Social connectedness and resilience post COVID-19 pandemic: Buffering against trauma, stress, and psychosis

The present study investigated psychosocial predictors of psychosis-risk, depression, anxiety, and stress in Croatia two years after the onset of the COVID-19 pandemic. Given the existing transgenerational war trauma and associated psychiatric consequences in Croatian population, a significant pandemic-related deterioration of mental health was expected. Recent studies suggest that after an initial increase in psychiatric disorders during the pandemic in Croatia, depression, stress, and anxiety rapidly declined. These findings highlight the role of social connectedness and resilience in the face of the global pandemic. We examined resilience and psychiatric disorder risk in 377 Croatian adults using an anonymous online mental health survey. Results indicate that there was an exacerbation of all mental ill health variables, including depression, anxiety, stress, and a doubled risk for psychosis outcome post-COVID pandemic. Stress decreased levels of resilience, however, those exposed to previous traumatic experience and greater social connectedness had higher resilience levels. These findings suggest that individual differences in underlying stress sensitization of Croatian population due to past trauma may continue to influence mental health consequences two years after COVID-19 pandemic. It is essential to promote the importance of social connectedness and resilience in preventing the development of variety of mental health disorders.


Introduction
The COVID-19 pandemic has severely disrupted every aspect of daily life, resulting in countless economic, social, and behavioral changes. Two years after the outbreak of the COVID-19 pandemic, research has primarily focused on elucidating the effects of the first pandemic wave on general and mental health (Orfei et al., 2022;Taylor et al., 2021;Penninx et al., 2022). Numerous studies have reported substantial increases in psychiatric morbidity, including anxiety, depression, insomnia, and post-traumatic stress disorder (PTSD) (Wang et al., 2020;Goldberg et al., 2022;Raina et al., 2021), as well as a dramatic increase in loneliness and psychosis prevalence (Carvalho et al., 2020), which has further elucidated the role of loneliness as a significant and important risk factor for psychosis-risk (Gizdic et al., 2022;Tso and Park, 2020). The development of psychotic symptoms, depression, stress, and anxiety symptoms in individuals with no history of psychiatric disorders is supported by evidence indicating an increased incidence of first-case psychopathology in COVID-19 patients (Taquet et al., 2021;Desai et al., 2021;Cao et al., 2022). One year post-pandemic, an increased prevalence of fatigue, sleep problems, memory loss, and concentration difficulties was reported globally (Boscolo-Rizzo et al., 2021;Liu et al., 2022;Han et al., 2022), as well as a persistent increase in anxiety, stress, and depression (Lakhan et al., 2020;Brooks et al., 2020;Shah et al., 2021;Joshi et al., 2021;Pierce et al., 2020;Meaklim et al., 2023), and psychotic-like symptoms (Lim et al., 2020;Taquet et al., 2021;Brown et al., 2020;Wu et al., 2021). Even after the lockdown restrictions were eased, general physical and mental health has deteriorated since the beginning of COVID-19 (Patel et al., 2022;Vadivel et al., 2021).
Furthermore, a number of studies across the globe have also identified a variety of risk factors for psychosis during the COVID-19 pandemic, including younger age, female gender identity, unemployment, loneliness, and a history of trauma (Tso and Park, 2020;Dean et al., 2021;Bauer et al., 2021;Lee et al., 2021;Proto and Quintana-Domeque, 2021). Previous exposure to trauma, in particular, is predicted to increase the prevalence of psychopathology and mental disorders during COVID-19 (Gizdic et al., 2022), given that trauma is widely predictive of nearly all subclinical and clinical psychopathology and negative outcomes (Lu et al., 2013;Aux em ery, 2012). Although the intensity of trauma triggers may diminish over time (Howell et al., 2015), it is important to consider their continued association with poor wellbeing, particularly among Croatians who have experienced war and natural disasters (e.g., earthquakes). War-related trauma and post-traumatic reminders have had a devastating and lasting effect on the mental health and quality of life of this population (Babi c-Banaszak et al., 2002;Vukojevi c et al., 2020;Jefti c et al., 2021). Specifically, the first wave of the COVID-19 pandemic increased the prevalence of nearly all psychopathology symptoms in Croatia (Joki c Begi c et al., 2020;Gizdic et al., 2022). Surprisingly, these rates decreased after a few months (from May to July 2020), when restrictions were partially relaxed (Ajdukovi c et al., 2020). In comparison to other European countries (and parts of Asia), Croatia seemed to have a relatively lower incidence of depression, stress, and anxiety (Newby et al., 2020;Park et al., 2020;Rossi et al., 2020). Although continuous increases in symptoms were anticipated through 2020, these results are suggestive of high levels of resilience and adaptability among this population throughout the pandemic.
As a result, researchers have examined the concept of resilience and discovered that it may serve as a protective factor not only against trauma exposures but also against the development of psychopathology symptoms (Pietrzak et al., 2011), despite the fact that patterns of vulnerability levels vary among individuals (Sominsky et al., 2020). Psychological resilience is an active, process-oriented defense mechanism that appears to be derived in part from having meaningful, supportive, and functional social networks. For instance, individuals with a higher degree of social connectedness and a lower level of loneliness tend to have a higher level of general wellbeing and are better protected against mental health issues during the COVID-19 pandemic Killgore et al., 2020;Groarke et al., 2020). However, social connectedness appears to be an especially important protective factor among trauma survivors. In some cases, individuals within a traumatized group become more resilient to adversity to the extent that their functioning is sometimes enhanced following exposure to adversity (Ayed et al., 2019;Finstad et al., 2021). During stressful and uncertain events, for instance, people tend to imitate the behavior and emotions of those around them (Duan et al., 2019), indicating that there is a collective social impulse that protects us. The 2020 study by Vukojevi c et al. suggested that, when people are together in a catastrophic situation, the catchphrase "we are in this together" has a deeper meaning due to the protective effect of crowd influence on our psyche. A possible explanation can be found in the Croatian experience of war (as well as the recent earthquake). Shared pain during a shared experience of disaster can unite people and inspire them to help each other, which promotes solidarity and increases social resilience, ultimately resulting in better mental health outcomes (Bastian et al., 2014;Garcia and Rime, 2019). According to Bastian et al. (2014) study, shared pain can increase cooperation and social bonding by acting as "social glue." In this study, we investigated a) psychosocial predictors of general and mental health in the Croatian population two years after the COVID-19 pandemic, and b) the role of mental health status and social connectedness in influencing resilience among Croatian individuals. We hypothesized based on our previous research (Gizdic et al., 2022), mental health symptomatology and social connectedness would play a significant role in resilience levels, particularly among those who had experienced trauma in the past.

Participants and procedures
All participants were Croatian adults (aged 18 and above) who completed an online survey created via Survey Monkey in Croatian that was distributed via online platforms and channels (such as the university emailing lists, social media platforms etc.) and in person. Before starting the survey, participants were informed of the study goals and aims, introduced to the type of questions and amount of time for completion of the study, as well as their ability to stop at any time. Participation was anonymous, voluntary, and open to everyone aged 18 and up (detailed in Gizdic et al., 2022). The survey took an average of 17 min to complete (74%). Data collection occurred between February and May 2022, following two peak waves of the COVID-19 pandemic (1.7 years after data collection during the first wave; survey 1 ran from July to September 2020; Gizdic et al., 2022). This study received exempt status from the Vanderbilt University Institutional Review Board (Vanderbilt IRB exempt #200337).

Measures
Following the previous survey (Gizdic et al., 2022), we repeated the same patterns of questions with slight modifications and addition of new scales. The present survey consisted of 159 questions regarding participant demographics, COVID-19 concern, general and mental health, including well-validated mental health measures of depression, anxiety, stress, psychosis, social connectedness, and social isolation. We also inquired about COVID-19 vaccination hesitancy, resilience, and exposure to trauma.
General information regarding COVID-19 diagnosis, concern, vaccination, and dosage, was requested to examine the overall effects of the pandemic on participants' daily lives. Ratings were given with appropriate responses to each item (i.e., for level of COVID concern, questions were scored on a 4-point Likert scale, ranging from 0¼not at all concerned to 4¼extremely concerned). Participants self-reported the changes in their current living situation, employment, number of days feeling positive emotions (love, happiness, and hope), as well as changes in their general health. To better understand previous trauma exposures, we included questions asking about adversity in childhood-emotional, physical, and sexual abuse; emotional and physical neglect; and included the Brief Trauma Questionnaire (BTQ; Schnurr et al., 2002)-a 10-item, self-report questionnaire that asks general trauma questions (e.g., Have you ever been in an active war zone or served in a job that exposed you to war-related casualties?) with follow up questions rating the severity of each traumatic event endorsed (e.g., If so, did you think your life was in danger or were you possibly seriously injured?).
The Short Scale for Measuring Loneliness (the UCLA Loneliness-short; Hughes et al., 2004) was used to assess subjective feelings of loneliness and social isolation; the Social Network Index (SNI; Cohen, 1997) was used to assess social connectedness including social network quality, size, and diversity (e.g., number of social high contact roles, embedded social networks, and regular people contacts); depression, anxiety, and stress subscales was assessed with Depression, Anxiety and Stress Scale -21-item version (DASS-21; Lovibond and Lovibond, 1995), and psychosis risk and distress was assessed with the Prodromal Questionnaire-16 (PQ-16; Ising et al., 2012).
We also included a measure of vaccination hesitancy (adult Vaccine Hesitancy Scale, aVHS; Akel et al., 2022) asking participants about their own hesitancy and perceptions of effectiveness, reliability, and potential risks of vaccinations (e.g., Vaccines are important to my health). The responses were rated on a 5-point Likert scale ranging from strongly disagree to strongly agree. We added the 4-item Brief Resilient Coping Scale (BRCS; Sinclair and Wallston, 2004), which assesses participants' levels of resilience (i.e., successful recovery from stressful situations). vaccination hesitancy, resilience, and social networks, in determining general and mental health. In the first step, independent variables for age, gender, social distancing adherence, childhood abuse and neglect, general traumas, and COVID-19 concern were used to form the basic model. In the second step, the full model included social network diversity, size, embedded social networks, loneliness, vaccination hesitancy, and resilience. For each dependent variable (e.g., self-reported general health, days feeling happy, feeling hopefully, and loving, DASS depression, stress, and anxiety, and PQ-scores), the change in R 2 between the basic model and full model was used to examine whether adding social network variables, loneliness, and/or vaccination hesitancy and resilience explained more variance after controlling for age, gender, trauma, social distancing, and COVID concern.
To achieve the second goal of the study, we again tested the relationships between resilience, trauma, social networks, loneliness, vaccination, and mental health variables. However, to gain further clarity on the role and directionality of resilience as a factor in wellbeing, we repeated the regression analysis but used resilience as a dependent variable to examine whether psychosocial variables predict the levels of resilience. In the first step, the same independent variables as in the previous model were entered (e.g., age, gender, etc.) as a basic model. In the second step, social network diversity, social network size, and embedded social networks, as well as loneliness, vaccination hesitancy, DASS scales, and PQ total and distress, were included in the full model. After controlling for age, gender, traumas, social distancing, and COVID concern, the change in R2 between models was used to determine if adding social network variables, loneliness, vaccination hesitancy, DASS scales, and PQ total and distress explained more variance in resilience. A Bonferroni correction of p < 0.0045 was applied to both analyses to minimize Type I Errors.

Results
A total of 377 Croatian adults (78% females; mean age ¼ 29.2, SD ¼ 12.31) participated in the study. Table 1 displays descriptive statistics for all study variables. Two years after the pandemic, participants reported overall good general health (42%), but they were still concerned with the pandemic (54%). 57% of participants received a COVID-19 vaccination, with a slight decrease in average general health before (mean ¼ 2.31; SD ¼ 1.02) and after vaccination (mean ¼ 2.29; SD ¼ 1.24). Overall, participants reported a relatively high number of days when they felt love, happiness, or hope ( Table 2). Questions assessing social connectedness, levels of loneliness and social isolation, and resilience were completed by approximately 83-86% of participants, whereas the DASS was completed by 79% of participants and the PQ-16 was completed by 77% of participants, of whom 28% were at high risk for psychosis (see Table 3).
In the first goal of the study, we examined the psychosocial predictors of health two years after the first wave of the pandemic. Concern with COVID, childhood abuse and neglect, general trauma, loneliness, and vaccination hesitancy were negatively associated with general health status. On the other hand, SNI embedded social network and resilience were positively associated with overall general health. COVID concern, age, and loneliness decreased the number of days when participants felt happy. Loneliness decreased the number of days when the participants felt hopeful and loving. SNI embedded social network and resilience both increased the number of days feeling happy and hopeful. Social distancing, SNI high contact role, and resilience all increased the number of days when participants felt love. Furthermore, age and resilience were found to be negatively associated with DASS depression, DASS stress, and DASS anxiety. Childhood abuse and neglect, loneliness, and COVID concern were found to be positively associated with DASS stress. Only childhood abuse and neglect and loneliness were linked to depression and anxiety symptoms from DASS. With respect to psychosis-risk, there was a negative relationship between age and psychosis symptoms and related distress. In contrast, there was a positive association between psychosis symptoms (PQ-16 score) and the following: childhood abuse and neglect, loneliness, and vaccination hesitancy. Similarly, there was a positive association between childhood abuse and neglect, loneliness with levels of distress surrounding psychosis symptoms (PQ-16 distress). These findings suggest that greater vaccination hesitancy, childhood abuse and neglect, and increased loneliness all contribute to an increased risk of psychosis (Table 4). A second set of analyses revealed that stress was negatively associated with levels of resilience, whereas general trauma and SNI high contact role were positively associated with resilience (Table 5). As such, although stress decreased levels of resilience, those with previous exposure to general trauma and greater social connectedness (i.e., a high number of people in their social network) had increased resilience levels.

Discussion
The present study sought to investigate the long-term mental health consequences of the COVID-19 pandemic in the Croatian population, emphasizing the importance of psychosocial factors in determining mental wellbeing. We specifically highlighted the effects of previous traumatic experience and the important role of social connectedness in resilience-a particularly relevant topic for the Croatian population given the country's previous transgenerational war trauma and natural disasters. Although most participants reported good general health and an increase in the number of days, they felt positive emotions compared to our previous study (Gizdic et al., 2022), the current findings show that people are still concerned about the COVID-19 pandemic, even two years after the first wave. There was a higher level of vaccination hesitancy within this population. Nonetheless, according to Think Global Health (2021), the average vaccination rate in the European Union (EU) is 65%, and in comparison, to Croatia, Bulgaria, for example, had only 22% of its population vaccinated and a very high death rate. Thus, this may appear to be a matter applicable to the global population rather than Croatia in particular.
Concern about COVID, vaccination hesitancy, but also past trauma and increased loneliness post-pandemic may have contributed to a decline in overall general health. In turn, the number of embedded social networks (i.e., the number of different network domains in which a participant is active) and resilience levels led to better general health and more days when participants felt happy and hopeful. The number of days participants felt love increased with social network diversity (i.e., the number of people with whom the participant has regular contact), resilience, and, unexpectedly, with social distancing adherence. Social distancing measures have been put in place throughout the pandemic to curb the spread of the COVID-19 virus. In many places, social distancing is seen as a pro-social behavior-one that protects the community from COVID-19 (Wider et al., 2022). Evidence suggests that widespread experience of hardship or pain increases cooperation, collaboration, and social bonding (Bastian et al., 2014). Given the history of shared trauma experienced by the Croatian population (e.g., war, earthquake), it is possible that increased social distancing adherence is viewed as extremely pro-social, collaborative, and benevolent behavior, thereby increasing feelings of love in participants' daily lives. Furthermore, Croatian social contacts are relatively reserved; for example, culturally normative public interpersonal greetings do not typically involve physical contact (such as hugging or kissing). As a result, it is possible that adherence to social distancing conforms to Croatian social norms and expectations and may involve less significant change in daily life routines than other aspects of the pandemic.
Overall, our findings are consistent with previous research on the detrimental and enduring effects of the COVID-19 pandemic on mental health, indicating that the pandemic and related social isolation, as well as past trauma, continue to have a large and pervasive impact on individual wellbeing (Patel et al., 2022;Vadivel et al., 2021). However, our results also highlight social connectedness (i.e., social network domains) and resilience as promising protective factors in preventing the further development of unfavorable mental and general health outcomes.
In addressing the first aim of the study, we noted a drastic increase in the prevalence of stress, depression, and anxiety symptoms. Surprisingly, the rate of psychosis risk post-pandemic nearly doubled when compared to the prevalence of high-risk psychosis rates at the beginning of the pandemic (Gizdic et al., 2022). These findings reflect the impact of ongoing and continued stress caused by the COVID-19 pandemic and suggest the pandemic will continue to have long-lasting consequences on individuals' functioning and wellbeing (Goldberg et al., 2022). These results are also supported by previous findings from the Severe Acute Respiratory Syndrome (SARS) outbreak in 2003, which showed that almost 82% of SARS survivors continued to experience poor mental health and related outcomes, including stress disorders such as PTSD (Mak et al., 2010). While the effects of viral pandemics on stress around the world are clear, it is also important to consider the nuances of populations with high exposure to adversity, such as the Croatian population.
Early adversity has been shown to leave neurobiological vulnerabilities that make individuals more sensitive to future stress (Read et al., 2014;Crist obal-Narv aez et al., 2016;Russell et al., 2018;Smith and Pollak, 2020), thereby increasing the risk of developing anxiety disorders, depression, and other broad dimensions of psychopathology     (Vaessen et al., 2017;Stroud, 2020;Wade et al., 2019). The stress sensitization model sheds light on the link between stress and the prevalence of affective disorders (Post, 1992;Stroud, 2020). According to this model and considering previous war-and natural disaster-related trauma, the Croatian population would be expected to be more sensitive to the changes caused by the pandemic relative to other populations. As a result, it appears that childhood adversity and subsequent stress exposure, such as the COVID pandemic, exacerbated depression, anxiety, and stress, particularly psychosis symptoms. These findings may add to the evidence of an underlying mechanism of increased stress-sensitivity. Contrary to our expectations, resilience had no effect on levels of psychosis, but it did lead to a decrease in stress levels. Thus, it may be plausible to think that building on resilience levels would lead to decreased stress levels and an amelioration of sensitivity to further stress. Loneliness is another important factor to consider in this interplay. Following our previous findings (Gizdic et al., 2022), this psychosocial factor remains a highly important risk factor in predicting a variety of symptom developments, particularly psychosis, even after two years of the pandemic, while social connectedness appears to serve as both a preventive and protective factor. As a result, strengthening social networks may have plausible effects on alleviating psychopathology symptoms, reducing levels of loneliness, and protecting against future stress.
There are several limitations to our study. First, despite the relatively large sample size, many participants did not complete the entire questionnaire, resulting in a smaller sample size for some of the measures (e.g., PQ-16 and DASS). Second, the majority of the sample consisted primarily of female participants, which may have limited its generalizability. Regardless, the study enabled a comprehensive investigation of multiple psychosocial predictors of psychopathology and psychosis-risk following two years after the COVID-19 pandemic, with evidence of long-term adverse effects of the pandemic and highlighting the significance of resilience and social connectedness.
To conclude, investigating the long-term mental health consequences of the COVID-19 pandemic and emphasizing the importance of psychosocial factors on mental wellbeing may help further detect the potential underlying mechanism of stress-sensitivity. We specifically highlighted the effects of previous traumatic experiences as well as the critical role of social connectedness in association to levels of resilience. Therefore, to mitigate the mental health consequences of large-scale traumatic events such as the pandemic in the future, it would be crucial to implement public health strategies that enhance and support social connectedness and resilience, especially for psychosis-a particularly relevant topic for the Croatian population given the lack of prodromal data and the country's history of exposure to transgenerational war trauma (including early exposures) and natural disasters.

Funding sources
Gertrude Conaway Vanderbilt Endowment fund partly supported this project.

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