From this global survey with 1,762 completed questionnaires involving 26 countries or regions, we found that the overall proportion of individuals having a low resilience level was more than one-quarter of the population, with a higher proportion in the Americas and European countries than in Asia Pacific regions. Younger individuals, female participants, those with a poorer financial situation in the past 6 months, and people with multimorbidity were associated with a low resilience level. We also found that low resilience was associated with the practice of various unhealthy behaviours and risk factors, which could represent maladaptive coping lifestyles.

There are few surveys that have measured the resilience levels of the general population performed on a global scale. Killgore et al. (2020) conducted an online questionnaire which collected data from 1,004 English speaking participants during the third week of the COVID-19 stay-at-home guidance in all 50 states of the US [20]. They used the Connor-Davidson Resilience Scale (CD-RISC) [26] and other validated surveys, including the Beck Depression Inventory-II (BDI-II), the Zung Self-Rated Anxiety Scale (SAS), and the Multidimensional Scale of Perceived Social Support (MSPSS), to measure psychiatric symptoms [20]. They found a significant lower psychological resilience level during the first weeks of the COVID-19 lockdown in the US when compared with published normative data for CD-RISC, implying a possible adverse influence of the pandemic conferring acute alterations in emotional outlook and perceived support. Also, they demonstrated that low resilience was associated with poorer mental health outcomes such as severe depression, anxiety, worry about the effects of COVID-19, greater difficulty coping with emotional challenges, and suicidal ideation. Another study among 898 young adults aged 18 to 30 years in the US from April to May 2020, which is one month after declaration of a state of emergency, found that up to 72% of the study participants, of which 81.3% were women, had low resilience [21]. High resilience level (score ≥ 30 in CD-RISC-10) was significantly associated with lower depression, anxiety and post-traumatic stress disorders. The findings of these studies, in conjunction with our results, highlight the low resilience levels of the general populations and the importance to enhance it given its potential psychological consequences.

We found that higher proportion of participants in Western countries reported lower resilience level than in Asia Pacific regions. In general, Asian individuals were less likely to report high levels of mental health symptoms than individuals from European and American regions [21]. Furthermore, Asian and Latinx immigrants, compared to participants born in the U.S., are less likely to endorse psychological distress [27, 28]. It has been speculated that other experiences such as ethnic identity, social networking, family cohesion, and even religiosity could act as a protective factor for mental health [29, 30].

We found that young participants aged less than 30 years suffered from lower resilience than the elderly population. Previous evidence from both Western and Asian countries demonstrated that younger people, e.g., 18–24 years old (and not exceeding 40), had the greatest increase in rates of psychological distress during the pandemic [31,32,33,34]. Younger age (under 35 years old) has been shown to be a significant mediator for stress-anxiety mediation models [35], implying that younger adults may be more vulnerable to stress and anxiety during this pandemic. In general, younger individuals fared the worst when they experienced depression, stress, and anxiety symptoms. An Australian survey conducted in April 2020 also found that younger people aged < 45 years were the most vulnerable group to psychological distress [33]. Loneliness and financial distress have been linked to poorer depression and anxiety outcomes, respectively, in younger adults. On the contrary, older adults demonstrated higher resilience than other age-groups, which could be due to their greater ability to savour life experiences [36, 37].

Furthermore, female individuals tend to have lower resilience during the pandemic. This is compatible with a recent comprehensive review on sex differences in resilience [38]. Females demonstrate increased vulnerability in times of stress, which could be attributed to gender-, sex hormone-, and sex chromosome-life span interactions in producing resilience. This gender difference could also be due to internalization of trauma, generalization of fear cues, anhedonia, passive coping, and blunting of corticosteroid response to stress among women [39], although the influence of sex on risk and resilience to stress could be complex that varies according to the type, timing and duration of the stressor as well as development with its associated changes in brain structure and function. The findings are in line with recent literature from the perspective of social inequality that suggests additional gendered vulnerabilities and stresses placed on women during the lockdown and thereafter, ranging from decreased health care access, increased unemployment, domestic violence, and higher risks of exposure and infection via work in health care industries [40,41,42].

Last but not least, people with multimorbidity were more likely to have low resilience score. Despite the fact that the massive global effort and resources directed to COVID-19 have completely overshadowed the pandemics of noncommunicable diseases in the twenty-first century, multimorbidity remains a crucial element of both diminished resistance to coronavirus infection and diminished resilience as found in this research. Individuals with multimorbidity suffer from physical challenges as well as social-psychological feelings of stress, anxiety, depression, loneliness, low self-esteem, social isolation, and changes in social roles [43]. Our findings indicate that their resilience level was likely to be low, which could be explained by the need for continuous efforts in maintaining healthy levels of functioning following adversity, which is a dynamic process but not a personality trait [44]. Given the possible widening of inequalities in both income and welfare, further research are necessary to illustrate whether the impact of inequality on the resilience of societies within developing economies may differ from that in industrialized nations during the course of globalization.

We also observed the presence of associations between low level of resilience and maladaptive coping behaviours such as low level of exercise, greater uptake of unhealthy food, increased consumption of alcohol and medications. The practice of these maladaptive coping behaviours represents various unhealthy lifestyles, which could be conceptualised as ‘industrial diseases’ that are related to poor diet, alcohol, gambling, drug and tobacco related diseases as these are directly associated with the vectors, i.e., the unhealthy commodity industries [45]. These industries are known to target those with least education, employment and income, and are known to target them during the COVID-19 pandemic, with a wide range of national and international policies that continue to ensure intergenerational disadvantage within countries and between countries. Therefore, it is highly likely that the association between low resilience and maladaptive coping behaviours observed in our study is mediated by the activities of these unhealthy commodity industries. This further calls for a need to shape the policy strategies for mitigating the negative impact of these unhealthy commodity industries, and for minimising the resultant impact of highly ineffective and inequitable policies among people across the globe.

This survey involved a large number of countries and represents a global collaboration from researchers in various study sites. The survey was devised and pilot-tested by an expert panel, and we used a published, validated resilience survey with good internal consistency and test–retest reliability. Nevertheless, there are several limitations that should be addressed. Firstly, we are unable to capture the response rate as this consecutive sampling strategy did not provide the number of participants who received the survey invitation. Furthermore, casual relationships between low resilience and the unhealthy behaviours could not be inferred, as there is a possibility of reverse causality in this cross-sectional study. In addition, the survey responses were received from a long period of time lasting for approximately two years. The number of new COVID-19 cases and mortality due to the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) fluctuated with time, and it is unknown whether the severity of the pandemic, its related public health measures, and the changing social distancing measures in different COVID-19 waves might exert an impact on the resilience levels. Furthermore, the generalizability of our study findings to other settings should be interpreted with caution, as we did not adopt a random sampling strategy and not all countries provided a large number of responses. Selection bias might exist as the current survey tended to attract participants who were regular internet users. Also, this study did not evaluate the reasons of low resilience among participants with the associated factors, and the mechanistic aspects of how low resilience might lead to unhealthy behaviours examined in this survey. Further in-depth work is required to examine the detailed pathways where resilience might influence personal behaviours in the pandemic.

Our study findings have identified a subgroup of individuals who may be at a higher risk for low resilience level. In particular, it is worrying that resilience is not only a stand-alone observation – it could also be closely associated with more unhealthy behaviours as demonstrated by results of this global survey. Interventions targeting these at-risk participants are needed. For instance, a previous survey [20] identified that more time spent on outdoors in the sunshine for at least 10 min; daily exercise; more extensive support from family and friends; care from a close significant other; and improvement of spiritual health such as prayers could be predictive of better resilience. Physicians could offer anticipatory care during their clinical consultations to provide resources related to interventions for patients with low resilience, such as the adoption of a holistic approach and behavioural modification to enhance self-efficacy. Proactive measures to improve and sustain resilience, including building of coping skills and implementation of social support networks, could be crucial [46]. Policy-makers should consider community-based programmes that could target these at-risk individuals by evidence-based resilience-enhancing strategies through a concerted, multidisciplinary effort, as interventions to improve resilience is often multi-pronged.

From a global health perspective, the risk factors identified for low resilience in this study, such as advanced age, female gender, poor financial situation, and the presence of chronic conditions, and the consequences of low resilience levels that spans from food consumption to social engagement were known to be related to globalization-related economic and social inequalities [47]. The increasing gaps in social protection alongside the widening of inequalities across different socio-economic strata within and between countries have been revealed and exacerbated by the COVID-19 pandemic [48, 49]. Public health and social welfare systems may therefore need to be re-oriented with a joint focus to tackle health inequalities following the WHO’s Health in All Policies approach. This may warrant multi-sectorial efforts to ensure the timely and equitable delivery of appropriate, accessible, and affordable health and social care products to groups in socioeconomically vulnerable or marginalized circumstances who are more likely to suffer from negative mental health outcomes and poor lifestyle habits.