PTSD only improved in the group with a clinical diagnosis of PTSD (30).Only a single study each found statistically significant   improvements in burnout (31), mood (24), and coping self-efficacy (29) with small effect sizes following mindfulness-based stress reduction, group-based CBT, and psychological first aid training, respectively; a single study also explored quality of life (25), with no statistically significant effects found using a CBT and positive psychology-based app.Risk of bias in randomized controlled trials.Risk of bias in quasi-experimental studies.
Generally, psychoeducational interventions promoted positive effects on mental health outcomes, although many changes were not statistically significant (Figure 5, Table 2).Both a virtual reality-based stress management intervention and resilience coaching program studies with low risk of bias found moderate to large statistically significant reductions in anxiety (40,44); two app-based psychoeducation interventions found small, non-significant decreases (35,36).Four studies explored burnout.Two group-based psychoeducation interventions resulted in small, statistically significant decreases (44,45).In contrast, a single, group-based session did not result in a statistically significant change.HCWs who used a CBT-based app reported slightly higher burnout scores at the end of the intervention (35).Moderate-to-large reductions in depression and PTSD symptoms were observed after using a psychoeducation app and groupbased workshops (36,37,39).However, differences were generally not statistically significant; a fourth study of an app-based intervention found no change in either outcome (35).Concerning positive mental health outcomes, group-based resilience workshops and a mental wellbeing app resulted in moderate improvements in quality of life, although with some concerns of risk of bias (36,45).A similar improvement was found following a resilience-focused group-based workshop, although the findings were not statistically significant (41).Resilience measures improved following group-based resilience training (38,44) and group-based stress reduction (42) with moderate to large effect sizes.A single resilience training workshop and app-based intervention did not result in statistically significant improvements in resilience among HCWs (36,41).
Statistically significant decreases in depression were only found following mindfulness training plus daily meditation (46); moderate effect sizes were found following transcendental meditation (50), a meditation app (51), and meditation retreats following Hurricane Katrina (56).In one RCT of brief mindfulness meditation and a quasiexperimental study of a meditation retreat, depression scores were higher in the intervention vs. comparator group (52,55).Concerning PTSD symptoms, a meditation retreat following home practices resulted in statistically significant reductions (56); however, reductions were not significant following the use of a mindfulness app (51).Ba Duan Jin exercises (53), laughter yoga (54), and mindfulness breathing and music therapy (57) resulted in large and statistically significant improvements in quality of life; these improvements were not significant following a mindfulness breathing intervention delivered via WhatsApp (48) or the use of a meditation app (51).Only one study explored the impact of transcendental meditation on resilience, which did not find a statistically significant difference between groups (50).

Organizational interventions
Three different interventions were delivered at the organizational level during the 2003 SARS pandemic in Taiwan (58) and the COVID-19 pandemic in the UK (59) and Italy (60), all with low or some concerns concerning the risk of bias (Figure 7, Table 2).These interventions included enhanced training and surveillance infection prevention and control measures (58,60), reorganization of care and staffing levels (58,60), and supports such as yoga, mindfulness, and general mental health support (58,59).The effect of these changes was evaluated after 1 month (59) and up to 5 months (60).An intervention focused on staffing, infection prevention and control measures, and mental health supports resulted in statistically significant reductions in anxiety and depression; effect sizes were moderate, and the risk of bias was low (58).An intervention delivering a mental health support hub within the hospital resulted in a large improvement in mood but was not statistically significant, and there were some concerns about the risk of bias (59).Finally, an intervention including proactive management, reorganized care settings, an increase in staffing, and online education resulted in small but statistically significant improvements in quality of life among nurses (60).

Mental health outcomes by delivery mode
When intervention categories were combined, few clear patterns emerged for differences in effect size by mode of delivery (in-person,  virtual, individual, and group-based, Figures 8-13).With respect to anxiety, interventions that were delivered in a group-based format were more consistently statistically significant than interventions delivered individually, with effect sizes generally moderate, and in-person compared to virtually delivered interventions had more consistently positive findings for reductions in depression.In the remaining comparisons, there were either similar findings by mode of delivery or too few studies in one or more categories to make meaningful comparisons.

Discussion
This review examined the use and effectiveness of strategies implemented at the individual and organizational levels to mitigate the psychological impacts experienced by those responding to a public health emergency.Recent systematic reviews focused on the COVID-19 pandemic have reported on the growing prevalence of adverse mental health outcomes in HCWs, which adds to the body of literature from Harvest plot of changes in anxiety by mode of delivery.previous public health emergencies (63)(64)(65)(66)(67)(68).In addition to documenting the burden, there is also a need to identify effective interventions that can be used in recovery from the early, acute phases of the pandemic and a knowledge of the types of interventions that could be efficiently implemented in future public health emergencies.This review found that psychotherapy had the strongest effect on decreasing several negative mental health outcomes and that psychoeducation may also be effective.Mind-body interventions seem particularly promising for improving quality of life, but more high-quality research is needed.These findings are consistent with the findings of a rapid synthesis of strategies to support the mental health and resilience of the public health workforce (including studies that were not focused on public health emergencies) (69).This review found that effective interventions were multi-sessional, built on   Harvest plot of changes in quality of life by mode of delivery.

FIGURE 13
Harvest plot of changes in resilience by mode of delivery.
part in in-person programming.An important factor that is not considered in research-tested interventions is the cost of delivery.Out-of-pocket expenses for mental health treatment by a qualified professional may be prohibitive for many HCWs.Similarly, there is a need for more cultural competency in psychological interventions (i.e., CBT and PFA).Differing cultural perspectives may lead to conflicting views on how some think about and understand mental health and wellness, and research literature has highlighted the importance of considering cultural components in the spread and scale of evidence-based mental health interventions (71).Recognizing diversity within the workforce is a critical step in providing adequate mental health support.
Based on the visual inspection of the harvest plots, the data suggest that group-based interventions may be more effective than individually delivered interventions, including one-on-one intervention delivery or delivery via the internet or mobile applications, suggesting that a sense of group cohesion is relevant to effective interventions.This is in line with data from qualitative reviews, which have identified feelings of camaraderie and team building as important components that may mediate the efficacy of the intervention on individuals' mental health (72,73).A recently published scoping review synthesized evidence on the key concerns of healthcare providers with respect to COVID-19 (74).Based on the findings, the authors recommended timely and personalized mental health support, particularly through interdisciplinary teams, to minimize burnout; this recommendation is supported by the findings presented here.
To answer our research question, we employed a comprehensive search strategy of multiple databases and gray literature and followed best practices for evidence synthesis, following the Cochrane Handbook for Systematic Review of Interventions to reduce bias in our study.The certainty of the findings of this review is limited by the quality of the included single studies.Most studies were quasiexperimental, using a pre-test and post-test design rather than RCTs.This is not surprising given that many were evaluations of real-world programs that were implemented in response to an unexpected public health emergency and were not pre-designed for the purpose of determining intervention efficacy.Thus, positive findings from studies without an appropriate control group may be confounded by other factors.At the same time, these real-world evaluations may be more generalizable than studies that are conducted in tightly controlled trial conditions.Within the included studies, effect sizes do appear to be larger in studies with a higher risk of bias.A large loss to follow-up was observed across both RCTs and quasi-experimental studies, which may overinflate the effect estimates if those who withdraw from a study are those who are not experiencing a benefit of the intervention.Heterogeneity within categories was high, precluding the use of statistical techniques such as meta-analyses to estimate effect sizes more precisely.These factors should be considered when using the evidence presented here in decision-making.
In addition, studies measuring anxiety, burnout, depression, and PTSD included in this review rarely provided a diagnostic criterion for inclusion.It is possible that the types of interventions needed to effectively treat a clinical diagnosis of a mental health disorder such as generalized anxiety disorder or major depressive disorder may be different than the type of intervention needed to reduce the risk of developing such disorders.Further studies should clearly describe recruitment strategies and eligibility criteria and perform subgroup analyses by baseline scores to better understand whether tailoring of interventions may be needed based on one's baseline mental health status.This is particularly important for including diverse populations who are less likely to acknowledge psychological symptoms and report being concerned about the stigma associated with mental illness (75).Finally, most of the interventions included were of relatively short duration; the medium-or long-term impact of these interventions is unknown.

Conclusion
This review adds to a growing body of evidence on strategies to mitigate the adverse psychological effects of frontline HCWs responding to a variety of public health emergencies, including pandemics and natural disasters.This review highlights the need for ongoing, rigorous research and evaluation of mental health programming implemented at the individual and organizational levels to mitigate the mental health risks in future pandemics or public health emergencies.Strategies to improve the mental health and wellbeing of HCWs following the COVID-19 pandemic will not be a "one size fits all" approach; however, the data to date elucidate promising areas to invest in to support the mental health of HCWs.The author(s) declare financial support was received for the research, authorship, and/or publication of this article.This scoping review was funded by the Public Health Agency of Canada, as part of the work of the National Collaborating Centre for Method and Tools (NCCMT).The views expressed herein do not necessarily represent the views of the Public Health Agency of Canada.

FIGURE 4
FIGURE 4Harvest plot of the effects of psychotherapy on mental health outcomes.

Finally
, two interventions were identified that did not fall into one of the categories above.Giordano et al. used trained music

FIGURE 5
FIGURE 5Harvest plot of the effects of psychoeducation on mental health outcomes.

FIGURE 6
FIGURE 6Harvest plot of the effects of mind-body interventions on mental health outcomes.

FIGURE 7
FIGURE 7Harvest plot of the effects of organizational interventions on mental health outcomes.

FIGURE 9
FIGURE 9Harvest plot of changes in burnout by mode of delivery.

FIGURE 11 Harvest
FIGURE 11Harvest plot of changes in Post traumatic stress disorder by mode of delivery.

FIGURE 10 Harvest
FIGURE 10Harvest plot of changes in depression by mode of delivery.

TABLE 1
Description of interventions.

TABLE 1 (
Continued) Centers for Disease Control and Prevention; CG, control group; COVID-19: Coronavirus Disease 2019; EAP, employee assistance program; HCW, healthcare worker; IG, intervention group; ICU, intensive care unit; NHS, National Health Service; NR, not reported; PFA, psychological first aid; PTSD, post-traumatic stress disorder; RCT, randomized controlled trial; SARS, Severe Acute Respiratory Syndrome; UK, United Kingdom; USA, United States of America.