The demographic characteristics of the study population are detailed in Table 2 [near here]. The Chi-squared test, based on 2000 replicates, underscores substantial disparities in the demographic profiles of PNM, based on their migration status as IDP or REF, particularly with regard to their previous places of residence (χ2 = 13.73, p = 0.005). Fisher’s Exact Test further confirms these disparities, indicating significant differences in the distribution of previous residence types between groups (p = 0.003). In detail, a larger percentage of REF (25.0%) compared to IDP (14.3%) originated from the capital city. While both groups showed a significant representation from larger cities – 50.0% of IDP and 33.3% of REF – a higher proportion of IDP are associated with this type of area. Conversely, REF had a greater representation from small towns (32.1% compared to 12.5% for IDP) and fewer REF originated from villages (9.5%) compared to IDP (23.2%), indicating a more rural background among the IDP group. Additionally, there are a difference in the origins of new mothers compared to pregnant women based on Chi-squares test (χ2 = 6.19, p = 0.002), which aligns with Fisher’s Test results (p = 0.0004). New mothers more commonly came from big cities (40%) and less frequently from small towns (24.3%), whereas pregnant women were more likely to come from small towns (49%) and less from big cities (29.4%).
Table 2
Comparison of sociodemographic characteristics of PNM groups
Variables | Pregnant | New Mothers | χ2 (p) |
IDP | REF | IDP | REF |
Respondents, n | 22 | 29 | 56 | 84 | |
Level of education, n(%) | χ2 = 8.55, p = 0.104 | χ2 = 3.16, p = 0.743 | 4.45 (0.487) |
Incomplete secondary | 1 (4.5) | 2 (6.9) | 1 (1.8) | 1 (1.2) |
Full secondary education | 4 (18.2) | 2 (6.9) | 10 (17.9) | 8 (9.5) |
Incomplete higher education | 3 (13.6) | | 5 (8.9) | 6 (7.1) |
Bachelor | 4 (18.2) | 11 (37.9) | 12 (21.4) | 22 (26.2) |
Masters | 9 (40.9) | 14 (48.3) | 28 (50.0) | 46 (54.8) |
Scientific degree | 1 (4.5) | | | 1 (1.2) |
Marital status, n(%) | χ2 = 1.45, p = 0.933 | χ2 = 1.50, p = 0.705 | 4.60 (0.326) |
Single | 1 (4.5) | 1 (3.4) | 1 (1.8) | 5 (6.0) |
Partnered/Married | 14 (63.6) | 20 (69.0) | 45 (80.4) | 64 (76.2) |
Divorced/Separated | 6 (27.3) | 5 (17.2) | 9 (16.1) | 13 (15.5) |
Widowed | 1 (4.5) | 1 (3.4) | 1 (1.8) | 2 (2.4) |
Other (Unspecified) | | 1 (3.4) | | |
Place of residence, n(%) | χ2 = 0.09, p = 0.999 | χ2 = 13.73, p = 0.005 | 6.19 (0.002) |
The capital | 5 (22.7) | 6 (20.7) | 8 (14.3) | 21 (25.0) |
Big city | 6 (27.3) | 9 (31.0) | 28 (50.0) | 28 (33.3) |
A small town | 11 (50.0) | 14 (48.3) | 7 (12.5) | 27 (32.1) |
Village | | | 13 (23.2) | 8 (9.5) |
Financial situation before war, n(%) | χ2 = 2.95, p = 0.262 | χ2 = 2.53, p = 0.513 | 4.31 (0.255) |
Very good | 2 (9.1) | 7 (24.1) | 11 (19.6) | 20 (23.8) |
Good enough | 13 (59.1) | 19 (65.5) | 32 (57.1) | 52 (61.9) |
Not very good | 4 (18.2) | 2 (6.9) | 3 (5.4) | 4 (4.8) |
Very bad | | | | 4 (4.8) |
Financial situation after war, n(%) | χ2 = 3.26, p = 0.635 | χ2 = 4.09, p = 0.423 | 2.13 (0.772) |
Much better | 1 (4.5) | | | 3 (3.6) |
Improved | 2 (9.1) | 1 (3.4) | 2 (3.6) | 2 (2.4) |
Unchanged | | 1 (3.4) | | 3 (3.6) |
Somewhat worsened | 7 (31.8) | 13 (44.8) | 19 (33.9) | 30 (35.7) |
Much worsened | 4 (18.2) | 6 (20.7) | 16 (28.6) | 23 (27.4) |
Predominantly, individuals in both categories are either partnered or married (74.8%), with no significant differences (p = [0.326; 0.933]), highlighting that family ties likely remain stable despite the challenges of displacement. At the same time, there is a shift in education level among the discussed sample in general (p = [0.104; 0.743]) with prevalence of respondents participating with completed higher education – bachelor’s (25.7%) or master’s degree (50.7%) as well. Moreover, the Chi-squared test indicates no significant differences in financial status for different groups both before and after the Russian invasion (p = [0.255; 0.772]). This finding is further underscored by a low Gini Coefficient of 0.153, indicating minimal inequality among respondents. The substantial proportion of individuals reporting worsened financial conditions after full-scale invasion among all subgroups – 61.7% from general sample – provides a stark contrast, pointing to the severe economic difficulties faced by war affected PNM.
The analysis of whether the current pregnancy is the first one for the respondents shows a marginal trend towards significance (р = 0.082) suggesting a possible variation in family planning or birth intervals that does not reach conventional levels of statistical significance, but is still shaping migration processes (See Table 3 [near here]). There is lack of data on other types of pregnancy despite singleton – indicated for only 6,3% of respondents – which reveals possible limitations of presented design. Still, we can assume that the incidence of complications during pregnancy presents a stark contrast. Particularly, among new mothers, where the chi-square values reveal a significant difference (χ2 = 13.66, p = 0.004). New mothers REF had significantly fewer complications compared to IDP new mothers. Concerns regarding medical care access were primarily reported only by new mothers that had to refuge from Ukraine (3.6%). The ratio of chosen main stressors, highlighted at Fig. 3 [near here], shows prosocial tendency among the PNM population. Naturally, REF are less likely to refer to safety issues (12.4%) compared to IDP (33.3%). Still REF are more often worried about the war outcomes on relatives (35.7%) and community (27.2%), unlike the IDP (30.2%, 15.9%, respectively).
Table 3
Comparison of birth-related characteristics of PNM groups
Variables | Pregnant | New Mothers | χ2 (p) |
IDP | REF | IDP | REF |
Respondents, n | 22 | 29 | 56 | 84 | |
Pregnancy type, n(%) | χ2 = 0.08, p = 0.999 | χ2 = 1.66, p = 0.729 | 0.91 (0.893) |
Singleton | 21 (95.5) | 27 (93.2) | 49 (87.5) | 81 (96.4) |
Twin | | 1 (3.4) | 1 (1.8) | 2 (2.4) |
N/A | 1 (4.5) | 1 (3.4) | 5 (8.9) | 1 (1.2) |
Other Children | χ2 = 3.47, p = 0.077 | χ2 = 3.41, p = 0.079 | 3.48 (0.082) |
First child | 16 (72.7) | 14 (48.3) | 35 (62.5) | 43 (51.2) |
Not first | 5 (22.7) | 14 (48.3) | 17 (30.4) | 41 (48.8) |
N/A | 1 (4.5) | 1 (3.4) | 4 (7.14) | |
Complications | χ2 = 0.34, p = 0.911 | χ2 = 13.66, p = 0.004 | 6.74 (0.529) |
None | 16 (72.7) | 20 (68.9) | 25 (44.6) | 55 (65.5) |
One | 3 (13.6) | 5 (17.2) | 11 (19.6) | 16 (19.1) |
Two | 2 (9.1) | 4 (13.8) | 8 (14.3) | 8 (9.5) |
Three or more | 1 (4.6) | | 11 (19.6) | 4 (4.7) |
N/A | | | 1 (1.8) | 1 (1.2) |
Though PNM subgroups assess war’s impact on mental health differently, they seek mental health care similarly (See Table 4 [near here]). Slight differences among new mothers and pregnant women regarding this question with p-values reaching for significance threshold (p = 0.059) are not supported by Fisher’s Exact Test results (p = 0.083), indicating dispersion disproportion among groups. 92.9% of IDP and 91.4% of REF as well stated that they did not intend to seek help. The highest percentage of women who received treatment due to mental health problems was among REF (7.6%); in contrast to IDP (4.8%).
Table 4
Comparison of mental health seeking behaviour among PNM groups
Variables | Pregnant | New Mothers | χ2 (p) |
IDPs | REFs | IDPs | REFs |
Receiving mental health care, n(%) | χ2 = 0.16, p = 0.999 | χ2 = 2.69, p = 0.265 | 1.55 (0.059) |
Yes | 2 (11.1) | 2 (7.1) | 1 (2.2) | 6 (7.8) |
No | 17 (88.9) | 26 (92.9) | 42 (93.3) | 70 (90.9) |
I don’t want to answer | | | 2 (4.4) | 1 (1.3) |
Figure 4 [near here] represents preferred ways of coping with the stressful impact of war. PNM selected sociable contacts and activities more often, although their sociability may be predicted by multiple routes. The REF subgroup primarily focused on ‘being in touch’, as noted by one of participants, saving ability to interact and staying up to date informed. REF most commonly cope by chatting with peers (30.5%), spending time on social media (14.5%), and monitoring international news (11%). IDP, unlike REF, preferred prosocial coping including aiding others (4.3% compared to 2.1% among REF), communicating with peers (30.1%) or other PNM (16.2%), and decreasing social media usage (10.0%). IDPs may have been more focused on ‘being present’, as a means of restoration.
Pregnant women tend not to use substances to cope with stress, whereas 3.1% of new mothers prefer this option. Pregnant women more frequently opt for social interactions, while new mothers are more inclined to choose eating comfort food (13.4%), communicating with other pregnant women or parents (18.6%), and increasing time spent on social networks (10.3%). Additionally, new mothers are less likely to choose coping strategies that enhance self-care (such as sleep, meditation, sports, or reading) compared to pregnant women (4.1% compared to 11.2%).
An ordinal logistic regression model was fitted using the cumulative link model to assess the predictors of self-reported distress levels (See Table 5 [near here]). The absence of the Hauck-Donner effect in any of the estimates indicates that the coefficient estimates are reliable. Based on this model, we can assume main perceived stressors in post-invasion experiences of PNM from Ukraine. Concerns regarding the financial consequences of war on respondents’ households were significantly associated with higher levels of general distress in daily life (℮ = 0.382, p < 0.001, OR = 1.465). The perceived stress from financial burdens potentially linked to participants’ original place of residence, as detailed in Fig. 5 and Fig. 6 [near here]. Additionally, respondents who reported negative changes in their financial situation experienced a 75.4% increase in perceived distress (℮ = 0.562, p < 0.001, OR = 1.754).
Table 5
Ordinal modelling results for perceived distress level predictors
Variable | Estimate | Std. Error | z value | Pr(>|z|) | P-value |
Financial situation changes | 0.562 | 0.169 | 3.320 | 0.001 | < 0.001 |
Finances-related stress | 0.382 | 0.055 | 6.967 | 3.23e− 12 | < 0.001 |
Lack of support sources | 0.143 | 0.052 | 2.727 | 0.006 | < 0.01 |
Disruption of social support network | 0.313 | 0.051 | 6.130 | 8.81e− 10 | < 0.001 |
Original place of residence | 0.249 | 0.060 | 5.121 | 0.001 | < 0.001 |
Higher ratings of lacking support sources were also associated with increased levels of tension (℮ = 0.143, p < 0.01, OR = 1.154), indicating that the odds of reporting higher distress levels rose by 15.4% for each unit increase in the lack of support sources. Similarly, each one-level increase in the assessment of disruption in the social support network was associated with a 36.7% increase in distress levels (℮ = 0.313, p < 0.001, OR = 1.367). From this point, we can assume that social support as well as financial stability could possibly serve as protector factors of wartime related distress in PNM (OR > 1). Moreover, based on the cumulative link model, the effect of various predictors directly on respondents’ perceived mental well-being was evaluated (see Table 6 [near here]). Concerns about a child’s health due to the war were a primary negative predictor of mental well-being (℮ = 1.095, p < 0.001, OR = 0.334). This indicates that those with concerns about their child’s health due to the war are 66.4% more likely to report a worsening in mental well-being and are likely more ready to relocate. The place of origin residence (℮ = 0.780, p < 0.001, OR = 0.458) was also a predictor of well-being. Those who had migration status of refugee were 54.2% of rose in distress levels, as visualised in Fig. 7 [near here].
Table 6
Ordinal modelling results for perceived mental well-being predictors
Variable | Estimate | Std. Error | z value | Pr(>|z|) | P-value |
Concerns on child’s health | 1.095 | 0.314 | 3.489 | 0.001 | < 0.001 |
Inability to receive medical care | 0.679 | 0.294 | 2.309 | 0.021 | < 0.05 |
Concerns on family involvement | 0.630 | 0.308 | 2.044 | 0.041 | < 0.05 |
Original place of residence | 0.780 | 0.147 | 2.439 | 0.041 | < 0.001 |
Additionally, individuals who encountered difficulties accessing medical care (℮ = 0.679, p < 0.05, OR = 1.971) and those who expressed concerns about potential changes in family support (℮ = 0.630, p < 0.05, OR = 0.533) reported lower well-being following the full-scale invasion. The odds ratio for the access to medical care variable, contrasting with concerns about changes in family support, was greater than 1, suggesting that the perceived stressors associated with these changes were predominantly related to family support, while access to medical care could serve as a protective factor for PNM during wartime. Overall, the regression analysis indicates that financial burdens, the original location of the household, concerns about a child’s health, access to medical care, and expectations regarding future family involvement in childcare were significant predictors of perceived war-related distress, ultimately contributing to reports of deteriorated mental health.