Key Points
Question
To what extent do standard socioeconomic factors explain health disparities in a minoritized youth population that shares ethnicity and citizenship with the host society?
Findings
In this cross-sectional study of 876 693 adolescents, youths of North Korean refugee origin living in South Korea exhibited 4-fold higher suicide attempts and 10-fold higher drug use than peers. Conventional socioeconomic factors accounted for less than 10% of these disparities.
Meaning
These findings suggest standard socioeconomic frameworks may not fully capture fundamental drivers of inequality in integrated minoritized populations, indicating a need to address unmeasured structural determinants beyond economic support.
This cross-sectional study investigates the extent to which socioeconomic factors explain health disparities in a minoritized, refugee youth population that shares ethnicity and citizenship with the host society.
Abstract
Importance
Conventional frameworks often attribute migrant health disparities to language barriers, legal exclusion, or socioeconomic deprivation. However, it remains unclear whether health inequities persist when these structural barriers are removed. Isolating the impact of historical and intergenerational trauma requires examining a unique minoritized population that shares full citizenship and linguistic heritage with the host society.
Objective
To quantify the explanatory power of conventional demographic, socioeconomic, and family determinants on health disparities in a fully integrated refugee group and to estimate the magnitude of the unexplained residual disparity attributed to nonsocioeconomic factors.
Design, Setting and Participants
This cross-sectional study analyzed data from the Korea Youth Risk Behavior Survey (2011 to 2024), a nationally representative survey. The study used North Korean refugee-origin youths (NKRY) as a strategic case study of a population sharing ethnicity and citizenship with the host country (South Korean-origin youth, SKY), compared alongside international-origin youths (IOY).
Main Outcomes and Measures
Self-reported mental health outcomes and health risk behaviors. Fairlie decomposition analysis was performed to calculate the percentage of the disparity explained by observed conventional determinants vs the unexplained residual component.
Results
Among 876 693 participants, SKY (n = 785 462; 392 731 [51.0%] male; median age, 14.97 years), NKRY (n = 547; 317 [58.25%] male; 15.61 years), and IOY (n = 13 800; 6675 [49.45%] male; 14.55 years) were included. NKRY had substantially poorer outcomes compared with SKY. Adjusted odds ratios for suicide planning (aOR, 4.35; 95% CI, 2.97-6.35), suicide attempts (aOR, 4.27; 95% CI, 2.78-6.56), and drug use (aOR, 14.09; 95% CI, 7.48-26.51) remained high. Decomposition analyses revealed that conventional socioeconomic factors explained less than 10% of these disparities. For suicide-related outcomes, observed factors accounted for less than 5% of the gap, leaving more than 95% unexplained. Conversely, disparities in the IOY group were largely explained by these conventional factors.
Conclusions and relevance
In this cross-sectional study, substantial health inequities persisted even in a minoritized population granted full legal and linguistic integration, with the vast majority of disparities remaining unexplained by conventional socioeconomic metrics. These findings challenge the sufficiency of standard social determinant frameworks and suggest that for displaced populations worldwide, policies must address unmeasured structural drivers such as historical trauma and intergenerational instability.
Introduction
Social determinants of health (SDH) frameworks explain many health disparities. Yet, substantial inequalities in mental health persist among minoritized youths even after accounting for conventional socioeconomic factors.1,2,3 Household resources and school contexts are critical, but they do not tell the whole story.4,5 Significant gaps often remain in outcomes such as suicidality, depression, and substance use.2,6 Most studies attribute these residual disparities to unmeasured confounding or incomplete data. Consequently, policy interventions continue to focus primarily on filling socioeconomic deficits.7,8
Furthermore, it remains unclear whether these conventional models adequately explain disparities for minoritized populations who already share citizenship and institutional access with the majority. In such groups, where common confounders like legal or linguistic barriers are mitigated,9 examining these populations can help eliminate primary structural constraints and shed light on these unexplained residual disparities.
Youths of North Korean refugee origin (NKRY) in South Korea represent a distinctively informative case study for global health. They account for less than 0.1% of the total population (cumulative No. >34 000 as of 2024). Yet, they represent a distinct and socially significant minoritized group. Under the South Korean Constitution, NKRY are recognized as nationals and granted full citizenship immediately upon arrival. NKRY’s immediate citizenship and shared heritage make them an ideal population for providing a stringent test of the SDH framework.10
Specifically, this unique context allows us to isolate the impact of cumulative life-course exposures, such as traumatic migration routes and the intergenerational legacy of political repression and famine. These factors persist beyond the reach of conventional socioeconomic explanations. At the same time, these youths are shaped by structural histories of famine, political repression, and displacement. This applies whether they were born in North Korea, third countries, or South Korea. They share the same institutional environment as their peers but remain exposed to the intergenerational transmission of trauma.7
Prior research suggests that health outcomes in such populations are driven by structural determinants. These are not reducible to standard household-level indicators.11,12,13 Yet, empirical research has rarely quantified exactly how much of the disparity remains unexplained. To bridge this gap, this study analyzes nationally representative data. We aim to quantify mental health and behavioral disparities between NKRY and their South Korean peers. Specifically, we use decomposition analysis to empirically evaluate the limits of conventional SDH frameworks. We assess whether observable demographic, family, and school factors can account for these disparities, or whether substantial inequalities persist beyond the scope of prevailing models.
Methods
Data Collection and Sample
This cross-sectional study analyzed data from the Korea Youth Risk Behavior Survey (KYRBS). This is a nationally representative, school-based survey conducted annually by the Korea Disease Control and Prevention Agency. We utilized data from 2011 to 2024. The KYRBS employs a multistage, stratified cluster sampling design to ensure national representativeness. Sampling weights account for selection probabilities and nonresponse. All analyses incorporated this complex survey design.14 Data were collected using anonymous, self-administered questionnaires. The study was approved by Korea University’s institutional review board and the Korea Disease Control and Prevention Agency obtained informed consent during primary data collection. The institutional review board exempted this analysis of anonymized secondary data. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
Study Population and Independent Variables
A total of 876 693 adolescents were included across the study period. We classified participants into 3 groups based on parental origin. The groups were South Korean-origin youth (SKY; n = 785 462), NKRY (n = 547), and international-origin youth (IOY; n = 13 800). Adolescents with missing parental origin data were excluded. For primary analyses, we compared SKY and NKRY. SKY served as the reference group. NKRY were defined as adolescents with at least 1 parent from North Korea. Importantly, all NKRY in this sample are legally recognized citizens. They attend regular South Korean schools. IOY were included as a contextual comparator. This enabled us to differentiate disparities specific to NKRY from those generally observed in other migrant backgrounds. Detailed information on the composition of the family structure is provided in the eTables 1 in Supplement 1.
Outcome Variables: Mental Health and Health Behaviors
Mental health outcomes included perceived stress, poor self-rated health, and sleep problems. We also measured unhappiness, depressive symptoms, and loneliness. Suicidality was assessed through 3 distinct items: suicidal ideation, suicide planning, and suicide attempts. Anxiety was measured using the Generalized Anxiety Disorder-7 scale.
Health risk behaviors included current smoking (cigarettes, e-cigarettes, heated tobacco), current alcohol use, and drug use. We also assessed physical inactivity, sexual experience, and violence victimization. All outcomes were coded as binary variables. Anxiety was dichotomized using standard cut-off scores.15 Detailed information on survey items, question wording, and years of inclusion or exclusion for each variable is provided in the eTable 2 in Supplement 1.
Explanatory Variables: Conventional Social Determinants.
We selected routinely measured factors common in national surveys. These represented conventional social determinants. Variables included sex, age, and family structure. Socioeconomic status was measured via perceived family economic status and parental education. Perceived family economic status was included given evidence that adolescents may lack accurate knowledge of objective household income and that subjective economic position is a robust predictor of mental health outcomes.16 School context included grade, perceived academic achievement, and age-grade difference. Age-grade difference served as a proxy for disrupted educational trajectories. All explanatory variables were selected a priori to represent conventionally measured social contexts, thereby allowing assessment of their explanatory limits.
Statistical Analysis
We first summarized characteristics and outcomes using survey-weighted proportions. We then examined associations using survey-weighted logistic regression. Crude models were estimated first. Subsequently, we ran adjusted models controlling for age, sex, family structure, school grade, age-grade difference, academic achievement, economic status, and parental education. Adjusted odds ratios (aORs) and 95% CIs are reported, with detailed model results presented in the eTable 3 and 4 in Supplement 1.
To quantify the explanatory power of these variables, we conducted Fairlie decomposition analyses.17 This method decomposes the gap between 2 groups into an explained component and an unexplained component. The explained component reflects differences in the distribution of measured variables. The unexplained component captures the residual disparity. We estimated the contribution of individual variables and calculated 95% CIs using bootstrap resampling. All analyses were conducted using R version 4.4.2 (R Project for Statistical Computing). Statistical significance was defined as a 2-sided P < .05 or 95% CIs that did not overlap 1.00 for the aORs.
Results
Table 1 presents the weighted characteristics of the study population. Among 876 693 participants, SKY (n = 785 462; 392 731 [51.0%] male; 14.97 years), NKRY (n = 547; 317 [58.25%] male; 15.61 years), and IOY (n = 13 800; 6675 [49.45%] male; 14.55 years) were included. NKRY exhibited markedly disadvantaged profiles compared with SKY. Single-parent families were significantly more common among NKRY (136 [23.63%]) than SKY (47 183 [5.54%]). Educational disruption was also evident. Pronounced age-grade differences of 3 years or more were observed in 14 NKRY (2.97%; 95% CI, 1.17%-4.77%) compared with only 0.04% SKY (301; 95% CI, 0.04%–0.05%). Perceived family economic status was more polarized among refugee youths. Parental educational attainment was generally lower. In contrast, IOY showed intermediate profiles. They were sociodemographically closer to SKY than to NKRY.
Table 1. Sociodemographic and Economic Characteristics of Youth by Parental Origin, 2011 to 2024.
| Variables | SKY (n = 785 462) | NKRY (n = 547) | IOY (n = 13 800) | |||
|---|---|---|---|---|---|---|
| No. and weighted % (95% CI) |
No. and weighted % (95% CI) |
No. and weighted % (95% CI) |
||||
| Family structure | ||||||
| Single-parent household | 47 183 (5.54 [5.49-5.60]) |
136 (23.63 [19.63-27.63]) |
1071 (8.34 [7.79-8.88]) |
|||
| 2-parent household | 738 279 (94.46 [94.40-94.51]) |
411 (76.37 [72.37-80.37]) |
12 729 (91.66 [91.12-92.21]) |
|||
| Sex | ||||||
| Male | 392 731 (51.00 [50.88-51.13]) |
317 (58.25 [53.57-62.94]) |
6675 (49.45 [48.50-50.40]) |
|||
| Female | 392 731 (49.00 [48.87-49.12]) |
230 (41.75 [37.06-46.43]) |
7125 (50.55 [49.60-51.50]) |
|||
| Age, mean (SD) | 14.97 (1.75) | 15.61 (1.81) | 14.55 (1.71) | |||
| School grade | ||||||
| Middle school, year 1 | 136 055 (16.17 [16.09-16.25]) |
65 (11.25 [8.43-14.07]) |
3469 (22.83 [22.06-23.60]) |
|||
| Middle school, year 2 | 135 135 (16.32 [16.24-16.40]) |
99 (17.15 [13.73-20.57]) |
2953 (19.59 [18.87-20.32]) |
|||
| Middle school, year 3 | 135 033 (16.58 [16.50-16.67]) |
99 (20.04 [16.19-23.89]) |
2644 (18.38 [17.66-19.10]) |
|||
| High school, year 1 | 128 851 (17.03 [16.95-17.12]) |
87 (15.09 [11.83-18.34]) |
1907 (15.66 [14.95-16.38]) |
|||
| High school, year 2 | 127 087 (16.94 [16.85-17.02]) |
82 (12.96 [9.98-15.94]) |
1530 (12.57 [11.92-13.22]) |
|||
| High school, year 3 | 123 301 (16.96 [16.87-17.05]) |
115 (23.52 [19.20-27.85]) |
1297 (10.97 [10.34-11.60]) |
|||
| Age-grade difference | ||||||
| None | 408 850 (51.99 [51.87-52.10]) |
201 (37.79 [33.20-42.39]) |
7400 (53.19 [52.24−54.13]) |
|||
| 1 y | 373 439 (47.6 [47.48-47.72]) |
163 (28.77 [24.54-33]) |
6002 (43.25 [42.31-44.19]) |
|||
| 2 y | 1029 (0.13 [0.12-0.13]) |
18 (2.42 [1.13-3.72]) |
70 (0.56 [0.42-0.71]) |
|||
| ≥3 y | 301 (0.04 [0.04-0.05]) |
14 (2.97 [1.17-4.77]) |
36 (0.35 [0.22-0.47]) |
|||
| Missing | 1843 (0.25 [0.24-0.26]) |
151 (28.04 [23.76-2.32]) |
292 (2.65 [2.32-2.98]) |
|||
| Perceived academic achievement | ||||||
| High | 97 837 (12.35 [12.27-12.43]) |
151 (28.44 [24.19-32.68]) |
1151 (8.32 [7.80-8.85]) |
|||
| Middle-high | 198 203 (25.23 [25.12-25.33]) |
65 (13.48 [10.05-16.91]) |
2757 (19.77 [19.02-20.53]) |
|||
| Middle | 224 346 (28.65 [28.54-28.76]) |
103 (18.40 [14.66-22.14]) |
3984 (28.19 [27.35-29.04]) |
|||
| Middle-low | 183 554 (23.41 [23.31-23.52]) |
106 (17.26 [13.84-20.68]) |
4031 (29.32 [28.46-30.18]) |
|||
| Low | 81 522 (10.37 [10.29-10.44]) |
122 (22.42 [18.51-26.33]) |
1877 (14.38 [13.71-15.06]) |
|||
| Perceived family economic status | ||||||
| High | 72 441 (9.36 [9.29-9.43]) |
116 (22.19 [18.28-26.10]) |
937 (7.38 [6.87-7.88]) |
|||
| Middle-high | 216 909 (28.03 [27.92-28.14]) |
51 (11.03 [7.80-14.27]) |
2326 (17.24 [16.52-17.95]) |
|||
| Middle | 370 668 (46.87 [46.75-46.99]) |
152 (26.63 [22.48-30.79]) |
6975 (49.44 [48.50-50.39]) |
|||
| Middle-low | 101 818 (12.77 [12.69-12.85]) |
120 (20.50 [16.74-24.25]) |
2803 (20.08 [19.32-20.83]) |
|||
| Low | 23 626 (2.97 [2.93-3.01]) |
108 (19.65 [15.93-23.36]) |
759 (5.87 [5.41-6.33]) |
|||
| Maternal education, y | ||||||
| ≤9 | 17 401 (2.07 [2.04-2.11]) |
115 (23.81 [19.70-27.92]) |
1099 (7.98 [7.47-8.49]) |
|||
| 12 | 263 311 (33.44 [33.32-33.55]) |
125 (19.37 [15.77-22.96]) |
3502 (25.71 [4.88-26.54]) |
|||
| ≥14 | 359 719 (47.16 [47.04-47.28]) |
116 (21.22 [17.28-25.15]) |
3807 (28.96 [28.09-29.82]) |
|||
| Unknown | 123 353 (14.86 [14.78-14.95]) |
142 (25.78 [21.68-29.89]) |
5162 (35.32 [34.42-36.22]) |
|||
| Missing | 21 678 (2.47 [2.43-2.51]) |
49 (9.83 [6.99-12.66]) |
230 (2.03 [1.75-2.32]) |
|||
| Paternal education, y | ||||||
| ≤9 | 18 867 (2.21 [2.17-2.24]) |
136 (27.75 [23.39-32.12]) |
1583 (11.14 [10.54-11.74]) |
|||
| 12 | 212 781 (26.49 [26.39-26.59]) |
89 (15.87 [12.45-19.30]) |
4107 (29.39 [28.53-30.26]) |
|||
| ≥14 | 397 899 (52.59 [52.47-52.70]) |
93 (17.12 [13.61-20.62]) |
2970 (23.60 [22.78-24.41]) |
|||
| Unknown | 130 410 (15.64 [15.56-15.73]) |
142 (25.44 [21.39-29.50]) |
4299 (29.57 [28.72-30.42]) |
|||
| Missing | 25 505 (3.07 [3.03-3.12]) |
87 (13.81 [10.60-17.02]) |
841 (6.30 [5.82-6.78]) |
|||
| Maternal nativity | ||||||
| Born outside Korea | NA | 467 (85.13 [81.77-88.49]) |
12 729 (90.91 [90.35-91.48]) |
|||
| Born in Korea | 763 784 (97.53 [97.49-97.57]) |
31 (5.04 [3.01-7.07]) |
841 (7.05 [6.55-7.56]) |
|||
| Missing | 21 678 (2.47 [2.43-2.51]) |
49 (9.83 (6.99-12.66]) |
230 (2.03 [1.75-2.32]) |
|||
| Paternal nativity | ||||||
| Born outside Korea | NA | 401 (74.84 [70.74-78.94]) |
2615 (22.06 [21.25-22.87]) |
|||
| Born in Korea | 759 957 (96.93 [96.88-96.97]) |
59 (11.35 [8.30-14.4]) |
10 344 (71.64 [70.76-72.51]) |
|||
| Missing | 25 505 (3.07 [3.03-3.12]) |
87 (13.81 [10.60-17.02]) |
841 (6.30 [5.82-6.78]) |
|||
Abbreviation: IOY, international-origin youth; NA, not applicable; NKRY, North Korean refugee origin youth; SKY, South Korean youth; ; .
Table 2 details the prevalence of adverse outcomes. NKRY showed consistently poorer mental health profiles. The largest disparities were in suicide-related outcomes. The prevalence of suicide attempts among NKRY was 16.0% (86 individuals). This is starkly higher than the 2.92% (23 122 individuals) observed in SKY. Elevated prevalence was also found for depression, anxiety, and severe stress. Risk behaviors followed a similar pattern. Conventional cigarette use was 40.64% (220 individuals) among NKRY vs 16.0% (123 101 individuals) among SKY. Alcohol use was 55.35% (313 individuals) vs 40.19% (312 575 individuals). Drug use showed the most dramatic difference, with prevalence rates more than 10 times higher in the refugee group. IOY generally exhibited intermediate prevalence, with much smaller gaps relative to the host population.
Table 2. Disparities in Mental Health and Health Behavior Among Youth by Parental Origin, 2011 to 2024.
| Variables | SKY (n = 785 462) | NKRY (n = 547) | IOY (n = 13 800) | |||
|---|---|---|---|---|---|---|
| No. and weighted % (95% CI) |
No. and weighted % (95% CI) |
No. and weighted % (95% CI) |
||||
| Perceived stress | ||||||
| Average or below | 477 686 (60.86 [60.75-60.98]) |
304 (56.28 [51.60-60.95]) |
8459 (61.09 [60.17-62.01]) |
|||
| Above average | 307 776 (39.14 [39.02-39.25]) |
243 (43.72 [39.05-48.40]) |
5341 (38.91 [37.99-39.83]) |
|||
| Perceived health | ||||||
| Average or better | 728 097 (92.67 [92.61-92.73]) |
475 (86.46 [83.26-89.67]) |
12 455 (89.93 [89.36-90.51]) |
|||
| Below average | 57 365 (7.33 [7.27-7.39]) |
72 (13.54 [10.33-16.74]) |
1345 (10.07 [9.49-10.64]) |
|||
| Sleep problems | ||||||
| No sleep problems | 461 989 (58.41 [58.30-58.53]) |
304 (55.04 [50.34-59.74]) |
8730 (62 [61.08-62.92]) |
|||
| Sleep problems | 323 473 (41.59 [41.47-41.70]) |
243 (44.96 [40.26-49.66]) |
5070 (38 [37.08-38.92]) |
|||
| Unhappiness | ||||||
| No | 507 621 (89.94 [89.85-90.03]) |
320 (83.56 [79.43-87.68]) |
6759 (88.34 [87.52-89.15]) |
|||
| Yes | 56 247 (10.06 [9.97-10.15]) |
66 (16.44 [12.32-20.57]) |
874 (11.66 [10.85-12.48]) |
|||
| NCa | 221 594 (NA) | 161 (NA) | 6167 (NA) | |||
| Depression | ||||||
| No | 569 344 (72.48 [72.37-72.59]) |
327 (58.24 [53.54-62.95]) |
9868 (70.77 [69.9-71.64]) |
|||
| Yes | 216 118 (27.52 [27.41-27.63]) |
220 (41.76 [37.05-46.46]) |
3932 (29.23 [28.36-30.10]) |
|||
| Loneliness | ||||||
| No | 174 520 (82.79 [82.61-82.97]) |
96 (79.00 [71.26-86.74]) |
5883 (80.66 [79.63-81.69]) |
|||
| Yes | 36 098 (17.21 [17.03-17.39) |
28 (21.00 [13.26-28.74]) |
1367 (19.34 [18.31-20.37]) |
|||
| NCa | 574 844 (NA) | 423 (NA) | 6550 (NA) | |||
| Anxiety | ||||||
| No | 182 196 (86.45 [86.29-86.61]) |
92 (73.73 [65.29-82.16]) |
6237 (85.58 [84.66-86.50]) |
|||
| Yes | 28 422 (13.55 [13.39-13.71]) |
32 (26.27 [17.84-34.71]) |
1013 (14.42 [13.5-15.34]) |
|||
| NCa | 574 844 (NA) | 423 (NA) | 6550 (NA) | |||
| Suicide ideation | ||||||
| No | 674 836 (85.88 [85.80-85.97]) |
397 (73.05 [68.87-77.22]) |
11 670 (83.98 [83.28-84.68]) |
|||
| Yes | 110 626 (14.12 (14.03-14.2]) |
150 (26.95 (22.78-31.13]) |
2130 (16.02 (15.32-16.72]) |
|||
| Suicide planning | ||||||
| No | 748 951 (95.36 [95.31-95.41]) |
445 (80.37 [76.54-84.20]) |
12 934 (93.33 [92.85-93.81]) |
|||
| Yes | 36 511 (4.64 [4.59-4.69]) |
102 (19.63 [15.80-23.46]) |
866 (6.67 [6.19-7.15]) |
|||
| Suicide attempt | ||||||
| No | 762 340 (97.08 [97.04-97.13]) |
461 (84.0 [80.57-87.43]) |
13 148 (94.87 [94.44-95.30]) |
|||
| Yes | 23 122 (2.92 [2.87-2.96]) |
86 (16.0 [12.57-19.43]) |
652 (5.13 [4.70-5.56]) |
|||
| Height, mean (SD), cm | 165.1 (8.44) | 166.04 (9.07) | 163.27 (8.54) | |||
| Weight, mean (SD), kg | 57.68 (12.18) | 59.50 (12.03) | 56.98 (13.22) | |||
| Conventional cigarette | ||||||
| No | 662 361 (84 [83.91-84.09]) |
327 (59.36 [54.69-64.02]) |
11 961 (85.24 [84.54-85.93]) |
|||
| Yes | 123 101 (16.0 [15.91-16.09]) |
220 (40.64 [35.98-45.31]) |
1839 (14.76 [14.07-15.46]) |
|||
| Electronic cigarette | ||||||
| No | 727 585 (92.33 [92.26-92.39]) |
374 (65.75 [61.20-70.31]) |
12 737 (91.09 [90.52-91.65]) |
|||
| Yes | 57 877 (7.67 [7.61-7.74]) |
173 (34.25 [29.69-38.8]) |
1063 (8.91 [8.35-9.48]) |
|||
| Heated tobacco product | ||||||
| No | 296 552 (96.76 [96.69-96.83]) |
161 (77.8 [70.96-84.64]) |
8594 (96.17 [95.71-96.64]) |
|||
| Yes | 9637 (3.24 [3.17-3.31]) |
37 (22.2 [15.36-29.04]) |
307 (3.83 [3.36-4.29]) |
|||
| NCa | 479 273 (NA) | 349 (NA) | 4899 (NA) | |||
| Alcohol use | ||||||
| No | 472 887 (59.81 [59.70-59.92]) |
234 (44.65 [39.95-49.35) |
9362 (66.06 [65.15-66.97) |
|||
| Yes | 312 575 (40.19 [40.08-40.30]) |
313 (55.35 [50.65-60.05]) |
4438 (33.94 [33.03-34.85]) |
|||
| Drug use | ||||||
| No | 167 097 (98.87 [98.82-98.93]) |
92 (86.96 [79.62-94.31]) |
6090 97.84 [97.40-98.28]) |
|||
| Yes | 1884 (1.13 [1.07-1.18]) |
12 (13.04 [5.69-20.38]) |
115 2.16 [1.72-2.60]) |
|||
| NCa | 616 481 (NA) | 443 (NA) | 7595 (NA) | |||
| Physical activity (1 week) | ||||||
| ≥72h | 526 176 (67.38 [67.27-67.49]) |
332 (59.14 [54.47-63.81]) |
9143 (66.66 [65.77-67.55]) |
|||
| <72h | 259 286 (32.62 [32.51-32.73]) |
215 (40.86 [36.19-45.53]) |
4657 (33.34 [32.45-34.23]) |
|||
| Sexual experience | ||||||
| No | 750 384 (95.39 [95.33-95.44]) |
380 (68.77 [64.37-73.18]) |
12 975 (93.0 [92.49-93.51]) |
|||
| Yes | 35 078 (4.61 [4.56-4.67]) |
167 (31.23 [26.82-35.63]) |
825 (7.0 [6.49-7.51]) |
|||
| Experiences of violence | ||||||
| No | 686 207 (96.51 [96.47-96.56]) |
371 (72.45 [68.06-76.84]) |
12 411 (92.91 [92.4-93.41]) |
|||
| Yes | 25 083 (3.49 [3.44-3.53]) |
132 (27.55 [23.16-31.94]) |
886 (7.09 [6.59-7.60]) |
|||
| NCa | 74 172 (NA) | 44 (NA) | 503 (NA) | |||
| Asthma | ||||||
| No | 723 308 (92.0 [91.93-92.06]) |
462 (83.88 [80.38-87.38]) |
12 858 (92.82 [92.33-93.32]) |
|||
| Yes | 62 154 (8.0 [7.94-8.07]) |
85 (16.12 [12.62-19.62]) |
942 (7.18 [6.68-7.67]) |
|||
| Atopic dermatitis | ||||||
| No | 599 437 (76.18 (76.07-76.28]) |
457 (82.51 (78.86-86.16]) |
11 457 (82.69 (81.97-83.40]) |
|||
| Yes | 186 025 (23.82 (23.72-23.93]) |
90 (17.49 (13.84-21.14]) |
2343 (17.31 (16.6-18.03]) |
|||
Abbreviation: IOY, international-origin youth; NA, not applicable; NC, not collected; NKRY, North Korean refugee–origin youth; SKY, South Korean youth; .
NC indicates outcomes not assessed in specific survey years due to changes in Korea Youth Risk Behavior Survey questionnaire content. Survey years for each outcome are provided in eTable 2 in Supplement 1.
Figure 1 displays the results of multivariable logistic regression. Even after adjusting for all conventional socioeconomic and school factors, NKRY remained at substantially elevated risk. Compared with SKY, NKRY had over 4-fold higher odds of suicide planning (aOR, 4.35; 95% CI, 2.97-6.35). Suicide attempts showed a similar magnitude of risk (aOR, 4.27; 95% CI, 2.78-6.56). Suicidal ideation remained 2-fold higher (aOR, 2.00; 95% CI, 1.47-2.72). Significant risks persisted for behavioral outcomes as well. The adjusted odds ratio for drug use was 14.09 (95% CI, 7.48-29.69). Sexual experience (aOR, 7.01; 95% CI, 4.94-9.94) and violence victimization (aOR, 5.90; 95% CI, 4.03-8.65) also remained significantly elevated. In contrast, associations for IOY were generally weaker and less consistent.
Figure 1. Adjusted Associations Between Parental Origin and Mental Health and Health Behavior Outcomes Among Youth: Survey-Weighted Logistic Regression Analysis, 2011 to 2024.
The reference group for all variables is South Korean youths.
Figure 2 presents the core findings from the Fairlie decomposition. Mental health disparities were overwhelmingly driven by the unexplained component. For suicide-related outcomes, conventional determinants failed to account for the gap. For suicidal ideation, 98.9% of the disparity was unexplained (absolute contribution, 12.0 percentage points [pp] of the total 12.2 pp gap).
Figure 2. Fairlie Decomposition of Disparities in Mental Health and Health Behaviors Between South Korean (SKY) and North Korean Refugee-Origin Youth (NKRY), 2011 to 2024.
Demographic factors include sex, age, and age grade difference, socioeconomic status (SES) and family factors include perceived family economic status, perceived academic achievement, family structure, maternal and/or paternal education level.
For suicide attempts and planning, the unexplained proportion exceeded 100%. This indicates a counterbalancing association. Some observed characteristics of NKRY (such as age or specific school factors) actually narrowed the gap statistically. However, unmeasured determinants exerted a much larger force in the opposite direction. Across most mental health indicators, more than 80% of the disparity remained unexplained. A similar pattern was observed for violence victimization and substance use. The only exception was alcohol use, where controlled factors explained 48.9% of the total 11.1 pp gap, and age explained a meaningful share (28.4% of the explained gap; absolute contribution, 3.2 pp) of the disparity. Unlike NKRY, IOY disparities were largely not significant. The exceptions were suicide attempts and violence victimization. Conventional SDH factors explained 38.4% and 9.3% of these gaps, respectively (eTables 5 and 6 in Supplement 1).
Discussion
The most striking finding of this cross-sectional study is not merely the magnitude of health disparities. It is the failure of conventional models to explain them. We analyzed a minoritized population that shares citizenship and language with the host society. Yet, we found that standard socioeconomic, family, and school variables accounted for a negligible fraction of the health gap. For suicide planning and attempts, the unexplained component exceeded 95%. This suggests that even if we could perfectly equalize the economic and family conditions of these youths to match their South Korean peers, the disparity would barely move. This unexplained component is not random statistical noise. It signals a systematic blind spot in our current SDH frameworks. These frameworks capture current resources. They fail to capture accumulated history.
Conventional Socioeconomic Explanations
Our findings empirically validate the structural turn in health equity research. Scholarship has shifted from focusing on individual behaviors to identifying structural drivers like historical trauma and political exclusion.11,13,18,19 However, empirical quantification of this limit has been rare.
Our decomposition analyses confirmed that widely used sociodemographic determinants explained only a negligible fraction of the observed disparities. To illustrate this limit concretely, our estimates suggest that even if the socioeconomic and family conditions of NKRY hypothetically matched those of their South Korean peers, the disparity in suicidal ideation would decrease by only a negligible margin (explaining just 1.1% of the total gap). Similarly, family economic status accounted for only roughly 5% of the suicide disparity.
The limited explanatory power of parental education likely reflects institutional nonequivalence. A degree obtained in North Korea lacks the social capital and transferability of one from South Korea. Standard statistical adjustments treat these qualitatively different backgrounds as identical, masking the functional disadvantage these families face. In some models, such as suicide attempts, the unexplained component even exceeded 100%. This indicates a counterbalancing association: while some observed demographic traits of NKRY may be statistically protective, unmeasured structural drivers—such as intergenerational trauma—are so potent that they completely overwhelm these factors. This confirms that disparities in this population are driven by forces that persist far beyond the reach of demographic and economic equalization.
Interpreting the Unexplained Disparity
The predominance of the unexplained component invites careful interpretation. Rather than pointing to a single omitted factor, it signals the presence of interrelated processes that operate beyond the scope of conventional family and socioeconomic indicators. These disparities should not be viewed as mere residual error, but as empirical signals pointing toward structural drivers not readily observable in standard survey data.20
First, these youths carry the weight of historical trauma. Specifically, the unresolved legacy of geopolitical conflict on the Korean Peninsula and the prolonged fear of persecution create a unique form of vulnerability.21 This vulnerability is not reducible to contemporaneous household poverty but stems from accumulated life-course exposures. Second, despite sharing ethnicity with the host society, many of these youths navigate a liminal identity. They often occupy a space where they feel fully accepted by neither North nor South.22 As observed in global refugee research, this identity ambiguity acts as an unmeasured stressor that persists even when legal integration is achieved. Furthermore, recent NKRY born in South Korea or third countries blend in seamlessly. They effectively become an invisible minoritized group. While outward cultural stress may decrease, navigating a hidden identity amid mixed societal attitudes generates distinct minority stress. Future research should examine these evolving dynamics. Third, intergenerational transmission plays a critical role. Youths remain highly permeable to the family emotional climate. Consequently, parental histories of displacement and instability can exert profound impacts on adolescent mental health, even without the child’s direct exposure to the initial trauma.20,23
These dynamics act as unmeasured structural drivers of inequality. Similarly, research on other minoritized populations reports health impacts that standard socioeconomic indicators fail to capture.20 Although rooted in the unique context of the Korean Peninsula, our findings provide a critical basis for understanding health disparities in populations shaped by unmeasured factors, including displacement and trauma. To capture these invisible inequities, national youth surveys must incorporate parental life-course indicators, such as adverse childhood experiences. This will enable a more comprehensive understanding of health disparities beyond conventional models.
NKRY as a Theoretical Model
The rapid demographic transformation of this population offers a critical theoretical insight. The population is no longer composed primarily of direct defectors; it has shifted toward a generation largely born in South Korea or third countries.24 At first glance, this might suggest increasing convergence with South Korean peers in terms of language and material conditions. However, our trend analysis reveals that health risks remain elevated despite this shift (eFigure in Supplement 1). This highlights a fundamental disconnect: institutional integration does not equate to psychosocial protection.
Despite sharing citizenship, compulsory schooling, and formal access to health care with the majority, these adolescents remain embedded in life trajectories shaped by parental displacement and family insecurity.10 Even among youths born in South Korea, the intergenerational transmission of parental trauma and the psychosocial burden of navigating a minoritized identity exert enduring influences on stress regulation and emotional well-being.20 Educational indicators illustrate this persistent vulnerability. While dropout rates initially declined, recent data indicate a rebound to 1.9% in 2022 to 2.2% in 2024, suggesting renewed instability despite sustained policy efforts.25 This instability is compounded by the structural devaluation of parental human capital; unrecognized North Korean degrees often limit parents to precarious employment, creating a structural ceiling for family mobility.
Furthermore, longitudinal qualitative data reveal a phenomenon of forced agency. NKRY often prioritize stable vocational paths, such as nursing or social work, over personal aptitude to mitigate family financial strain.24 While influenced by South Korean cultural norms, this reflects a collective survival strategy shaped by the legacy of displacement.24,26 Other displaced populations may exhibit comparable mechanisms. Establishing global generalizability requires accumulating granular evidence like the present study. These compressed transitions to adulthood heighten psychological strain by sacrificing developmental exploration.27 Critically, standard SDH metrics capture educational outcomes but overlook this underlying process of developmental compromise.
Practitioners should address this by screening for parental adverse childhood experiences. They must also identify specific factors within their unique contexts that hinder a youth’s developmental exploration.28 Furthermore, schools and clinics should implement cultural safety frameworks. This means recognizing survival-driven choices as legitimate adaptations rather than developmental failures. Institutional support should pursue 2 main goals. First, professional counseling must help youths secure exploratory opportunities. Second, support systems should help them process their circumstances to thrive despite intergenerational pressures. Ultimately, validating diverse life trajectories beyond conventional upward mobility models can effectively mitigate the distinct stress of being part of a minoritized group.
The findings of this study have several implications for research on minoritized youth health. The large unexplained component indicates that key determinants shaping vulnerability remain insufficiently captured by prevailing SDH models. Consequently, further expansion using conventional socioeconomic or school-related variables alone is unlikely to substantially reduce unexplained disparities.
These results highlight the need for better measurement strategies. Historically focused on physical health, instruments like the KYRBS must evolve to include comprehensive mental health assessments. National surveys face practical limits in adding minoritized group-specific measures like acculturative stress. Therefore, developing dedicated minoritized group surveys is essential. Meanwhile, researchers must pragmatically use existing data to identify vulnerable youths. Moreover, while our model emphasizes structural factors, other unmeasured social determinants, such as neighborhood or peer contexts, could also play a substantial role. Addressing this requires complementary approaches. Longitudinal designs are necessary to capture intergenerational processes. Qualitative or mixed-methods research is also crucial. For example, interviewers with shared cultural backgrounds enable context-sensitive measurement.29 Without these advances, disparities will be repeatedly documented but rarely understood.
Furthermore, this study highlights the importance of analytic approaches that explicitly interrogate explanatory limits. Decomposition methods are critical for distinguishing between disparities attributable to observed contexts and those that persist beyond them. In their absence, research risks overemphasizing observable economic disadvantages while obscuring structurally patterned processes.30,31
From a policy perspective, interventions focusing narrowly on improving family socioeconomic resources or school performance are unlikely to substantially reduce mental health disparities. While economic support remains important, policies must pivot toward early psychosocial support that addresses the root causes of vulnerability. Specifically, interventions should prioritize family-based programs designed to interrupt the intergenerational transmission of trauma,32 rather than focusing solely on the individual child. In schools, implementing cultural safety frameworks alongside standard curriculum can help mitigate the distinct minoritized stress navigated by these youths.33
Taken together, these findings hold universal implications. As societies become more diverse, reliance on conventional indicators risks rendering invisible the distinct needs of populations shaped by displacement. Recognizing this limitation is essential for moving beyond narrow, short-term solutions toward policies that address the deeper sources of risk.
Limitations
Several limitations should be acknowledged. First, the cross-sectional design precludes causal inference. Although modeled as parallel outcomes, violence and health behaviors likely interact bidirectionally as both outcomes and social determinants of mental health. However, the use of nationally representative data spanning 15 years allows for robust assessment of population-level patterns and consistency across cohorts. Second, socioeconomic conditions were measured using perceived family economic status rather than objective income. While this introduces uncertainty, it is unlikely to fully account for our findings. Importantly, international-origin youths share similar vulnerabilities in socioeconomic reporting yet exhibited substantially smaller unexplained components. This divergence suggests that measurement bias alone cannot explain the large residual disparities observed in refugee youths. Third, the survey lacked direct measures of psychosocial and historical exposures. Nevertheless, this dataset remains the only nationally representative source including youths of North Korean refugee origin. It offers a unique opportunity to empirically evaluate the limits of conventional social determinants frameworks. While this study used a broad IOY category, future qualitative research should compare NKRY with specific subgroups like youths of Chinese or Vietnamese descent. Comparing voluntary migration groups with those shaped by displacement will provide vital clues for developing targeted interventions. Finally, although the small sample size of refugee-origin adolescents limited statistical power, the consistency of decomposition results across multiple outcomes and survey waves strengthens confidence in the robustness of our conclusions. Furthermore, variables introduced between 2018 and 2024 rely on subsamples smaller than 200. These include loneliness, anxiety, heated tobacco, and drug use. These findings reflect a recent cohort and require cautious interpretation. Nevertheless, they still highlight extreme disparities.
Conclusions
This cross-sectional study found that substantial mental health and behavioral disparities among minoritized youths can persisted beyond what conventional demographic, socioeconomic, family, and academic determinants were able to explain. Using youths of NKRY as a theoretically informative case, our findings demonstrate that these unexplained gaps reflected the limits of prevailing SDH frameworks, pointing to the need for structural interventions that better capture cumulative, intergenerational, and life-course processes shaping minoritized youth health.
eTable 1. Detailed Family Structure of Youths by Parental Origin
eTable 2. Survey Items Used to Assess Socioeconomic Status, Mental Health, and Health Behaviors
eFigure. Trends in Mental Health and Health Behavior Among Youth by Parental Origin, 2011-2024
eTable 3. Crude Associations Between Parental Origin and Mental Health and Health Behavior Outcomes Among Youth: Survey-Weighted Logistic Regression, 2011-2024
eTable 4. Adjusted Associations Between Parental Origin and Mental Health and Health Behavior Outcomes Among Youth: Survey-Weighted Logistic Regression Analysis, 2011-2024
eTable 5. Fairlie Decomposition of Disparities in Mental Health and Health Behaviors Between North Korean Refugee Origin and South Korean Youth
eTable 6. Fairlie Decomposition of Disparities in Mental Health and Health Behaviors Between International Origin Youth and South Korean Youth
Data Sharing Statement
References
- 1.GBD 2019 Mental Disorders Collaborators . Global, regional, and national burden of 12 mental disorders in 204 countries and territories, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet Psychiatry. 2022;9(2):137-150. doi: 10.1016/S2215-0366(21)00395-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Cosma A, Stevens G, Martin G, et al. Cross-national time trends in adolescent mental well-being from 2002 to 2018 and the explanatory role of schoolwork pressure. J Adolesc Health. 2020;66(6S)(suppl):S50-S58. doi: 10.1016/j.jadohealth.2020.02.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Reitsma MB, Kendrick PJ, Ababneh E, et al. ; GBD 2019 Tobacco Collaborators . Spatial, temporal, and demographic patterns in prevalence of smoking tobacco use and attributable disease burden in 204 countries and territories, 1990-2019: a systematic analysis from the Global Burden of Disease Study 2019. Lancet. 2021;397(10292):2337-2360. doi: 10.1016/S0140-6736(21)01169-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Viner RM, Ozer EM, Denny S, et al. Adolescence and the social determinants of health. Lancet. 2012;379(9826):1641-1652. doi: 10.1016/S0140-6736(12)60149-4 [DOI] [PubMed] [Google Scholar]
- 5.Reiss F. Socioeconomic inequalities and mental health problems in children and adolescents: a systematic review. Soc Sci Med. 2013;90:24-31. doi: 10.1016/j.socscimed.2013.04.026 [DOI] [PubMed] [Google Scholar]
- 6.Williams DR, Lawrence JA, Davis BA. Racism and health: evidence and needed research. Annu Rev Public Health. 2019;40:105-125. doi: 10.1146/annurev-publhealth-040218-043750 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Braveman P, Egerter S, Williams DR. The social determinants of health: coming of age. Annu Rev Public Health. 2011;32:381-398. doi: 10.1146/annurev-publhealth-031210-101218 [DOI] [PubMed] [Google Scholar]
- 8.Holdroyd I, Vodden A, Srinivasan A, Kuhn I, Bambra C, Ford JA. Systematic review of the effectiveness of the health inequalities strategy in England between 1999 and 2010. BMJ Open. 2022;12(9):e063137. doi: 10.1136/bmjopen-2022-063137 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Khazanchi R, Evans CT, Marcelin JR. Racism, not race, drives inequity across the COVID-19 continuum. JAMA Netw Open. 2020;3(9):e2019933-e2019933. doi: 10.1001/jamanetworkopen.2020.19933 [DOI] [PubMed] [Google Scholar]
- 10.Lee R, Lee SM, Hong M, Oh IH. Psychiatric health risks in North Korean refugee youths resettled in South Korea. JAMA Netw Open. 2025;8(5):e2512941-e2512941. doi: 10.1001/jamanetworkopen.2025.12941 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Dawes DE. The political determinants of health. JHU Press; 2020. doi: 10.56021/9781421437903 [DOI] [Google Scholar]
- 12.Gkiouleka A, Huijts T, Beckfield J, Bambra C. Understanding the micro and macro politics of health: Inequalities, intersectionality & institutions - a research agenda. Soc Sci Med. 2018;200:92-98. doi: 10.1016/j.socscimed.2018.01.025 [DOI] [PubMed] [Google Scholar]
- 13.Castañeda H, Holmes SM, Madrigal DS, Young ME, Beyeler N, Quesada J. Immigration as a social determinant of health. Annu Rev Public Health. 2015;36:375-392. doi: 10.1146/annurev-publhealth-032013-182419 [DOI] [PubMed] [Google Scholar]
- 14.Kim Y, Choi S, Chun C, Park S, Khang YH, Oh K. Data resource profile: the Korea youth risk behavior web-based survey (KYRBS). Int J Epidemiol. 2016;45(4):1076-1076e. doi: 10.1093/ije/dyw070 [DOI] [PubMed] [Google Scholar]
- 15.Spitzer RL, Kroenke K, Williams JB, Löwe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med. 2006;166(10):1092-1097. doi: 10.1001/archinte.166.10.1092 [DOI] [PubMed] [Google Scholar]
- 16.Currie C, Molcho M, Boyce W, Holstein B, Torsheim T, Richter M. Researching health inequalities in adolescents: the development of the health behaviour in school-aged children (HBSC) family affluence scale. Soc Sci Med. 2008;66(6):1429-1436. doi: 10.1016/j.socscimed.2007.11.024 [DOI] [PubMed] [Google Scholar]
- 17.Fairlie RW. An extension of the Blinder-Oaxaca decomposition technique to logit and probit models. J Econ Soc Meas. 2005;30(4):305-316. doi: 10.3233/JEM-2005-0259 [DOI] [Google Scholar]
- 18.Kickbusch I. The political determinants of health—10 years on. British Medical Journal Publishing Group; 2015. doi: 10.1136/bmj.h81 [DOI] [PubMed] [Google Scholar]
- 19.Karatekin C, Gresham B, Barnes AJ, Corcoran F, Kritzik R, Mason SM. Re-politicizing the WHO’s social determinants of health framework. Health Promot Int. 2024;39(5):daae122. doi: 10.1093/heapro/daae122 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Sangalang CC, Vang C. Intergenerational trauma in refugee families: a systematic review. J Immigr Minor Health. 2017;19(3):745-754. doi: 10.1007/s10903-016-0499-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Kwon H. After the Korean War: An Intimate History. Cambridge University Press; 2020. doi: 10.1017/9781108768313 [DOI] [Google Scholar]
- 22.Lankov A. Bitter taste of paradise: North Korean refugees in South Korea. Journal of East Asian Studies. 2006;6(1):105-137. doi: 10.1017/S1598240800000059 [DOI] [Google Scholar]
- 23.Choudhury S, Piera Pi-Sunyer B, Blakemore SJ. A neuroecosocial perspective on adolescent development. Annu Rev Dev Psychol. 2023;5(1):285-307. doi: 10.1146/annurev-devpsych-120321-011511 [DOI] [Google Scholar]
- 24.Nam SD, Kim JS, Kim S, et al. Second-Cycle of the Education Longitudinal Study on Youth with the North Korean Background (VII) . Han’guk Kyoyuk Kaebarwŏn; 2022. [Google Scholar]
- 25.North Korean Youth Education Support Center . Education statistics for students with North Korean background in 2025. 2025. Accessed June 10, 2026. https://hub4u.or.kr/board/policy/view/1142.do?searchword=&searchkey=articleSubject&page=1&order=1
- 26.Titzmann PF. Growing up too soon? Parentification among immigrant and native adolescents in Germany. J Youth Adolesc. 2012;41(7):880-893. doi: 10.1007/s10964-011-9711-1 [DOI] [PubMed] [Google Scholar]
- 27.Kumaga R. The role of intergenerational trauma and adultification in children and young people. Reimagining Race in Psychology. Routledge; 2024:131-145. [Google Scholar]
- 28.Ramsden I. Cultural safety and nursing education in Aotearoa and Te Waipounamu. Victoria University of Wellington Wellington; 2002. [Google Scholar]
- 29.Willis GB. The practice of cross-cultural cognitive interviewing. Public Opin Q. 2015;79(S1):359-395. doi: 10.1093/poq/nfu092 [DOI] [Google Scholar]
- 30.Koh K, Elder TE, Grady SC, Darden JT, Vojnovic I. Explained and unexplained racial and regional inequality in obesity prevalence in the United States. Ethn Health. 2020;25(5):665-678. doi: 10.1080/13557858.2018.1442559 [DOI] [PubMed] [Google Scholar]
- 31.Jackson JW, VanderWeele TJ. Decomposition analysis to identify intervention targets for reducing disparities. Epidemiology. 2018;29(6):825-835. doi: 10.1097/EDE.0000000000000901 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Mental health of refugees and migrants: risk and protective factors and access to care. World Health Organization; 2023. [PubMed] [Google Scholar]
- 33.Teja Z, Worrell FC. School-based interventions for refugee children and adolescents. In: Worrell FC, Hughes TL, Dixson DD, eds. The Cambridge Handbook of Applied School Psychology. Cambridge University Press; 2020:634-654. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eTable 1. Detailed Family Structure of Youths by Parental Origin
eTable 2. Survey Items Used to Assess Socioeconomic Status, Mental Health, and Health Behaviors
eFigure. Trends in Mental Health and Health Behavior Among Youth by Parental Origin, 2011-2024
eTable 3. Crude Associations Between Parental Origin and Mental Health and Health Behavior Outcomes Among Youth: Survey-Weighted Logistic Regression, 2011-2024
eTable 4. Adjusted Associations Between Parental Origin and Mental Health and Health Behavior Outcomes Among Youth: Survey-Weighted Logistic Regression Analysis, 2011-2024
eTable 5. Fairlie Decomposition of Disparities in Mental Health and Health Behaviors Between North Korean Refugee Origin and South Korean Youth
eTable 6. Fairlie Decomposition of Disparities in Mental Health and Health Behaviors Between International Origin Youth and South Korean Youth
Data Sharing Statement
