*2.3. Statistical Analyses*

Statistical analyses were performed using STATA MP in version 16. Descriptive statistics were used to examine participants' time pressure, adverse mental health, family separation, sociodemographic, and postmigration characteristics. Differences in time pressure and adverse mental health by family separation, sociodemographic, and postmigration factors were analyzed using chi-squared, Fisher's exact, and *t*-test. Multiple logistic regression [41] was applied to examine whether family separation, sociodemographic, and postmigration factors were potential risk factors for (i) being time-stressed and (ii) having a higher risk for adverse mental health. The null hypothesis was that family separation was not a potential risk factor for (i) being time-stressed and (ii) having a higher risk for adverse mental health considering the main relevant confounders of sociodemographic and postmigration factors. Cases presenting a missing value for at least one of the modeling variables were excluded from analyses (listwise deletion). Nevertheless, we could not find any specific characteristics for incomplete answers. Independent variables considered in the multivariable analyses were age, sex (as a dummy variable (DV)), country of origin (DV), residence status (DV), length of stay since arrival (DV), separation from spouse or partner (DV), and separation from child or children (DV). Hosmer–Lemeshow goodness of fit test was used to evaluate the logistic regression models. We computed odds ratios (ORs) and 95% confidence intervals (CIs), and the significance was set at *p* < 0.05.

#### **3. Results**

Table 1 presents the characteristics of the study population. The majority were male, most immigrated from war-affected Middle Eastern countries, and the mean (SD) length of time since arrival in Germany was 29.9 (11.6) months. More than 30% of participants with a spouse or partner, and about 18% with a child or children reported separation. About 38.9% of participants subjectively experienced time pressure always (19.4%) or often (19.9%), with a mean (SD) MCS score of 39.6 (13.9). Participants who subjectively experienced time pressure sometimes (28.2%), seldom (8.7%), and never (23.8%) had a mean (SD) MCS score of 47.5 (14.3). Overall mean (SD) MCS score was 44.5 (14.6).


**Table 1.** Characteristics of the study population (*n* = 208).

*n*, quantity; %, proportion; SD, standard deviation; <sup>a</sup> Algeria (14.3%), Eritrea (14.3%), Nigeria (28.6%), Somalia (7.1%), Ghana (14.3%), Morocco (7.1%), Egypt (14.3%); <sup>b</sup> Azerbaijan (5.3%), Bangladesh (21.0%), India (5.3%), Lebanon (15.8%), Palestine (10.5%), Russia (10.5%), Tajikistan (5.3%), Turkey (10.5%), stateless (15.8%).

Table 2 provides detailed information on sociodemographic and postmigration factors stratified by time pressure and adverse mental health groups. More than half of the female participants were classified in the higher risk for adverse mental health group, while most male participants reported less time pressure and better mental health. Frequency in the time-stressed group increased with length of stay since arrival in Germany from 29.2% (<12 months) to 44.3% (>36 months). Nevertheless, the majority of participants who reported family separation were in the not time-stressed group but in the higher risk for adverse mental health group.


**Table 2.** Time pressure and adverse mental health groups across independent variables (*n* = 208).

*n*, quantity; %, proportion; SD, standard deviation; *p*, *p*-value; <sup>a</sup> Algeria, Eritrea, Nigeria, Somalia, Ghana, Morocco, Egypt; <sup>b</sup> Azerbaijan, Bangladesh, India, Lebanon, Palestine, Russia, Tajikistan, Turkey, stateless.

> Concerning family separation, it was found that the majority of respondents with a partner and at least one child were not separated (76.5%), while 11.8% of respondents who were separated from their spouse or partner were also separated from their child or children.

Table 3 shows the regression model between being time-stressed as the dependent variable and independent variables (*n* = 129), which had an acceptable fit (Hosmer and Lemeshow statistic: χ <sup>2</sup> = 1.87, df = 8, *p* = 0.985). Participants with insecure residence status were more likely to be time-stressed as compared to those with a secure status.


**Table 3.** Results of logistic regression analysis on predictors for being time-stressed (*n* = 129).

OR, odds ratio; CI, confidence interval; <sup>a</sup> Algeria, Eritrea, Nigeria, Somalia, Ghana, Morocco, Egypt; <sup>b</sup> Azerbaijan, Bangladesh, India, Lebanon, Palestine, Russia, Tajikistan, Turkey, stateless.

Table 4 displays the final model with factors associated with higher risk for adverse mental health (*n* = 124), also with an acceptable fit (Hosmer and Lemeshow statistic: χ <sup>2</sup> = 8.63, df = 8, *p* = 0.375). Separation from at least one child was strongly associated with a higher risk for adverse mental health. Furthermore, female sex and insecure residence status were also found to be positively associated with a higher risk for adverse mental health, while a length of stay of more than three years since arrival was negatively associated.


**Table 4.** Results of logistic regression analysis on predictors for higher risk for adverse mental health, *n* = 124.

OR, odds ratio; CI, confidence interval; <sup>a</sup> Algeria, Eritrea, Nigeria, Somalia, Ghana, Morocco, Egypt; <sup>b</sup> Azerbaijan, Bangladesh, India, Lebanon, Palestine, Russia, Tajikistan, Turkey, stateless.

#### **4. Discussion**

This study aimed to investigate the association of family separation with time pressure and adverse mental health considering the main relevant confounders of sociodemographic and postmigration factors among refugees in North Rhine-Westphalia, Germany. This study showed that separation from at least one child was positively associated with a higher risk for adverse mental health; however, separation from spouse or partner and/or child or children was not associated with being time-stressed.

In our sample, about 38.9% of refugees were classified as being time-stressed; however, one study examining the relationship between socioeconomic characteristics and time pressure in the German general population based on SOEP data in its 2002 wave version using the same instrument [37] showed a lower overall prevalence of being time-stressed (35.3%). Therefore, time pressure is prevalent in our studied refugee sample. As this study may be the first to explore the association between family separation and subjectively perceived time pressure in a resettled refugee population, we did not find support for our hypothesis that refugees separated from their family members experience time pressure

always or often. Nevertheless, it is interesting to note that a study primarily focusing on the mental health consequences of family separation for refugees found that separation from family members was a major stressor because family reunification was one of the refugees' primary needs [22]. This stressor could be exacerbated in times of the COVID-19 pandemic, as asylum and resettlement processes are disrupted by lockdowns [42].

Concerning our second hypothesis that refugees who are separated from their family members experience adverse mental health, we did find support that separation from at least one child might be a source of health inequalities among resettled refugees. Our findings show that separation from a child may be a risk factor for mental illness, which is consistent with previous research highlighting the negative impact of child separation on migration-related stress [43], distress [22], and mental health and well-being [44]. In the contrast, prior research shows that children who were separated from their parents report greater symptoms of anxiety [45], depression [46], and psychotic disorders [47]. The findings of our study extend previous work by demonstrating the adverse impact of child separation in resettled refugees from diverse language and ethnic groups. We also observed that Afghans and Iranians in particular were at higher risk for mental health problems. However, origin itself was not a significant influencing factor in the multivariable analyses.

Concerning sociodemographic and postmigration factors, multiple logistic regression revealed that insecure residence status was positively associated with both being timestressed and having a higher risk for adverse mental health. This is consistent with the literature, as previous research showed that insecure residence status can cause postmigration stress [48] and pose a psychological risk [32,49,50]. A study comparing hair cortisol concentration (HCC) of recently fled asylum seekers with and without PTSD found no difference in HCC; however, compared with permanently settled immigrants, recently fled asylum seekers showed higher HCC [51]. This finding has important policy implications, as refugees with insecure residence status suffer from stress, which in turn can negatively affect mental and physical health [52,53]. From a public health perspective, it is therefore critical that legal restrictions on refugees' access to health care be lifted, regardless of their residence status. The relation of adverse mental health with female sex is established by prior studies [31,54]. As female refugees represent a minority, the lack of gender-specific reception and housing conditions must be addressed. Generally, life in reception centers is more difficult for single women since they lack male protection [55]. In addition, length of stay longer than 36 months since arrival was found to be negatively associated with a higher risk for adverse mental health. A study assessing the prevalence and risk factors for mental distress among refugees in Germany showed that a shorter duration of residence permission was shown to be associated with more severe symptoms of PTSD [56]. Another study focusing on the association between length of stay in asylum centers and mental disorders found that a longer length of stay was associated with an increase in cases of mental disorders [57]. Further data collection and analysis are needed to draw a conclusion. Together these findings suggest that in addition to family separation, sociodemographic and postmigration factors pose major risks for health and well-being among refugees.

There are several limitations to our research. First, we utilized data from the FHS, and selection bias may be an issue because participants in the FHS were self-selected. This means that individuals who were not interested in health issues have decided not to participate in the FHS. In addition, the research questions for this study were developed after the data were collected, so we lack some important information that could be helpful to further understand the relationship between family separation and time pressure. Asking about premigration stressors and reasons for time pressure since their arrival might be beneficial to further understand time pressure as a potential determinant of health among refugees. Another limitation arises in connection with the cross-sectional design. Thus, we were unable to examine the temporal relationship of family separation with time pressure and adverse mental health; longitudinal design studies are recommended. Moreover, there is a need to disentangle the pathways between family separation, time pressure, and mental health through mediation analyses. In addition to that, both outcome variables were

dichotomized for regression analyses, resulting in a loss of information [41], particularly for the MCS score; however, the sample mean MCS score was within the range of an optimal cut-off point to screen for adverse mental health such as depressive and anxiety disorders [40] and more than five points below the respective general population mean; thus, it was within the range of five to ten points as a minimally important difference [58]. Since our data only included a convenience sample of 208 refugees from East-Westphalia-Lippe, caution should be taken on the generalizability of the results. It should also be noted that the refugees who were able to come to Germany from Africa and the Middle East might be in relatively better health than refugees who settled in neighboring countries of their home country. Finally, our sample was heterogeneous in terms of origin, causes of flight, and ethnocultural family orientation. Despite its limitations, this study provides new data on subjectively perceived time pressure as a postmigration stressor and on the mental health of refugees from the Middle East and Africa resettled in North Rhine-Westphalia, Germany.
