Anxiety and depressive symptoms among migrants during the COVID-19 pandemic in Norway: A two-wave longitudinal study

This study was designed to examine the impact of the COVID-19 pandemic on the mental health of migrants living in Norway. We conducted a longitudinal two-waves survey among a sample of 574 migrants and multilevel modelling was used to analyse anxiety, health anxiety and depressive symptoms. Demographic and psychological predictors were investigated. The levels of anxiety, health anxiety and depressive symptoms among migrants decreased from the lockdown (strict social distancing protocols) to phaseout. Reductions in maladaptive coping strategies were related to parallel reductions in anxiety, health anxiety, and depression, and a reduction in loneliness was related to a reduction in depression. The results indicate that the elevated levels of anxiety, health anxiety and depressive symptoms among migrants in the first phase of the pandemic may be temporary.


Introduction
The coronavirus disease 2019  pandemic has led to a public health emergency, thousands of deaths, generalized economic depression, unemployment, worldwide quarantines and national lockdowns (Jakovljevic et al., 2020;Sani et al., 2020;Tian et al., 2020;Zhang et al., 2020). While the impact of COVID-19 on mental health has been extensively documented in the general population (Kumar and Nayar, 2021;Khan et al., 2020;Xiong et al., 2020), important gaps remain in the literature regarding the specific consequences of the pandemic on particular population groups, such as migrants. Migrants in this sense is defined as persons born outside the residing country with legal residence (first generation migrants) or person born in the residing country with legal residence (second generation migrants). This also holds for previous epidemics and pandemics, where few studies have focused on the mental health symptoms of migrants in pandemic contexts. This is surprising given the size of this group worldwide. Almost 4% of the world's population lives outside its country of origin (United Nations Department of Economic and Social AffairsPopulation Division, 2020).
Several studies have shown that anxiety, depression and posttraumatic stress disorder tend to be more prevalent among these groups than the general population, even years after resettlement in a host country (Bogic et al., 2015). While the lack of properly disaggregated data makes it difficult to quantify their suffering, some evidence of migrants being disproportionately affected in terms of both disease transmission and outcomes and the consequences of mitigation measures is emerging . Migrants are overrepresented in COVID-19 laboratory-confirmed cases, hospital admissions, intensive care treatment and death statistics in all countries with available data (Aldridge et al., 2020;Erdal et al., 2020). This could suggest that COVID-19 lockdowns affect the mental health of migrants. However, so far, mainly cross-sectional studies have investigated the impacts of COVID-19-related lockdowns and limited longitudinal studies exist that investigate the mental health symptoms of migrants.
On 12 March the most radical measures since the Second World War, including school closure and travel bans, were imposed in Norway. In the beginning of the pandemic, the majority of identified COVID-19 cases were Norwegian-born tourists and their close contacts, most of them returning from skiing destinations in the Alps, and the prevalence of infection among migrants in Norway was low . Later, confirmed cases in migrants increased and have remained stable at 35-50%more than twice the proportion of the migrant population (18%) (SSB, 2020). It has been an unequal burden of disease in Norway, and knowledge about groups at especially high risk is important to direct efforts to prevent disease transmission, severe illness and death and to plan and adapt health services accordingly. This align with the call for action of gathering information regarding the impact of the COVID-19 pandemic on mental health and the risk and protective factors of vulnerable groups is an immediate multidisciplinary research priority (Holmes et al., 2020). The combination of age, gender, employment status, emotional support, worry/rumination and levels of preexisting mental health issues may influence the impact of the COVID-19 pandemic on migrants.
Drawing on theoretical and empirical considerations, several factors were included in the current research. Emotional support from others has been found to be protective during the COVID-19 pandemic (Bu et al., 2020;Groarke et al., 2020). This is important to study given the government-initiated physical distancing protocol which potentially decrease the opportunity of emotional support. Studies have indicated that for young migrants, family and community are protective resources that can promote psychosocial well-being (Merrill Weine et al., 2014). Against this background, face-to-face contact with friends and relatives within the limits of the distancing rules during this pandemic seems essential for this group and may serve as a protective mechanism fostering competence, support and behaviours that can promote their psychological health.
Other actionable factors, and thus potential targets of intervention, include worry and rumination. These factors are given central roles in the self-regulatory executive function model (S-REF) and derived from the metacognitive therapy model (Wells and Matthews, 1996). In this model, emotional problems and disorders are linked to the activation of a particular maladaptive style of thinking called the cognitive attentional syndrome (CAS). This involves a range of maladaptive strategies to regulate negative feelings and thoughts, including worry and rumination, attentionally focusing on a threat and unhelpful coping behaviours that backfire (e.g., thought suppression and situational avoidance).
CAS arises from metacognitive beliefs. Two categories of beliefs are important: (1) positive beliefs about the need to engage in aspects of the CAS (e.g., "Worrying helps me cope.") and (2) negative beliefs about the uncontrollability, dangerousness or importance of thoughts and feelings (e.g., "Some thoughts could make me lose my mind."). Metacognitive therapy based on the S-REF model has a solid evidence base in terms of outcomes (Normann and Morina, 2018) and mechanisms of change Hoffart et al., 2018). Worrying and rumination can keep someone in a negative and dissatisfied mental loop and lead to negative feelings, such as sadness, depression, anxiety and worrying both about one's overall health and the concern of the family during the present pandemic. One study found elevated worry and rumination level at the beginning of the first lockdown, then declined but increased when UK returned to lockdown and higher levels were associated with a fiveand ten-fold increase in clinically meaningful rates of depression and anxiety in the general population (O'Connor et al., 2022). Migrants may experience compounding worry that concern family members, many of whom remain living in conflict-affected or low-resource countries where the virus is spreading and where health care is poor or nonexistent and no being unable to return to the country of origin in an emergency to support or protect family members (Rees and Fisher, 2020). This worrying and rumination can contribute further to mental distress for this population; furthermore, metacognitive beliefs are crucial factors in the maintenance and development of mental disorders. However, this is understudied in this group.
Sociodemographic and health factors can explain why the COVID-19 pandemic affect people differently. In the general population women have been almost three times likely than men to report dysfunctional anxiety due to the COVID-19 pandemic (Caycho-Rodríguez et al., 2021), unemployed people during the COVID-19 pandemic showed greater mental health deterioration (Posel et al., 2021), having a psychiatric diagnosis predicts higher psychological distress (Qiu et al., 2020), moreover studies have indicated that younger people have reported more adverse psychological adversities, such as anxiety, depression and post-traumatic stress symptoms (Conversano et al., 2020;Qiu et al., 2020). Thus, being a woman, younger, unemployed, and having a psychiatric diagnosis have all been identified as risk factors for mental distress during the pandemic in the general population. We know that migrants are disproportionately affected and it has been suggested that migrants and refugees may be among the most vulnerable to the mental health effects of COVID-19 (Holmes et al., 2020;Júnior et al., 2020;Mukumbang et al., 2020;Rees and Fisher, 2020), however we do not know whether the empirically established factors in the general population hold for the migrant population. It could be differences between the migrant population and general population when it comes to e.g. gender roles and the status of the elderly. In addition, one study found that as a group, older female migrants had an elevated health vulnerability as such (Krobisch et al., 2021). This needs to be empirically investigated.
The purpose of this preregistered study was to examine the impact of the COVID-19 crisis by investigating mental health symptoms (i.e., anxiety, health anxiety and depression) among migrants in Norway at two points in time: 1) during the strict government-initiated physical distancing protocols in the first weeks of the pandemic (T1) and 2) to the period of lightened protocols three months into the pandemic (T2). Demographic factors were investigated to identify subgroups with increased levels of anxiety, health anxiety and depression symptoms. Predictors such as emotional support, feelings of loneliness and unhelpful coping strategies were investigated.
The hypotheses were as follows: H1. There will be a significant decrease in levels of anxiety symptoms, health anxiety symptoms and depressive symptoms from the baseline period with the strictest mitigation protocols in place, during the first weeks of the pandemic (T1) to the period of lightened protocols three months into the pandemic (T2).
H2. Higher levels at T1 and less reduction from T1 to T2 of unhelpful coping strategies and loneliness will predict less change in symptoms of anxiety, health anxiety and depression from T1 to T2, above and beyond the influence of demographic variables (i.e., having a preexisting mental health condition, age, gender and education).
H3. Higher levels at T1 and more increase from T1 to T2 of emotional support will predict more change in symptoms of anxiety, health anxiety and depression from T1 to T2, above and beyond the influence of demographic variables (i.e., having a preexisting mental health condition, age, gender and education).

Study design and recruitment procedure
Because of infection guidelines concerning viral transmission and the time-sensitive nature of a study, we could not disseminate the survey through conventional methods such as access to postal services. Participants were recruited through an online survey disseminated through broadcasting on the national news channel of Norway, national radio stations, regional and local radio stations across the country, national newspapers, and regional and local newspapers across the country, additional local and regional media, social media sources in addition to dissemination to a random selection of Norwegian adults through a Facebook Business algorithm. These national, regional, and local advertisements contained brief messages about the possibility to partake in a study at the University of Oslo about mental health in the adult population, in this sense this study comes from a large national survey not specifically aimed at migrants. The information was given in Norwegian. There were no incentives present in the recruitment procedure. This recruitment procedure is elaborated in detail elsewhere (Ebrahimi et al., 2021a,b). The data collection period for T1, lasted seven days between March 31, 2020 and April 7, 2020, a timeframe in which all social distancing protocols were held constant during the two weeks before data collection and during the data collection week. Participants who participated at T1 were invited to participate at T2. The data collection period continued for three weeks, from June 22, 2020 to July 13, 2020. A total of 574 participated at T1, and of these, 260 responded at T2.
Approval from the Regional Committee for Medical Research Ethics was received before the commencement of this study (reference number: 125510). Participants could terminate the survey at any time, with no consequences. The study was preregistered at ClinicalTrials.gov after data collection, but before any analysis (Identifier: NCT04443764) and is part of the Norwegian COVID-19 Mental Health and Adherence Project (MAP-19).

Participants
Eligible participants were individuals older than 18 years of age who had provided their consent to participate in the study and who were firstor second-generation migrants living in Norway. The population we have investigated are either persons born outside Norway but residing in Norway with legal residence (first generation migrants) or person born in Norway with legal residence (second generation migrants). There was no significant difference between these two groups on either of the four outcome variables, Health

Outcomes
The PHQ-9 (Kroenke et al., 2001) was used to the measure symptoms of depression in accordance with the diagnostic criteria for a major depressive disorder. The questionnaire consisted of nine items; each was scored on a 4-point Likert scale (0-3), with a range of scores from 0 to 27. Higher scores indicate greater depression severity, and scores above 10 are considered the cutoff that indicates whether a person probably had a depressive disorder. Cronbach's alpha ranged from 0.86 (T1) to 0.90 (T2).
The GAD-7 (Spitzer et al., 2006) is a questionnaire consisting of seven items that measure symptoms of anxiety and worry. For this current study, the items were scored on a 4-point Likert scale (0-3), with scores ranging from 0 to 21. The cutoff for the Norwegian samples was found, yielding a cutoff of 8 and above for high sensitivity and specificity (Johnson et al., 2019). Cronbach's alpha ranged from 0.88 (T1) to 0.89 (T2).
The symptoms of health anxiety were measured using two items from the validated Health Anxiety Inventory (HAI) (Salkovskis et al., 2002) -Item 1: "I constantly have images of myself being ill" and Item 6: "I spend much of my time worrying about my health"-as well as an item measuring specific fear of being infected with coronavirus and an item measuring fear of dying from the coronavirus The items were scored on a 4-point Likert scale (0-3) asking about symptomatology for the last 2 weeks. Cronbach's alpha ranged from 0.75 (T1) to 0.78 (T2).

Predictors
The variables derived from the metacognitive model were measured using the Cognitive Attentional Syndrome-1 Questionnaire (CAS-1) (Wells, 2009). The CAS-1 questionnaires consist of 16 items. The first two, assessed on a scale from 0 to 8, are questions concerning the frequency of rumination and worry as well as concentration on threats. A further six items, assessed on a scale from 0 to 8, concern maladaptive behaviors used to cope with negative emotions and/or thoughts, e.g. thought and situation avoidance, drinking or substance abuse, and attempts to control thoughts or emotions. These items are the variable "unhelpful coping strategies" in the analysis. The last eight items are assessed on a scale from 0 to 100, four items that measure positive metacognitive beliefs (e.g. "worrying helps me cope") and four items measure negative metacognitive beliefs (e.g. "worrying too much could harm me"). CAS-1 has been proven to have satisfactory psychometric properties (Nordahl and Wells, 2019). Cronbach's alpha ranged from 0.88 (T1) to 0.90 (T2) on coping strategies, from 0.57 (T1) to 0.65 (T2) on positive metacognition and from 0.66 (T1) to 0.74 (T2) on negative metacognition (T2).
Emotional support was measured by three items: "I have sufficient social support (from fellow peers and others close to me); " "I feel emotionally close to the people whom I care about; " and "I feel that others close to me care about me and my well-being." The item was scored on a four-point Likert scale ranging from do not at all agree (0) to almost every day (4). Thus, higher scores indicate greater social support. Cronbach's alpha ranged from 0.75 (T1) to 0.87 (T2).

Statistical analysis
Repeated surveys, like the one used in this present study, typically have missing data. Given such missing data, full information maximum likelihood estimation (FIML) was utilised. FIML allows individuals to contribute with the proportion of data they have available in the estimation procedure. As a state-of-the-art approach in scenarios with missing data (Schafer and Graham, 2002), FIML yields more unbiased results than the other analytic methods (O'Connell et al., 2017).
In preliminary analyses and for each of the dependent variables (i.e., PHQ-9, GAD-7, and HAI), the combination of random effects and the covariance structure of residuals that gave the best fit for the "empty" model (the model without fixed predictors except the intercept) was chosen. The Akaike information criterion (AIC) was used to compare the fit of the different models. Models that resulted in a reduction in AIC greater than two were considered better (Burnham and Anderson, 2004). SPSS 27.0 was used for data analysis (IBM Corp., 2018).
First, H1 a decrease in mental health symptoms was tested by using GAD-7, PHQ-9 and HAI as dependent variables in a model using time (T1 period ¼ 0, T2 period ¼ 1) as a predictor. A model with a random intercept and a diagonal covariance structure turned out to have the best fit for all the dependent variables. Second, demographic-group variables were added as predictors. Only the demographic variables found to be significant were included in subsequent analyses. Third, the initial (T1) levels of loneliness, emotional support, coping strategies, positive metacognition and negative metacognition were added, together with the interactions of these constant covariates with time. These interactions represent tests of H2 about the covariates predicting changes in mental health symptoms, (i.e., Model 1).
Finally, the T2 levels of loneliness, emotional support, coping strategies, positive metacognition and negative metacognition as constant covariates were added, together with the interactions of these constant covariates with time. These interactions represent tests of H2 about the change in the covariates from T1 to T2, predicting a change in mental health symptoms from T1 to T2 (i.e., Model 2).

Demographic characteristics
Of the 10,061 respondents at T1, 574 met the study criteria. Of these, 365 (63.6%) were second-generation migrants. Of the 574 participants, 454 (79.1%) were women, 352 (61.32%) had higher education from a university, 296 (51.7%), were not married or in a civil union, 148 (25.8%) did not have children and 488 (85.0%) did not have a psychiatric diagnosis as shown in Table 1. Table 2 shows the mean levels of anxiety, health anxiety and depression for the entire sample and in each subgroup.
The mean level of depression was 8.44 (SD ¼ 5.78), anxiety 6.47 (SD Regarding the reported symptoms of depression (PHQ-9: cutoff of 10), 36.4% had symptoms meeting the clinical cutoff for probable depression at T1, compared to 13.8% at T2. Also, at T1, 34.8% had symptoms meeting the clinical cutoff for generalized anxiety disorder (GAD-7: cutoff of 8), compared to 12.2% at T2.
In the first model, the demographic variables of age, gender, having higher education and a psychiatric disorder were entered as predictors.

Discussion
This present study investigated the symptoms of anxiety, health anxiety and depression among migrants and changes in these symptoms during the strict government-initiated physical distancing protocols in the first weeks of the pandemic (T1) and to the period of lightened protocols three months into the pandemic (T2). As expected, the levels of anxiety, health anxiety and depression significantly decreased from T1 to T2, indicating that the overall symptoms among the migrants declined after the phaseout of the distancing protocols. Reduction of maladaptive coping strategies from T1 to T2 was associated with a reduction in anxiety, health anxiety and depression, and reductions of loneliness were associated with reductions of depression.
The data on mental health problems in the present sample of migrants are aligned with those of other studies of the general population that have examined the mental health impacts of the COVID-19 pandemic and related lockdowns, using the same measures employed in this study. This body of literature indicates that 18-31% of the general population report anxiety symptoms (GAD) and 17-28% report depressive symptoms (PHQ) beyond the threshold (China: Wang et al., 2020;Ireland: Hyland et al., 2020;Italy: Rossi et al., 2020;Norway: Ebrahimi et al., 2021a,b). One methodological stringent meta-analysis reviewing sixty-five longitudinal cohort predominantly European and Nort American studies displayed a statistically small overall increase in mental health symptoms show similar results from multiple countries (Robinson et al., 2022). Santomauro et al. (2021) estimated a significant increase in the prevalence of both major depressive disorder (with an estimated additional 53⋅2 million [95% uncertainty interval 44⋅8-62⋅9] cases worldwide-ie, a 27⋅6% [25⋅1-30⋅3] increase) and anxiety disorders (76⋅2 million [64⋅3-90⋅6] additional cases-ie, a 25⋅6% [23⋅2-28⋅0] increase) since before the pandemic. The authors concluded that there is important to strengthen mental health systems in most countries.
However, compared to other studies that have investigated anxiety (GAD) and depression (PHQ) among migrants, our sample had lower levels of mental health problems during pandemic time than other migrant samples during nonpandemic time. Several studies have found that the prevalence of depressive symptoms is three to five times higher than in the general population (Kurt et al., 2021;Leiler et al., 2019). There were a few unemployed participants (13%) and a high prevalence of higher-educated participants (61%) in our sample, compared to our study of the Norwegian population reporting the prevalence of anxiety and depression during the COVID-19 pandemic (Ebrahimi et al., 2021a, b). In that study there were 19% unemployed and 56% higher-educated participants. Thus, one possible explanation for this finding is participants in this current sample are less affected, have higher access to health care services and have fewer language barriers than migrants residing in the migration agency's housing facilities.
During the pandemic, Norwegian governments implemented campaigns translating vaccination-and pandemic-related information into the native languages of migrant groups. A previous study on Norwegian adults found no empirical support for the notion that immigrants display more scepticism towards vaccination compared to natives of Norway Table 4 Fixed effects estimates (top) and variance-covariance estimates (bottom) for models of predictors of anxiety (GAD-7). Note: T1 ¼ a period of one week (31 March-7 April 2020), starting nearly three weeks after the implementation of strict social distancing protocols in Norway (March 12, 2020). T2 ¼ a period of three weeks (22 June-13 July 2020), starting one week after the strict social distancing protocols were discontinued (June 15, 2020). GAD-7 ¼ generalized anxiety disorder. 1 ¼ T1; 2 ¼ T2; 3 ¼ Wald Z. *p < 0.05 **p < 0.01 ***p < 0.001.
K. Vrabel et al. Psychiatry Research Communications 3 (2023) 100115 (Ebrahimi et al., 2021a,b), in fact one study found that the majority of migrants reported high levels of trust in the Norwegian government and health authorities (Madar and Benavente, 2022).
The level of symptoms of anxiety, health anxiety and depression were reduced from T1 to T2, which may appear to be a function of several factors associated with the repeal of stringent lockdown restrictions. Physical distancing was one of the critical measures employed to limit the spread of COVID-19. However, physical distancing can have consequences for mental health and well-being in both the short and long term (Druss, 2020). Since migrants tend to have poor social support, physical distancing may significantly affect their mental health (Kirmayer et al., 2011). In this current sample, there was a high prevalence of employment. Various studies have found a significant protective effect of employment on mental health in general and particularly on depression and psychological distress (Masten and Obradovic, 2008). Employment has been found to be associated with improved self-esteem, greater well-being, increased social contact and independence, which contribute to good mental health (van der Noordt et al., 2014). This may have affected the migrants in this study since they had high prevalence of employment both in the first phase and three months into the pandemic.
Regarding our hypotheses about changes of predictors parallel to changes in anxiety, health anxiety and depression, only the association between coping strategies and mental health symptoms was supported in the context of the other predictors. A greater reduction in unhelpful coping strategies was associated with a greater reduction in anxiety, health anxiety and depression. These findings are consistent with studies identifying that ineffective coping strategies are associated with poorer mental health (Gurvich et al., 2020;Stanisławski, 2019) and studies revealing that the reliance on maladaptive coping strategies (e.g., alcohol consumption) having increased during the pandemic (e.g., Capasso et al., 2021;Taylor et al., 2021). The associations between maladaptive coping strategies and the trajectory of psychopathological symptoms are of particular concern. It may be that reduction in coping strategies may alleviate anxiety and depression, but studies with more time points are needed to address this question.
One unexpected finding emerged. More use of maladaptive coping strategies and more loneliness at T1 predicted a greater reduction in depression, anxiety and health anxiety. In our hypothesis 2 we expected that more use unhelpful coping strategies and more loneliness at T1 would predict less change in symptoms of anxiety, health anxiety and depression, in that sense this hypothesis was not met and a more counterintuitive findings emerged. This may be due to regression to the mean, reflecting that those with the highest scores on coping strategies, loneliness and depression at T1 were prone to approach towards the mean at T2. Alternatively, a recent analysis of individual change patterns of depression and anxiety in the total sample shows high variability of changes, indicating one sub-group who started high and normalized at T2, and another subgroup who started at a normal level but had increased substantially at T2 (Ebrahimi et al., 2021a,b). Further analyses of sub-groups could reveal more nuanced relationships between initial maladaptive coping and symptom change, but loss of power prevents us from doing this.

Strengths and limitations
A major strength of this study is that it captured the detrimental associations with government-initiated distancing protocols which are present across countries, making its findings generalisable across similar cultures employing similar distancing protocols. Another strength of this study is the large sample of the same individuals -migrants -who experienced identical interventions across the two measurement periods. We were interested in examining mental health in a whole subsample of persons with migrant background, this means that there are probably persons that were migrants in early childhood and adulthood in our sample. These groups may differ on experiences; however, we did not have the possibility to contrast these groups. In addition, the sample largely comprised well-educated female participants and a small sample size of men, which may indicate a bias in the sample. Another limitation of this study is that it was based on self-reported measures rather than clinician-administered interviews and full random sampling was not conducted, due to the urgency of the T1 data collection. Thus, those who chose to respond may have specific features that may affect the results. Major steps were taken to reduce this possible bias through the recruitment of participants across a variety of platforms which were more accessible to all age groups. Specifically, the sampling strategy aimed at reaching immigrants through national media, however this precludes an Note. Note: T1 ¼ a period of one week (31 March-7 April 2020), starting nearly three weeks after the implementation of strict social distancing protocols in Norway (March 12, 2020). T2 ¼ a period of three weeks (22 June-13 July 2020), starting one week after the strict social distancing protocols were discontinued (June 15, 2020). HAI ¼ Health Anxiety Inventory. 1 ¼ T1; 2 ¼ T2; 3 ¼ Wald Z. *p < 0.05 **p < 0.01 ***p < 0.001.
overview of how many people have may have seen the advertisement to participate in the study versus how many actually participated. There was a lot of missing data at T2. This was addressed by analyzing the data with maximum likelihood estimation, which gives less biased results if the data are missing at random.

Conclusion
The elevated levels of anxiety, health anxiety and depressive symptoms among migrants in the first phase of the pandemic may be temporary. Reductions in maladaptive coping strategies were related to parallel reductions in anxiety and health anxiety, and reductions in loneliness were related to reductions in depression.

Data availability
The ethical approval granted by the Regional Committees for Medical and Health Research Ethics in Norway (REK and NSD) and in accordance with the information given to the participants in the informed consent form regarding the use of data, does not allow the authors to submit the data to a public repository. In line with the ethics approval, the data are to be kept at a secure server only accessible by the authors of this study. The data are stored at the TSD-system, which is part of the long-term storage facility at the University of Oslo. TSD uses regular back-up thus the data is well secured. Data will be stored at least five years in accordance with the ethical approval granted by the Regional Committees for Medical and Health Research Ethics in Norway. Access to the data can be granted following ethical approval of suggested project plan for the use of data from NSD and REK. The data will then be anonymized and further stored at the TSD-system. Such requests are to be sent to Associate Professor, KariAnne Vrabel, Institute of Psychology, University of Oslo, email:karianne.vrabel@modum-bad.no; Associate Professor, Sverre Urnes Johnson, Department of Psychology, University of Oslo, email: s.u.johnson@psykologi.uio.no or to psychologist Omid V. Ebrahimi, email: omid.ebrahimi@psykologi.uio.no. Although the authors cannot make their study's data publicly available at the time of publication, all authors commit to make the data underlying the findings described in this study fully available without restriction to those who request the data, in compliance with the Psychiatry Research Data Availability policy. For data sets involving personally identifiable information or other sensitive data, data sharing is contingent on the data being handled appropriately by the data requester and in accordance with all applicable local requirements.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.