Psychological Interventions in a Pandemic Emergency: A Systematic Review and Meta-Analysis of SARS-CoV-2 Studies

Background: The study aim was to review the evidence and effectiveness of psychological interventions applied during Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) pandemic. Methods: A literature search was run from April 2020 to April 2021. The inclusion criteria were: (1) RCTs or observational studies; (2) professional health carers and patients who had contracted coronavirus during the pandemic; (3) adults and elderly people with a viral infection diagnosis; (4) suitable measures to assess intervention effectiveness on clinical status and psychological and behavioral aspects. An internal validity assessment was planned using robvis. Data were synthesized according to PICO criteria. Results: A total of 12 studies were selected. Studies measuring mental health outcomes demonstrated the greatest reduction in symptoms, with eight out of the twelve studies demonstrating a reduction in symptoms that reached statistical significance (p < 0.05) and four of the studies reaching a higher significance level of p < 0.01. The type of psychological intervention was predominantly cognitive behavioral therapy (CBT). All studies except one was run online. Conclusion: Psychological interventions have a benefit on mental health outcomes, even if performed online. In particular, CBT seems to be the psychological intervention that is used more and also seems to have a larger effect size on the mitigation of mental health symptoms and increasing resilience.


Introduction
In the past 10 years, several viral epidemics have occurred, such as influenza caused by the virus subtype H1N1 (2009), Middle East respiratory syndrome-MERS (2012), Ebola virus disease (2014) [1][2][3], and a novel virus Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) that emerged around the world at the end of 2019. Globally, response measures were taken, including self-isolation, social distancing, and quarantines, augmenting further the distress.
During the SARS-CoV-2 pandemic, innovative technological know-how displayed a great ability to back up public health interventions [8], permitting an effective psychological assistance for patients during the pandemic with a positive impact on mental health aspects. However, evidence-based advice on what would be effective in mitigating this impact is not yet clear. A rapid review was conducted to detect the most effective interventions for reducing nurses' psychological issues during outbreaks [9], but the 14 articles included were the recommendations of organizational and individual self-care interventions, except for a study which described a digital intervention and user satisfaction. Another rapid systematic review was conducted to identify coping strategies among healthcare workers [10], but it only included studies with at least 100 healthcare workers, and removed studies with smaller samples that might contribute to clinical practice and future research.
Therefore, the aim of this work was to review the evidence and effectiveness of psychological interventions applied during the SARS-COV-2 pandemic, including therapeutic outcomes and measurable psychological improvements (reduction in depression and/or anxiety and/or post-traumatic stress symptoms from baseline to the last available followup, evaluated using suitable measures to assess psychological and behavioral aspects and clinical status).
A literature search was run using Scopus, MEDLINE (Web of Science), CINAHL Complete, PsycINFO, and PubMed (from April 2020 to April 2021).
The PRISMA guidelines for systematic reviews were followed [11] and employed to evaluate the quality of the study. An inductive approach was applied to the analysis [12].
The search queries included terms (Severe Acute Respiratory Syndrome Coronavirus 2 OR Coronavirus OR SARS-CoV-2 OR COVID OR COVID-19) to determine the interest outcomes (viral infection AND [psychological treatment OR psychological first aid OR psychological support OR psychotherapy sessions]).
The inclusion criteria are shown below: (1) RCTs or observational studies; (2) professional health carers and patients who contracted coronavirus during the pandemic; (3) adults and elderly people with viral infection diagnosis; (4) suitable measures to assess intervention effectiveness on clinical status and psychological and behavioral aspects. The exclusion criteria were the following: (1) participants with severe mental health symptoms such as psychosis or suicidality; (2) studies that were not published in English, and (3) no psychological interventions.

Study Selection
Three reviewers applied eligibility criteria and selected studies for inclusion in the systematic review. Of them, two reviewers screened records for inclusion and one reviewer checked decisions in a blinded manner.
The characteristics to assess were the following: randomization, deviations from the intended intervention, missing data, outcome measurement, and selection of reported results.
The assessment was conducted at study and outcome level. The criteria used to assess the internal validity was planned using robvis (a visualization tool for risk of bias assessments in a systematic review) [13]. Inter-rater reliability was measured using Cohen's kappa, k with −1 (absence of an agreement) to 1 (perfect agreement) [14].
Three researchers extracted and checked the received data. Of them, two researchers independently extracted data and one researcher checked the extracted data and quality assessment.

Data Synthesis
The recorded data were inserted into one .xls spreadsheet and analyzed using R Ver. 2.8.1 statistical software package (The R Project for Statistical Computing; available at URL http://www.r-project.org/, accessed on 11 April 2022). The data to be extracted from the studies were the following: sample size, sample type (patients, clinicians, nurses, other health professionals), mean age ± sd, psychological intervention type, performed session number, session modality (onsite or online), performed measurements, and achieved outcomes (pre and post intervention). The main outcome(s) was reduction in depression and/or anxiety and/or post-traumatic stress symptoms from baseline to the last available follow-up, evaluated using suitable measures to assess psychological and behavioral aspects and clinical status.
The data were synthesized according to PICO criteria [15], consisting of the testing for a relationship between the size and precision of study effects, the risk ratios for individual studies combined using a random effects meta-analysis, and co-authors supplying a complete summary and accurate organization of all studies covering transparency, approach, methodology, and strengths/weaknesses [16,17].

Results
The details of the literature search and study selection procedure are outlined in the PRISMA flow diagram ( Figure 1). was measured using Cohen's kappa, k with −1 (absence of an agreement) to 1 (perfect agreement) [14].
Three researchers extracted and checked the received data. Of them, two researchers independently extracted data and one researcher checked the extracted data and quality assessment.

Data Synthesis
The recorded data were inserted into one .xls spreadsheet and analyzed using R Ver. 2.8.1 statistical software package (The R Project for Statistical Computing; available at URL http://www.r-project.org/. Accessed on April 11, 2022). The data to be extracted from the studies were the following: sample size, sample type (patients, clinicians, nurses, other health professionals), mean age ± sd, psychological intervention type, performed session number, session modality (onsite or online), performed measurements, and achieved outcomes (pre and post intervention). The main outcome(s) was reduction in depression and/or anxiety and/or post-traumatic stress symptoms from baseline to the last available follow-up, evaluated using suitable measures to assess psychological and behavioral aspects and clinical status.
The data were synthesized according to PICO criteria [15], consisting of the testing for a relationship between the size and precision of study effects, the risk ratios for individual studies combined using a random effects meta-analysis, and co-authors supplying a complete summary and accurate organization of all studies covering transparency, approach, methodology, and strengths/weaknesses [16,17].

Results
The details of the literature search and study selection procedure are outlined in the PRISMA flow diagram ( Figure 1). As shown in Figure 1, a total of 1416 articles were identified, of which 103 duplicates were deleted. After abstract evaluation, 1169 studies were excluded. Another 132 studies were excluded after a more comprehensive examination. Consequently, 12 published studies were suitable for this systematic review.

Methodological Quality
According to Cochrane's recommendations, the robvis tool was used in order to assess bias risk (Figure 2), rating bias as "low", "high", or "some concerns" against several domains. As shown in Figure 1, a total of 1416 articles were identified, of which 103 duplicates were deleted. After abstract evaluation, 1169 studies were excluded. Another 132 studies were excluded after a more comprehensive examination. Consequently, 12 published studies were suitable for this systematic review.

Methodological Quality
According to Cochrane's recommendations, the robvis tool was used in order to assess bias risk (Figure 2), rating bias as "low", "high", or "some concerns" against several domains. The nature of the interventions meant that most studies presented a trend towards a low risk of bias. At this point in the review, a meta-analysis was considered appropriate.

Data Extraction and Outcomes
A data extraction table was created in order to standardize the process (Table 1). Contacting the authors was unnecessary because no missing information was identified. The nature of the interventions meant that most studies presented a trend towards a low risk of bias. At this point in the review, a meta-analysis was considered appropriate.

Data Extraction and Outcomes
A data extraction table was created in order to standardize the process (Table 1). Contacting the authors was unnecessary because no missing information was identified.
All studies, except for Shaygan et al. [22], occurred in countries of high economic development.
The study aims were to treat mental health symptoms, with most participants having baseline symptoms classed as clinical disease.
The sample sizes ranged from 9 participants to 1500 participants. Seven studies included patients, of which three studies included people with a mean age ranging from 40.8 to 48.3 years [19,24,25], two studies included people with a mean age ranging from 34.6 to 36.8 years [22,26], and two studies included younger people with a mean age ranging from 26.1 to 28.5 years [23,28]. Another study included health care workers without personal identifying information being collected [18]. The last four studies included the general population/adults with a mean age ranging from 42.0 to 45.0 years [20,29] and students with a mean age ranging from 8.2 to 31.0 years [21,27].
Finally, the three no-RCT studies were a validation protocol study [18], a case-control study [19], and a pilot study [29].
Only in two studies was the type of psychological intervention a psycho-educational approach [21,22] with a session duration of 1 h per week [21] and per day [22].

Outcomes
Post-intervention, the mental health outcome scores were significantly improved compared to the pre-intervention scores for all twelve studies ( Table 2). The Cohen's d effect sizes (where calculated) were small to large (d = 0.21-2.89), indicating an increase in primary/secondary coping abilities [27,28], and a decrease in depression and anxiety [20,26,28], functional impairment [20], intolerance of uncertainty [20], and insomnia [20,26,29]. However, in two other studies the d ppc2 and df effect sizes were calculated showing, respectively, large and medium-large effect sizes in resilience increase and perceived stress reduction [22], and a very large effect size in depression and anxiety reduction [23].  The partial eta squared (ηp 2) effect size was calculated in three studies, revealing a significant large effect (ηp 2 = 0.20-0.89) on mental health condition with a decrease in depression and anxiety symptoms [21,24,25], functional and attention deficits [21], and traumatic symptoms [24]. A medium effect (ηp 2 = 0.07-0.08) was reported in insomnia reduction [25].

Discussion
There were no instances of the intervention group having worse scores post intervention. In contrast, in three of the eight studies with control groups the mental health scores worsened over the study period [22][23][24]. Only two study with one-month follow up period, post-intervention outcome scores were maintained [20,25].

Discussion
According to the study aim, the evidence and effectiveness of psychological interventions applied during the SARS-CoV-2 pandemic, including therapeutic outcomes and measurable psychological improvements, were reviewed.
In the meta-analysis, it showed that psychological interventions have a benefit on mental health outcomes, even if performed online. In particular, CBT seems to be the psychological intervention that is used more and seems to have a larger effect size on the mitigation of mental health symptoms and increasing resilience.
CBT has a positive impact on symptoms such as depression, anxiety, and perceived stress in health care workers, patients, general population, and students. The American Psychological Association and International Society of Traumatic Stress Studies specified to constantly monitor the mental health status of pandemic-affected populations and provide timely evidence-based trauma-centered psychotherapies, such as CBT [30][31][32].
However, beyond the above, the importance of enhancing the online psychological support has emerged, because of traditional face-to-face psychological facilities being unavailable during pandemic outbreak [32]. Tele-medical approaches needed to be implemented in public health strategies in order to support many people simultaneously and effectively [33]. The benefits of e-mental health approaches are great, and these innovative resources are currently more necessary than ever before [34].
Given the evidence for the risk of short-and long-term psychological consequences and their impact on job performance and quality of care [35], psychological packages should be urgently implemented to manage mental health care for medical staff working on the frontline against the pandemic [36].
The studies by Brooks et al. [37], in the context of the psychological effects on people due to other epidemics (Ebola and SARS), identify five main causes of stress related to lockdown: (1) duration of the lockdown; (2) fear of infection; (3) feelings of frustration and boredom; (4) inadequate devices or unobtainable masks; (5) inadequate information or an excessive increase in information sources.
To these factors, we can also add concerns related to the financial situation that have generated negative feelings in some individuals and symptoms attributable to psychoemotional stress.
Furthermore, it is necessary to identify how certain characteristics inherent in factors such as demand, control, and support play a role in modulating the perception of stress [38]. Therefore, it is necessary to bring the reflection on stress closer to the coping constructs, highlighting how individual motivation to a task influences the perception of the situation, thus modulating the level of stress. The contribution of Caplan and Van Harrison relates to the observation that, alongside stress, adaptation coexists as a possible adaptive response; this outcome is the result of an interaction between person and environment to which objective and subjective variables present both in the person and in the environment contribute. Regarding the included studies in our work, we found limited evidence for coping capacity. Predominantly, the studies have a low risk of bias, but we only dealt with the main effects. Moreover, not all the variables were comparable, in particular the heterogeneity of size effect assessment. It is necessary for future studies to improve design quality to increase the trust in their outcomes.
Therefore, psychological systems as well as medical systems of health care require huge investment and major upgrading in order to be prepared for future pandemics.

Conclusions
An excellent way to improve mental health is to focus on reducing social inequity through an integrated community-oriented system of care operating across health and social care systems, supported by operational research to guide implementers and policymakers to deal with current and future challenges. This requires political will at the highest policy-making level, and an increase in resources at all levels of health and social care systems [39,40].