Perceived stress during COVID-19: Community resilience three years before the pandemic as a protective factor.

Research indicates that stress increased across the globe after the outbreak of the COVID-19 pandemic. Community resilience has been suggested as a central protective factor for stress related to disasters and emergency crises. This study examined the contribution of community resilience reported three years prior to the onset of the COVID-19 pandemic, together with related worries and personal risk factors, to perceived stress among Israeli adults following the first wave of COVID-19 in Israel. We performed a two-period 3-year longitudinal study (Period 1 [P1]: July-September 2017; Period 2: [P2] May-June 2020). The final sample included 578 participants. Participants completed a community resilience self-report questionnaire during P1 as well as measures regarding perceived stress and COVID-19 worries during P2. Using linear hierarchical regression, we tested the additional explanatory effect of community resilience and found it to be negatively associated with perceived stress. While health-related worries were not significantly associated with perceived stress, worries related to the functioning of governmental and health institutions regarding the COVID-19 pandemic were significantly associated with perceived stress. Additionally, being single, living in a smaller residence and income reduction during the pandemic predicted higher perceived stress. The current study highlights the potential buffering role of community resilience in protecting against COVID-19 stress. Assessing community resilience may help identify vulnerable groups, and focusing on community building may be an effective strategy to mitigate stress in future disasters.

Introduction

The COVID-19 pandemic has had a devastating impact on the physical and mental health of the global population. Research indicates that the pandemic has caused significant stress among the general population across many different countries [[1], [2], [3], [4]]. Specifically, perceived stress, defined as a general perception that environmental demands exceed perceived capacity to cope [5], has occurred due to a multitude of factors, including the direct impact of COVID-19 illness, the threat of contagion, bereavement, restrictions to freedom, social isolation, and negative financial changes. Thus, not surprisingly, levels of perceived stress have been observed to have increased during the COVID-19 pandemic [[6], [7], [8]].

The literature suggests a number of factors buffering against stress, one of which is community resilience. Community Resilience describes a community's ability to function amidst crises or disruptions [9], and has been described as a fundamental element in emergency preparedness as well as a means of ensuring social stability in the face of disasters and other major crises [10,11]. One of the ways which community resilience reflects at the individual level is by perceived community resilience. Specifically, individuals with high levels of perceived community resilience may be able to receive needed support during a crisis, as well as prepare for taking action including helping others [11]. Perceived community resilience was found to act as a buffer for individual perceived stress in events such as war crises [12] and terror attacks [13]. Community resilience could thus potentially act as a protective factor for stress during the COVID-19 pandemic.

Several COVID-19 studies have suggested the critical role of communities in coping with the current crisis [14,15], and show that a lack of community resources can be detrimental to individuals [16]. On the other hand, as far as we know, perceived community resilience was little studied as a buffer for stress. Cross-sectional studies conducted during the pandemic have shown mixed results regarding whether perceived community resilience acts as a buffer for distress. For example, in Israel, lower community resilience was associated with higher levels of mental health problems [14], and distress symptoms were shown to be inversely related to a sense of community resilience across both Jewish and Arab populations [17]. Additionally, in another study in the US, higher levels of community resilience were associated with more mental health problems such as posttraumatic stress symptoms (PTSS) [18]. In contrast, in a study conducted in the Philippines, community resilience did not significantly buffer psychological distress in adults during COVID-19 [19], Therefore, the lack of studies regarding the role of community resilience on perceived stress, and these mixed results regarding the possible buffering role played by community resilience in distress, call for research in this field to examine the possible contribution of individual perceptions of community resilience on stress during the pandemic. Additionally, this role should be examined in the context of, and while accounting for, other risk factors cited in the current COVID-19 literature. An important methodological note lies in the fact that these COVID-19-related studies only assessed community resilience during the pandemic. Assessing community resilience during an emergency or crisis situation means that we are unable to disentangle community resilience from the crisis context. Therefore, we cannot determine whether community resilience prospectively protects individuals from stressors associated with that crisis. Thus, with the aim of examining if community resilience acts as a prospective buffer for stress and in order to help vulnerable communities in future disasters [10], we conducted a follow-up of a sample that provided reports of perceived community resilience three years before the pandemic.

According to Cohen and colleagues (2013), there are community resiliency factors which intersect between economic, social and environmental capital that may affect the functioning of the community and that may be protective against stress too [11]. One of the important community resilience related-factors that may impact perceived stress is how the population perceives the functioning of the government and whether they trust their preparation and actions in times of crisis. Specifically, it was found that trust in governmental actions to address COVID-19 are negatively associated with mental health burdens [20]. Additionally, findings have suggested that perceived trust and satisfaction regarding the government's pandemic responses provide protective effects from pandemic-induced depression and anxiety in the general population [21], as well as in medically vulnerable groups and those in quarantine [22]. Relatedly, those who reported lower trust in the healthcare system were found to have more psychological distress [22]. Furthermore, lack of trust in leadership was found to be the most sensitive component in the decline of national resilience following the COVID-19 pandemic [14]. Specifically, it is thus important to assess whether, together with community resilience as a buffer to perceived stress, governmental and healthcare-related COVID-19 worries predict perceived stress. In addition to studying these above-mentioned factors, background risk factors for perceived stress should be considered as well. Specifically, it has been found that being a woman, younger, single, and living with more children led to higher levels of perceived stress [23]. Furthermore, those with lower income [24], those who lost income as a result of the pandemic [25,26], as well as participants living with others were at risk for developing perceived stress [27]. Interestingly, confined living conditions (e.g., those who had smaller residences) has also been suggested as a predictor for a higher risk for developing stress [28].

Current study

We explored the buffering potential of community resilience three years before the pandemic on perceived stress together with demographic characteristics and specific COVID-19-related factors such as worries related to the pandemic (i.e., related to health and government functioning), size of residence where individuals were isolated and/or quarantined and loss of income following the pandemic as predictors of perceived stress. This study was conducted in Israel, which implemented a restrictive lockdown early on to manage the pandemic.

Methods

Study setting

The current study is part of a two-period study examining mental health in Israel in a large sample of Israeli citizens. The first period of data collection was conducted during July–September 2017 (P1), a relatively calm period without any major armed conflict, terror exposure, or health crisis, and the second period of data collection took place immediately following the first wave of COVID-19 and before the second wave in Israel between May 12 – June 26, 2020 (P2).

Israeli Covid-19 status information at the time of the study

On March 11, 2020, Israel enforced social distancing rules, and restricted group gatherings, and the following week, Israel declared a national state of emergency and went into a very restrictive lockdown. During April, May and June, lockdown restrictions were eased. By the final day of data collection, 23,109 people had been infected in Israel with COVID-19, 15,432 had recovered and 321 had died (https://datadashboard.health.gov.il/COVID-19).

Participants

Data were collected in both periods via a data collection company (existing panel according to quotas). Participants for P1 were selected from a panel of participants kept by the survey organization and through individuals responding to online adverts. The participants were either contacted by phone or by internet (according to the participants' preference) based upon representative population quotas (sex, age, and location), with deliberate oversampling of participants from ethnic minority backgrounds and immigrant participants. The total number of individuals surveyed at P1 was 1350. The sample for the current study were participants who participated in both data collection period (n = 578; Details of the selection flow is presented in Fig. 1 ). Of the final sample, 18% (n = 104) completed the questionnaire by phone. The inclusion criterion for this study was age 18 and above at P1. The mean age at P1 was 45.2 (SD = 16.15). Table 1 shows that the sample rates for gender, and ethnicity minority background in the final sample are similar to the rates in society as described by the Central Bureau of Statistics (CBS, 2021): women 50.03%, Jewish 78.9%, Non-Jewish 21.1%. The proportion of participants in each age group was also similar to the rates in society in 2017 (CBS, 2019). In our study, 11.1% were between ages 18–24 (15% according to the CBS), 18.5% were between ages 25–34 (20% according to the CBS), 21% were between ages 35–44 (20% according to the CBS), 17.5% were between ages 45–54 (16% according to the CBS), 13% were between ages 55–64 (18% according to the CBS), and 14% were above age 65 (15% according to the CBS). For more details about the sample, see Table 1. Only 0.3% (n = 5) of the participants reported that they had contracted COVID-19, however, 29% reported that they were considered at high risk for COVID-19, and 32% lived with someone considered high risk. Ethical approval was received from the University of Haifa's institutional review board for both periods (IRB ethical approval reference numbers: 211/17 and 236/20).

Fig. 1

Table 1

Sample Characteristics (N=578).

Characteristics	N M(SD) or %
Gender
Men	250	43%
Women	328	57%
Religion
Jewish	469	81%
Muslim	92	16%
Christian	12	2%
Other	5	.9%
Secular	288	50%
Traditional	135	73%
Religious	128	22%
Ultra-orthodox	27	5%
Age (in years)	578	45.2 (16.15)
Children P2
No	155	27%
Yes	421	73%
Missing	2	0.003%
Education P2
Not university educated	329	57%
University educated	249	43%
Relationship Status in P2
In a relationship	431	75%
Single	145	25%
Missing	2	0.003%
Income status
Lower than average	123	25%
Average	198	40%
Higher than average	171	35%
Missing	86	14%
Significant reduction in income during the crisis
Yes	284	51%
No	292	49%
Missing	2	0.003%
Contracted COVID-19	5	0.3%

Comparison of dropout and final sample at P2

In the dropout sample, women comprised 58% of the sample, and there were 56.6% Jewish participants and 41.1% non-Jewish participants. A total of 15% were between ages 18–24, 34% were between ages 25–39, 26% were between ages 40–54, 16% were between ages 55–69, and 9% were above age 70. In addition, there were no significant differences between the study sample (n = 578, who completed both periods) and the dropout sample (n = 791, who only completed P1) regarding community resilience scores (Study sample; M = 3.35, SD = 0.90, Dropout sample; M = 3.55, SD =  = 0.90.

Measures

All questionnaires were translated and back translated from English into Hebrew and Arabic, and completed according to participants’ preferences in both P1 and P2 (Hebrew 100%).

Background variables included age, gender, income, reduction in income during the pandemic, relationship status (in a relationship/single), children (Yes/No), size of property where one resides assessed by using an eight-point ordinal scale from 40 m2 up to 200 m2 or larger.

Health worries and Government/Healthcare system worries during COVID-19

were measured using four items taken from a study by Ref. [29]; and an additional 5 items developed for the current study. Regarding health concerns participants were asked six questions (3 from Ref. [29] and 3 added by the authors) to indicate how worried they were about contracting the virus and their friends and family becoming infected. Regarding the ability of the government and the healthcare system to manage the pandemic 3 items were used (2 from Ref. [29] and 1 added by the authors). All questions were assessed on a 5-point scale from 1 “not at all” to 5 “very much”). A mean score was calculated for each factor (please see details in the factor analysis of the questionnaire items for these factors in the supplementary Table 1). The internal reliability of the health measure was α = 0.81 and of the ability of the government and healthcare system to manage the pandemic measure was α = 0.75.

Perceived community resilience was assessed using the 10-item version of the Conjoint Community Resiliency Assessment Measure (CCRAM; [9]. This measure assesses five domains, namely social trust (e.g. “ There is trust among the residents of my town "), collective efficacy (e.g. “ I believe in the ability of my community to overcome an emergency situation”), leadership (e.g. “ The municipal authority of my town functions well”), emergency preparedness (e.g. “ My town is organized for emergency situations”) and attachment to place (e.g. “ I feel a sense of belonging to my town”), Participants responded on a 5-point scale ranging from 1 (does not agree at all) to 5 (totally agree). Most items were skewed to the right (skewness smaller than zero, Zskew<1.65; p(K–S)<0.05), yet a typical bias in items coded on a limited and discrete scale. A mean score was calculated, and the Cronbach alpha reliability was α = 0.91.

Perceived stress was measured by the Short Form Perceived Stress Scale (PSS-4, [30]. The questionnaire includes four items that assesses the degree to which personal situations during the past month were appraised as unpredictable, uncontrollable and overwhelming. For example: “In the last month, how often have you felt that you were unable to control the important things in your life?”. Items were rated on a Likert scale ranging from 0 (never) to 4 (very often) and a mean score was calculated for all items. The internal reliability of the original scale was α = 0.60 [30] and in the current study was moderate for a 4-item scale with α = 0.54.

Analysis plan

In order to examine the possible associations between perceived stress and potential protective and risk factors, we conducted Pearson correlations between these variables. Significance levels were adjusted by applying the conservative Bonferroni correction. Additionally, we used a t-test to examine differences between genders in relation to the outcome variables. A hierarchical linear regression analysis was used to assess predictors of perceived stress at P2. We tested the effect of risk factors: age, gender, having children, education level, relationship status, meters of living space (Step 1), and worries due to COVID-19 regarding health and government and health system (Step 2), and then the protective factor of community resilience at P1 (Step 3). Beyond regression coefficients, we provided the R change as an assessment of the additional contribution of each step to the overall explanatory power of the model. Missing data were low for almost all variables (less than 1%, apart from size of property which had 12% missing and Income status which had 14% missing); Statistical analyses were performed using SPSS Statistics 25.0.

Results

Descriptive analyses

Only 0.3% (n = 5) of the participants reported that they had contracted COVID-19 by the time of data collection. The mean score for health worries was 2.90 (SD = 1.00), and the mean score for government/healthcare system worries during COVID-19 was 2.80 (SD = 1.12). Bivariate correlations among study variables.

The bivariate correlations between the study's continuous variables are presented in Table 2 . First, perceived stress at P2 was significantly and negatively associated with community resilience at P1. In addition, perceived stress exhibited significant associations with all other study variables. Specifically, perceived stress was significantly and positively associated with both types of worries related to COVID-19, and size of residence. Secondly, no differences in gender, education level, or having children were found regarding perceived stress at P2. However, regarding P2, Jewish participants reported higher stress levels (M = 10.48, SD = 2.90) than non-Jewish participants (M = 9.83, SD = 3.20) (t(172) = 2.04, p < .05), participants not in an intimate relationship reported higher stress levels (M = 9.50, SD = 3.05) than those in relationships (M = 12.00, SD = 2.73), t(52.70) = −5.72, p < .001), and participants whose income changed negatively during the pandemic reported higher stress levels (M = 10.50, SD = 3.30) than those with no change or increased income (M = 9.50, SD = 2.94), t(563) = −3.80, p < .001).

Table 2

Bivariate correlation of the study variables.

	1	2	3	4	5
1. Age
2. Perceived Community resilience (P1*)	.20**
3. Size of property one resides in (square meters)	.23**	.10*
4. Health worries during COVID-19	−.02	−.05	−.04
5. Government/Healthcare system COVID-19 worries	.00	−.20**	−.08	.53**
6. Perceived Stress	−.10*	−.16**	−.14**	.19**	.25**

Note: * All variables except community resilience were examined at P2.

*p < .05, **p < .01.

Hierarchical linear regression

Table 3 presents the hierarchical linear regression model results. In Step 1, we found that married respondents reported lower levels of stress at P2 (β = −0.16, p < .01) as did owners of larger properties (Meter2: β = −0.11, p < .05), whereas income reduction during COVID-19 was associated with greater perceived stress at P2 (β = 0.15, p < .05). In Step 2, we found that higher levels of worries from public regulations due to COVID-19 were associated with higher perceived stress outcomes (β = 0.17, p < .01). However, health-related worries did not predict perceived stress. Results indicate that community resilience at P1 predicted lower perceived stress at P2 (Step 3: β = −0.09, p < .05). Steps 2 and 3 each add significant explanatory power over and above Step 1 (Step 2 ΔR : 0.05, p < .001; Step 3 ΔR : 0.08, p < .05).

Table 3

Hierarchical Linear regression predicting Perceived Stress P2.

	B	SE	beta	p-value
Step 1
Sex	−0.23	0.28	−.04	.405
Age	−0.01	0.01	−.05	.336
Children	−0.05	0.39	−.01	.901
Education	0.18	0.28	.03	.535
Relationship Status	−1.23	0.38	−.16	.001
Size of property when one resides (square meters)	−0.16	0.07	−.11	.016
Income reduction during COVID	0.95	0.28	.15	<.001
R2	.07			<.001
df	7500
F	5.55			<.001
Step 2
Health-related COVID worries	0.29	0.17	.09	.088
Government/Health system COVID worries	0.49	0.15	.17	.001
ΔR2	.05			<.001
R2	.12			<.001
df	9498
F	7.75			<.001
Step 3
Perceived Community resiliencea	−0.33	0.16	−.09	.038
ΔR2	.008			.038
R2	.13			<.001
df	10,497
F	7.45			<.001

Note: Sex: 0 = Male, 1 = Female, Education: 0 = Not university educated 1 = University educated; Children: 0 = No children, 1 = at least one child. Income changes during the pandemic: 0 = No change or increased income, 1 = Reduced income/.

All variables except perceived community resilience (assessed at P1) were examined at P2.

Discussion

The current study aimed to examine community resilience as a protective factor for stress related to the COVID-19 crisis. The main result indicated that components related to community such as perceived community resilience serves as a protective factor for perceived stress. Additionally individual socio-demographic factors, such as age, gender, number of children, relationship status, as well as worries about the government and the healthcare system, a reduction in income, and living space were found to be risk factors for stress.

These results are in line with previous studies suggesting that community resilience may play a significant role in buffering perceived stress for those exposed to threats of war [12] and terror [13]. Our current findings broaden the literature on community resilience to the COVID-19 pandemic context [31], and highlight the importance of perceived community resilience as an individual resource for coping with the threat created by the pandemic. Furthermore, our study results highlight the significant role of measuring community resilience before the crisis as a resilience factor for perceived stress beyond individual psychological risk factors. In this study, we examined two individual worries related to the current situation. Health related worries were not significantly associated with perceived stress in this study, which is surprising as they have been related to mental health problems [32] as well as perceived stress in other studies [24]. A possible explanation is that during the data collection the number of COVID-19 cases was relatively low. However, government/health system COVID-19 worries were significantly associated with perceived stress. This supports the understanding that individuals' perceptions of covid-19-related community services impacts individual stress during the pandemic. Moreover, since community resilience was examined in this study as individually perceived community resilience, future studies should examine the dynamic relationship between individuals' perceptions and environmental components of community resilience, namely, objective factors of resilience, including infrastructure, economic resources, availability of various services which assess objective aspects of community resilience [33]. In addition to community resilience, the study results examined individual risk factors for perceived stress, such as not being in an intimate relationship [23], increased financial insecurity [34], and condition of residence [28]. These are specific individual, but also community-related risk factors that can impact personal resilience. Our results indicate that COVID-19 interventions could be applied at the community-level [35], focusing on strengthening solidarity and practical cooperation and providing a personal sense of trust in the community. Organizational and community level components that support resilience, such as cooperation of the state with communities through community health services, local neighborhood committees, and volunteers' organizations, could reduce risk, support effective coping, and mitigate stress. Moreover, results from a rapid scoping review of community resilience during COVID-19 [36] suggest that the ability of the government to be effective in building community resilience also depends on the trust of the population in the government's ability to properly respond to the crisis. Examples of such interventions can be seen in Cuba [37] and in Singapore [38], such as identifying vulnerable individuals and helping community-based organizations establish on-going ties to help support these individuals' emotional and mental well-being especially during the COVID-19 pandemic. Therefore, we suggest that, besides the efforts to improve physical and psychological health, health officials take seriously the need to build enduring and trustworthy relationships with communities.

Limitations

The current study has a number of strengths, namely pre-COVID-19 reports of community resilience among a relatively large and diverse sample. Nevertheless, there are several limitations. First, this study used self-reports; future studies should consider in-person interviews, as well as assessments of more objective aspects of community resilience [11]. Second, future studies should consider clinical interviews.

Third, although the current sample is representative of the population in Israel regarding gender, ethnicity and age, it was not a randomized sample, somewhat limiting the generalizability of the results. Fourth, specific groups in the population are under-represented because they have less access to the internet and recruitment of respondents is often based on self-selection. Both under-coverage and self-selection may lead to biased estimates (Bethlehem, 2010) in the discussion. Fifth, being conducted early in the pandemic our results represent a specific phase in the crisis and might not be relevant for later periods. Sixth, in the current study the internal reliability of the PSS-4 was low, which might limit the robustness of the study findings. This measure might not adequately capture the specific stressors linked to the pandemic.

Study implications

The current study indicates that various factors were associated with perceived stress during COVID-19, including worries regarding government and health system functioning during the COVID-19 pandemic and low levels of perceived community resilience. It may be that strengthening community resilience in routine times can improve disaster preparedness. Furthermore perceived community resilience assessment measures may provide some indicator of potential individual and community reactions in the face of potential crises [10]. As suggested by Ref. [39] assessment, action planning, and the recognition of the uniqueness of each community are key to strengthening resilience. As community resilience is also related to perceived social support, to the strength of social connections, and to the physical and mental health of the public [11], policy makers may consider ways to develop community social support and address specific worries and concerns, for example by using local neighborhood teams to check up on individuals living alone, and provide online meetings, activities, support sessions and disseminating information and updates [34,40]. These efforts may contribute to reducing levels of perceived stress during the pandemic as well as other large-scale crises.

Funding

The current study was funded in part by the Israeli Ministry of Social Affairs. They were not involved in any aspect of preparing this manuscript.

Uncited references

[[41], [42], [43], [44], [45]].

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.