The Swedish Police Service’s deportations of unaccompanied, asylum-seeking refugee children: The role of coping and general mental health

Abstract The number of unaccompanied, asylum-seeking refugee children (UARC) coming to Sweden has increased 100 times during the last ten years. If children do not voluntarily return, the police are responsible for deportation. This study aims to describe police officers’ coping in the deportation of UARC and to investigate the associations between coping and general mental health in relation to the deportation of UARC among police officers by considering sociodemographic variables. Validity and reliability analyses were conducted for the use of the Ways of Coping Questionnaire (WOCQ) in the context of UARC. Mental health was assessed using the General Health Questionnaire 12 (GHQ-12). The GHQ-12 was the outcome (dependent variable), and the experience of deportations of UARC, the WOCQ and sociodemographic variables were the independent variables. A 5-factor structure with some important similarities and differences to the original version was confirmed. This study shows that the police officers utilize different coping strategies in the same complex situations during the deportation of UARC. The use of escape-avoidance and self-control increased the likelihood of psychological disturbance, whereas positive reappraisal had a protective effect. This research also demonstrates that coping strategies have a moderating effect on general mental health in police officers’ work with the deportations of UARC.


PUBLIC INTEREST STATEMENT
During the past ten years, the number of unaccompanied refugee children seeking asylum in Sweden has increased 100 times. Approximately 20 per cent of the children are denied asylum. If the child refuses to return to the country of origin, the police are responsible for the deportation. In deportations, the police officers must manage a traumatized child who is under stress. Encountering such children might have an influence on police officers' general mental health; and, the police officers' general mental health might have an influence on how the officers cope with deportations. In the context of the deportation, the officers utilize both emotional and problem-solving coping. Avoidant and self-controlling coping were associated with poor mental health, whereas positive reappraisal was associated with good mental health. More knowledge about police officers' coping might help the police to treat the children with kindness and decency, and listen to their needs.

Introduction
The number of unaccompanied, asylum-seeking refugee children (UARC) coming to Sweden has increased dramatically. For example, approximately 100 times more children arrived during 2015 than ten years earlier (The Swedish Migration Board, 2016). Approximately 20% of these children are denied a residence permit and are sent back to their country of origin (The Swedish Migration Board, 2015). If children do not voluntarily return, the police are responsible for deportation. The deportation of refugees should be performed with respect for their human rights and dignity in compliance with national and international law ( /115/EC, D, 2008SFS 2005SFS :716, 2005).
In the current situation, police officers must manage a child who is under stress. These children are often traumatized during their escape to Sweden or from their experiences before their escape (UNHCR, 2007(UNHCR, , 2017. In addition, many of them have had negative experiences with police officers in other countries on their way to Sweden and have been subjected to harassment and brutality (Vesterdahl, 2016). Studies have indicated that when the final decision is received, children worry about what will happen when they return to their country of origin (Kullander, Tönnes Lönnroos, Verständig, & Viblemo, 2016;Riddell, 2016). Many times, the child is afraid of getting killed just for living. Uncertainty about the future can also affect these children negatively. Children may become self-destructive, be aggressive towards others or attempt to escape from police officers (Kullander et al., 2016;Riddell, 2016). This behaviour requires police officers to be vigilant about the safety and security of the child as well as third parties and their own safety and security. At the same time, the officers must manage sad and lonely children who must leave their friends and an environment that has been a safe haven for them. In this complex situation, "the best interests of the child shall be a primary consideration" (UN General Assembly, 1989, p. 3). These demands may be contradictory, and police officers must cope with stressful situations.

Theoretical framework
In the context of medical science, the concept of stress was coined by Selye in 1936, who defined it as "the non-specific response of the body to any demand for change" (Selye, 1978). Selye's definition was complemented by Lazarus's (1966Lazarus's ( , 1993Lazarus's ( , 2000 studies on stress. Lazarus's pioneering and original work on stress recognized that subjective stress, not objective stress, causes problems. A life experience is stressful for individuals only if they perceive it as stressful. Thus, we can define stress as negative feelings and beliefs that arise when people feel unable to cope with demands from the environment (Aronson, Wilson, & Akert, 2013;Lazarus & Folkman, 1984).
The interaction between a person and his or her environment is mediated by "cognitive appraisal" and is called "coping" (Lazarus & Folkman, 1984). Lazarus and Folkman (1984) stated that cognitive appraisal occurs in the form of two main mechanisms that contribute to responses to stress: primary and secondary appraisal. Lazarus and Folkman (1984) defined coping as the sum of cognitive and behavioural efforts that are constantly changing and that are intended to handle particular demands, whether internal or external, that are viewed as taxing or demanding. Simply put, coping is an activity we perform to identify and apply solutions to stressful situations or problems that emerge because of stressors (Lazarus & Folkman, 1984). Coping has two major functions: regulating stressful emotions (emotion-focused coping) and changing the troubled environment that causes distress (problem-focused coping). These coping strategies can be used together or separately (Lazarus & Folkman, 1984).
According to the World Health Organization (WHO), work-related stress often has a negative effect on individuals' mental health, which might cause "early retirement from work, high absence rates, and low organizational productivity" (WHO, 2014b). The WHO defines mental health as: "a state of well-being in which every individual realizes his or her own potential, can cope with the normal stresses of life, can work productively and fruitfully, and is able to make a contribution to her or his community" (WHO, 2014a). Social, psychological and biological factors affect an individual's mental health.
Individuals' coping strategies have different effects on their general mental health. For example, positive reappraisal was found to be a protective factor for mental health disorder among Swedish patrolling police officers in an exploratory cross-sectional study (Padyab, Backteman-Erlanson, & Brulin, 2016). Around the late 1980s, research examined appraisal and coping among Swedish police officers in acute and stressful situations. The findings showed that police officers appraised situations as more solvable, felt less threatened and utilized more problem-solving coping than other people did (Larsson, Kempe, & Starrin, 1988). Another study on police officers in Australia found that emotion-focused coping contributed to negative work experiences, whereas problem-focused coping resulted in positive work experiences (Hart et al., 1995). This finding is supported by a study on Jamaican police officers that suggested that emotion-focused coping is related to increased levels of depression and anxiety (Nelson & Smith, 2016). In contrast, a study on US police officers found that emotion-focused coping had a buffering effect on distress in relation to stressful life events, whereas problem-solving coping had no buffering effect (Patterson, 2003). In a recent study on first responders in Sweden, the findings showed that approach coping behaviours, such as confronting a situation, were related to better well-being, whereas avoidant coping behaviour, such as avoiding talking about a stressful experience, were related to a decrease in well-being (Arble & Arnetz, 2016). This finding is consistent with another US study suggesting that dissociation is related to stressors and mental health disorders (Aaron, 2000).
This study aims to describe police officers' coping in the deportation of UARC and to investigate the associations between coping and general mental health in relation to the deportation of UARC among police officers by considering sociodemographic variables. Validity and reliability analyses were conducted to use the WOCQ in the context of UARC.

Material and methods
Until 2014, the Swedish Police Service was composed of 21 police county authorities. After 1 January 2015, it was reorganized into one National Police Authority (Swedish National Police Board, 2016). All 21 police authorities in Sweden were contacted in spring of 2014 and requested to participate in a survey about work stress in relation to the deportations of UARC. All but three of the police authorities participated in the study. After approval by the Regional Ethics Committee at Umeå University (2014/69-31Ö9), a total of 2113 survey questionnaires and prepaid return envelopes were mailed in a sealed envelope to the contact person at each authority. The contact person distributed them to the police officers, who responded to the survey and returned it in the prepaid envelope. A total of 714 police officers responded and returned the questionnaires. This sample size provided approximately 9 cases per item, which is considered sufficient to conduct factor analysis (Pallant, 2005). This convenience sample was used due to secrecy rules in the Swedish police organization. The survey included an introductory letter stating the purpose of the study, a consent form, and a selfadministered questionnaire used to collect information on sociodemographic data, coping strategies and the general mental health of the participants.

Instruments
Coping and mental health were assessed using two different questionnaires.

Ways of coping
The Ways of Coping Questionnaire (WOCQ), which was translated into Swedish, assesses coping strategies (Lazarus & Folkman, 1984). The WOCQ has been standardized and validated in Sweden in relation to different contexts, mainly health situations (Ahlström & Wenneberg, 2002;Lundqvist & Ahlström, 2006;Padyab et al., 2016). The instructions state that the respondents will be asked to answer each item in terms of how they coped with a specific scenario, such as, "Two days ago, you enforced the deportation of a 16-year-old asylum-seeking youth, who told you about the escape to Sweden and the fear of being forced to go back and being killed". The instrument contains 66 items with a recommended four-point response scale ranging from does not apply or not used (0) to used a great deal (3) (Folkman & Lazarus, 1988). A sample item was, "I talked to someone about how I was feeling" or "I made a plan of action and followed it". The scoring of the coping sub-scales was conducted according to the WOCQ manual (Folkman & Lazarus, 1988) as the mean values of items in each sub-scale.

General mental health
General mental health was measured by the General Health Questionnaire 12 (GHQ-12), which has been well validated and used in a large number of studies, (Arnetz, Arble, Backman, Lynch, & Lublin, 2013;Black, McCabe, & McConnell, 2013;Johnson et al., 2005;Marchand & Durand, 2011;Mofidi, Ghazinour, Araste, Jacobsson, & Richter, 2008). A study on a Swedish sample showed an internal consistency (Cronbach's α) between 0.80 and 0.84 for three different groups (Sconfienza, 1998). This 12-item screening instrument was developed to detect psychiatric disorders in community settings and non-psychiatric clinical settings (Arnetz et al., 2013;Banks et al., 1980;Goldberg & Williams, 1988). Each GHQ item has four answer options. The GHQ sum scores were based on the Likert score (item score 0-0-1-1; score range 0-12). Higher scores indicate poorer mental health. We used the original scoring by Goldberg with response categories scored "not at all" and "no more than usual" as "0" and "somewhat more than usual" and "much more than usual" as "1", giving a possible range from 0 to 12, with higher scores indicating poorer mental health. In accordance with Goldberg, Oldehinkel, and Ormel's (1998) recommendation that "if the mean is below 1.85 then the threshold of 1/2, from 1.85 to 2.7 a threshold of 2/3, and above 2.7 a threshold of 3/4 seems to work best for the GHQ-12", we decided to use the appropriate scores as the cut-off for the differentiation between individuals with and without psychological disturbance. The 1/2 recommendation implies, for example, that those answering positively to more than 2 questions would be considered a "case". In this study, the GHQ mean was 1.5, which is lower than 1.85, then the given cut-off of 2 is considered as psychological disturbance.

Other study variables
Experience working with UARC was identified using a single-item question in the sociodemographic questionnaire (yes/no). The sociodemographic questionnaire included questions about the participants' demographic and work characteristics (sex, age, marital status, education, working experience, type of duty and shift work). Sex was defined as male or female. Age was a continuous variable in years. Marital status had two categories, married (married and cohabiting) and single (single, widowed and divorced). Education was defined as a university degree or no university degree. Working experience was the total work experience as a police officer in years. Type of duty was defined as patrolling versus others. Shift work was defined as daytime work and shift work.

Statistical methods
The GHQ-12 was the outcome (dependent variable), and experience of deportations of UARC, the WOCQ and sociodemographic variables were the independent variables. The GHQ-12 was used as categorical (psychological disturbance "yes" or "no") variables. The construct validity of the WOCQ was investigated using confirmatory factor analysis (CFA) and exploratory factor analysis (EFA). In the first step, CFA was used to test the original eight-factor structure (Folkman, Lazarus, Dunkel-Schetter, DeLongis, & Gruen, 1986). In the second step, EFA was used to extract more appropriate factors based on the given scenario among police officers. To obtain the optimal number of factors, we used parallel analysis (Hayton, Allen, & Scarpello, 2004). In the third step, minor modifications were applied and CFA on the newly derived scales from EFA was utilized. The correspondence between the WOCQ items and the coping dimensions specified by Folkman and Lazarus's theory was evaluated. In this procedure, the items in each factor were evaluated in relation to the coping dimensions and compared to Folkman and Lazarus's (1988) labelling. Items that did not fit into the coping dimension and had a low loading factor were omitted from that factor scale. The decisions on some items were based on semantic and conceptual reasons rather than empirical reasons. CFAs followed Jöreskog's (2004) guidelines for the analysis of ordinal data. Diagonal weighted least-squares estimation was applied to polychoric correlations that were based on the asymptotic covariance matrix. The models were evaluated by means of the Bentler-Satorra chi-square score, root mean square error of approximation (RMSEA) (Steiger, 1990), and 90% confidence interval of RMSEA. Browne and Cudeck (1993) provided a guideline stating that values of RMSEA less than 0.05 indicate a close fit, values in the range of 0.05 to 0.08 indicate a fair fit, and values above 0.10 indicate a poor fit. The goodness-of-fit index (GFI) and adjusted goodness-of-fit index (AGFI) were used. Values exceeding 0.90 indicate a good fit of the model to the data. The comparative fit index (CFI) (Bentler, 1990) and incremental fit index (IFI) (Bollen, 1989) were also reported, with values equal to or greater than 0.90 denoting an acceptable fit to the model (Bentler & Bonett, 1980;Kline, 2005).
EFA was conducted using principal component analysis (PCA) with polychoric correlation and oblique rotation. Polychoric correlation was applied because the variables consist of ordinal data, and oblique rotation was used because the factors were expected to be dependent on each other (Folkman & Lazarus, 1988). The internal consistency of the scale was measured by calculating Cronbach's α, which ranges from 0 to 1, with values greater than 0.60-0.70 considered to have acceptable reliability (Tabachnick & Fidell, 2006).
Coping sub-scales, age and working experience were compared between police officers with and without psychological disturbance as well as between police officers with and without experience with forced repatriations of UARC using a t-test or Mann-Whitney test depending on the normal distribution. The Mann-Whitney test was used to compare GHQ between police officers with and without experience with UARC's forced repatriation. Pearson's Chi-square test for independence was used to calculate the association between categorical variables (gender, marital status, education, duty, work schedule) and experience with UARC's forced repatriation and psychological disturbance.
Univariate and multivariable logistic regression described the impact of WOCQ on the likelihood of being a case according to GHQ cut-off scores (that is, the presence of psychological disturbance, "yes" or "no"). Hierarchical logistic regression analysis was used to assess the predictive value of coping and other sociodemographic variables for psychological disturbance. The significance level was set at p < 0.05. We conducted EFA of the factor structure of the WOCQ and statistical analysis using SPSS 23, and we performed CFAs by means of LISREL 8.8.
An independent-samples t-test was conducted to compare the mean age, work experience and all the coping strategy scores for police officers with and without experience with unaccompanied, asylum-seeking refugee children's forced repatriation (Table 2). There was a significant difference in mean age for experience (44 ± 11.61) and lack of experience (41 ± 11.10; p = 0.011) and work experience (18 ± 13.94 vs. 15 ± 12.95, respectively, p = 0.014). There were significant differences in scores for escape-avoidance, planful problem-solving and self-control for experience and lack of experience. Escape-avoidance is used more often among those with no experience of UARC (0.58 ± 0.41 vs. 0.41, ± 0.42, respectively, p < 0.001). Planful problem-solving is also used more often among those Notes: F&L = factor labels from the solution of Folkman and Lazarus (1988); factor = factor label current study; PPS = planful problem solving; EA = escape-avoidance; D = distancing; SSS = seeking social support; CC = confrontive coping; SC = self-control; AR = accepting responsibility; PR = positive reappraisal.
with no experience of UARC (1.22 ± 0.64 vs. 1.02, ± 0.68, respectively, p < 0.001). Police officers without experience of UARC used more self-control compared to those with experience (1.64 ± 0.52 vs. 1.44 ± 0.58, respectively, p < 0.001), and distancing was used more often among those with no experience of UARC (1.05 ± 0.46 vs. 1.13 ± 0.46, respectively p = 0.050). A Pearson Chi-square test showed no significant associations among gender, marital status, education, duty, work schedule, psychological disturbance and experience with UARC's forced repatriation (Table 2). Table 3 shows a significant association between psychological disturbance and marital status (p = 0.001) as well as work schedule (p = 0.007). There were significant differences in scores for  escape-avoidance and self-control among individuals with psychological disturbance compared with those without psychological disturbance. Escape-avoidance is used more often among those with psychological disturbance compared with those without psychological disturbance (0.62 ± 0.43 vs. 0.51, ± 0.41, respectively, p = 0.004). Similar findings emerged for self-control, with a relatively higher score among those with psychological disturbance compared to those without (1.71 ± 0.52 vs. 1.57 ± 0.54, respectively, p = 0.003).

Logistic regression
Univariate logistic regression was performed to assess the bivariate associations between independent variables and the likelihood that respondents would report that they had psychological disturbance. The full model contained 13 independent variables: experience of UARC and five different coping strategies, age, gender, marital status, work experience, education level, duty and work schedule. In step one (model 1), univariate logistic regressions were conducted including all independent variables one at a time. In step two (model 2), a multivariable logistic regression containing experience with UARC and sociodemographic variables was conducted. In step three (model 3), a multivariable logistic regression containing all independent variables was conducted.
As shown in Table 4, four of the independent variables made a unique statistically significant contribution to model 1, namely, the coping strategies of escape-avoidance and self-control as well as marital status and work schedule. The strongest predictor of reporting psychological disturbance was being unmarried; these individuals were twice as likely to report psychological disturbance compared to married individuals (OR = 2.01, 95% CI: 1.34-3.04). The odds ratio for the coping strategy of escape-avoidance was 1.76 (95% CI: 1.19-2.61). This finding indicated that for each unit increase in reporting the coping strategy of escape-avoidance, respondents were almost 1.8 times more likely to report psychological disturbance. The odds ratio for self-control was 1.64 (95% CI: 1.17-2.29), indicating that for each unit increase in reporting the coping strategy of self-control, respondents were more than 1.6 times more likely to report psychological disturbance. The odds ratio for working shift was 1.58 (95% CI: 1.13-2.22). This finding indicated that respondents who performed shift work were approximately 1.6 times more likely to report psychological disturbance than those who did not perform shift work.
The odds ratio for marital status decreased from 2.01 (95% CI: 1.34-3.04) in model 1 to 1.89 (95% CI: 1.23-2.92) in model 2, indicating that respondents who were unmarried were approximately 2 times more likely to report psychological disturbance than those who were married when controlling for other variables. Contrarily, the odds ratio for work schedule increased from 1.58 (95% CI: 1.13-2.22) in model 1 to 1.96 (95% CI: 1.11-3.46) in model 2. This finding indicated that respondents who performed shift work were almost 2 times more likely to report psychological disturbance than those who did not perform shift work when controlling for sociodemographic variables.
Model 3 showed an increase in odds for escape-avoidance from 1.76 in model 1 to 2.32 (OR = 2.32, 95% CI: 1.26-4.28). This finding indicated that for each unit increase in reporting the coping strategy of escape-avoidance, respondents were more than twice as likely to report psychological disturbance when controlling for other factors in the model. In addition, the odds ratio for positive reappraisal showed a statistically significant contribution to the model. The odds ratio of 0.61 (95% CI: 0.394-0.950) for the coping strategy of positive reappraisal was less than 1, indicating that for each unit increase in reporting the coping strategy of positive reappraisal, respondents were 40% less likely to report psychological disturbance when controlling of other factors in the model. The odds ratio for self-control showed a slight increase compared to model 1 (OR = 1.64, 95% CI: 1.17-2.29 vs. 1.74, 1.09-2.80, respectively). The odds ratio for marital status slightly decreased (OR = 1.89, 95% CI: 1.23-2.92 vs. 1.78, 1.12-2.83, respectively), whereas work schedule remained unchanged compared to model 2.
Controlling for other factors in the model, the strongest predictor of reporting psychological disturbance was the coping strategy of escape-avoidance, followed by shift work. Self-control showed the same level of prediction for psychological disturbance. Marital status was the only significant variable that showed lower prediction for psychological disturbance in model 3 compared to model 1. The coping strategy of positive reappraisal was found to have a protective effect on psychological disturbance when controlling for other variables.

Discussion
The present study was designed to describe police officers' coping in the deportation work of unaccompanied, asylum-seeking refugee children and to investigate the associations between ways of coping and general mental health in relation to such work considering sociodemographic variables. In addition, the WOC questionnaire was validated in the context of UARC. To the best of our knowledge, this is the first study investigating police officers' coping in the deportation work of UARC. Thus, we have not found previous studies in this context to compare our main outcomes.
In the current report, some of the factor scales were similar to those identified by Folkman and Lazarus (1988). However, "Seeking social support", "Confronting coping" and "Accepting responsibility" could not be identified among the components in this study. A possible explanation for these differences may be related to our choice of using a specific but complex scenario in the questionnaire. In the current study, we used a scenario that was related to working with the deportation of UARC. The original WCQ asked the respondents to think of the most stressful situation they had experienced in the past week (Folkman & Lazarus, 1988). Another explanation may be that coping behaviours are situational and may vary by country, occupation, social group, gender, age, and situation, and they are largely influenced by individuals' previous experiences (Lazarus & Folkman, 1984). These preferences affect police officers' appraisal of the demands and resources to cope with this complex and stressful situation.
Coping behaviours are situational (Lazarus & Folkman, 1984). In the current study, both police officers with and without experience with deportations of UARC seemed to utilize a combination of emotional and problem-solving coping strategies. A possible explanation for this finding is that in deportation work, an officer should consider (1) how to solve their work task, that is, to deport the child and (2) how to care for a child who is under high stress. These contradictory demands may make police officers appraise the same situation from different angles. Accordingly, officers must alternate between emotional and problem-solving coping. This is a psychological perspective that can be explained from an organizational perspective using Lipsky's (2010) street-level bureaucracy.
According to Lipsky (2010), police officers constantly weigh different interests against one another. In the current situation, the police officers face two overarching political demands. First, the officers must consider efficiency; they must increase the number of forced deportations (Swedish Government, 2014). Second, they must consider dignity by caring for the unaccompanied asylumseeking refugee children's human rights ( /115/EC, D, 2008Council of Europe (COM), 2005). In this situation, police officers utilize their discretionary power to manage an assignment that involves these seemingly contradictory demands. The officers utilize their discretion to interpret increased efficiency and dignity. This interpretation makes it cognitively possible to combine increased efficiency and dignity (Hansson, Ghazinour, & Wimelius, 2015). More specifically, and from a psychological perspective, the current scenario consists of several steps that may be connected to either "increased efficiency" or "dignity" and that activate different coping strategies. The first step is the assignment to "enforce" the child. The response to this may be connected to "increased efficiency" and coping strategies such as for example "planful problem solving" and/or "positive reappraisal". These coping strategies are related to the assignment. The second step is the meeting with an overstrained child. The response to this may be connected to "dignity" and coping strategies such as "distancing" and/or "self-control". These coping strategies are related to the child as an individual. The third step is the content of the story the child is telling the police officer. The response to this may be connected to the complexity of combine "increased efficiency" and "dignity" and to coping strategies such example "escape-avoidance" and/or "positive reappraisal". These coping strategies are related to the child's story and what the child feels. This may explain why police officers seem to combine five different coping strategies. However, it is not possible to determine exactly which coping strategy is adaptive due to the changing and complex scenario.
The current deportation scenario is very complex. This complexity makes it likely that police officers think and act in different ways and that they perceive different demands and resources to manage a situation. The police officer must deal with a child who may be very sad and upset. This makes it necessary for the officer to both protect the child from harming himself or herself and to prevent the child from escaping. Simultaneously, the officer has the responsibility of helping the child say goodbye to his or her friends and packing their belongings. The police officer must provide emotional support in this situation that is stressful for the child. In this situation, police officers must manage their own emotions without being governed by them. Every police officer perceives a situation differently and has different resources to manage the situation. This makes it necessary to combine different coping strategies. This can be explained by Folkman et al.'s (1986, p. 993) description of coping: "it is process oriented, meaning that it focuses on what the person actually thinks and does in a specific stressful encounter, and how this changes as the encounter unfolds" and is "contextual, that is, influenced by the person's appraisal of the actual demands in the encounter and resources for managing them" (Folkman et al., 1986, p. 993).
Although officers utilize a combination of different coping strategies, we have found some patterns in the association between the experience of deportations of UARC and coping strategies. There seems to be a moderating effect on general mental health depending on which coping strategy police officers utilize in their work with forced deportations of UARC. Police officers with experience of UARC use less escape-avoidance, self-control and planful problem solving. According to the results from the final logistic regression model, police officers with psychological disturbance use more escape-savoidance, more self-control and less positive reappraisal. One interpretation of the inverse association between positive reappraisal and psychological disturbance and the association between escape-avoidance and psychological disturbance is that it is better for the officers' mental health to face the situation and talk about it but not to attempt to solve the problem or change the situation. The use of the coping strategies of escape-avoidance, self-control or positive reappraisal has a moderating effect on general mental health.
The findings in the multivariable logistic regression showed that self-control increased the risk of psychological disturbance. "Self-control" is characterized by efforts to regulate one's feelings and actions. This behaviour may have a negative effect on general mental health, especially if the work task in the situation conflicts with one's own values. These results seem to be consistent with other research that found that coping skills involving blame/denial and negative distraction were associated with high psychological stress (Kaur, Chodagiri, & Reddi, 2013). Furthermore, in the multivariable logistic regression, consistent with a previous study (Padyab et al., 2016), positive reappraisal decreased the risk of psychological disturbance, and escape-avoidance increased the risk of psychological disturbance. The latter corroborates the ideas of Arble and Arnetz (2016), Penley, Tomaka, and Wiebe (2002), and Violanti (1992), who suggested that avoidance is negatively associated with overall health outcomes, and Aaron (2000), who suggested that dissociation is related to stressors and mental health disorders. The results showed that the risk of psychological disturbance differs depending on which coping strategy police officers utilize.
Several reports have shown that emotion-focused coping is related to negative effects on health or work experience (Hart et al., 1995;Nelson & Smith, 2016), whereas Patterson (2003) found that emotion-focused coping had a buffering effect on distress. Acquadro Maran, Varetto, Zedda, and Ieraci (2015) found that humour (emotional coping) and planning (planful coping) were negatively correlated with perceived distress. In the current study, escape-avoidance, self-control and positive reappraisal can be defined as emotional coping. Thus, we suggest that emotional coping has a dubious effect on general mental health in this situation. These relationships may partly be explained by the way police officers interpret their demands and resources in the deportation work of UARC. More specifically, the officers utilize discretion to appraise a situation based on their individual perceptions of the situation . In the primary appraisal, the officer "evaluates if he or she has anything at stake in this encounter", and in the secondary appraisal, the officer "evaluates what if anything can be done to overcome or prevent harm or to improve the prospects for benefit" (Folkman et al., 1986, p. 993). Primary and secondary appraisals determine which coping strategy should be used in a specific situation.
Another finding was that "shift work" became a stronger predictor of psychological disturbance in the multivariable logistic regression compared to the univariate logistic regression. The marital status "single" was the strongest predictor of psychological status in the univariate logistic regression, but when we controlled for other variables in the model, being single became a weaker predictor and "shift work" became a stronger predictor of psychological disturbance. The most probable reason why the risk of psychological disturbance increased among shift workers and decreased among single officers may be because it is difficult to combine shift work with living with a partner.

Conclusion
This study has shown that in the deportation work of UARC, police officers utilize different coping strategies in the same complex situation. The research has also shown that coping strategies have a moderating effect on general mental health in the work with forced deportations of UARC. More specifically, the coping strategies of "escape-avoidance" and "self-control" have a negative effect on general mental health, whereas the coping strategy of "positive reappraisal" has a positive effect on general mental health. In addition, shift work and being single have a negative effect on general mental health.
The fact that coping is process oriented and contextual and that police officers utilize different coping strategies in the same situation makes it problematic to determine which coping strategy generally has a positive impact on officers' general mental health. However, Penley et al. (2002) suggested that problem-focused coping is positively associated with overall health outcomes, whereas avoidance is negatively associated with overall health outcomes. Furthermore, Nelson and Smith (2016) found that emotion-focused coping was positively associated with depression and anxiety in Jamaican police officers. In the current study, the scenario is complex, and officers can perceive the situation and appraise the demands in many different ways. Depending how the police officers perceive and appraise the situation, the same coping strategy may be either adaptive or maladaptive. Nevertheless, patterns of both adaptive and maladaptive coping are evident in this study. It is important to note that officers utilize different coping strategies simultaneously and that coping is situational, and it might be more adaptive to utilize positive reappraisal than escapeavoidance or self-control. This knowledge may contribute to the understanding of how different coping strategies affect general mental health.
The most important limitations of this study lie in the fact that we used self-reported, cross-sectional data. Another limitation is the convenience sample that we were required to use due to secrecy rules and the difficulty of collecting data from the police authority. Although a probability sample is preferable (Dawson & Trapp, 2004), the importance of this study makes this limitation reasonable. Another source of weakness in this study might be the response rate. However, it has been shown that even with only moderate response rates, studies on health issues, prevalence figures, and associations between variables are relatively unbiased (Søgaard, Selmer, Bjertness, & Thelle, 2004).
Despite these limitations, the findings of this research suggest the importance of being aware that different coping strategies moderate the effect of a situation on general mental health. We recommend education for police officers to make them aware of their coping behaviour in these situations. Education on coping theory for police students and officers might function to prevent harm to UARC. In addition, we recommend external counselling supervision. The purpose of this counselling supervision is to change perspectives by talking and listening to other individuals in the work team. It may also function as organizational social support in which the police officers can reflect in their teams. Future research might explore the moderating or mediating effect of different factors, such as, locus of control, burnout and personality on well-being of the Swedish Police Officers. Further qualitative measurement of coping need to be carried out in order to have a deeper understanding of coping process.