Delinquency Type As a Distributional Amplifier of Somatic Burden In Maltreated Adolescents: A Latent Class and Quantile Regression Approach

Background and Purpose: Adolescent delinquency and somatization co-occur at clinically significant rates, yet their joint patterning within child maltreatment exposure remains underexplored. Most prior studies have treated delinquency as a unitary construct and relied on mean-based analyses, obscuring how specific delinquency types interact with heterogeneous maltreatment profiles to produce differential somatic burden. Grounded in a bio-psychosocial framework, this study aimed to (1) identify latent child maltreatment profiles in a nationally representative Korean adolescent sample, and (2) examine how delinquency type conditions the association between maltreatment typologies and somatization across the somatic symptom distribution.

Methods: Data were drawn from the 2021 National Survey on Mental Health of Youth (N = 6,689), collected via stratified cluster sampling. Child maltreatment was assessed using eight indicators of physical abuse and neglect. Latent class analysis identified four maltreatment profiles. Three delinquency variables — status offenses, violent delinquency, and property offenses — were operationalized as dichotomous indicators. Joint maltreatment-delinquency exposure groups were entered as categorical predictors in simultaneous quantile regression models at the 30th through 90th percentiles, with bootstrapped standard errors (1,000 replications), adjusting for sex, age, family structure, socioeconomic status, perceived stress, self-esteem, and school adjustment.

Results: LCA identified four maltreatment profiles: Least Victimization (56.4%), Moderate Dual Victimization (31.1%), Predominantly Neglected (6.7%), and Predominantly Abused (5.8%). Quantile regression revealed pronounced distributional heterogeneity. Among adolescents in the Least Victimization class, status and violent delinquency were significantly associated with elevated somatization across all quantiles (status: β = 0.67–2.80; violent: β = 0.93–3.38), suggesting that delinquency itself carries somatic correlates independent of maltreatment severity. Most strikingly, in the Predominantly Neglected class, violent delinquency exhibited markedly amplified associations at the upper tail of the distribution (τ = .80: β = 11.63, 95% CI [4.06, 19.20]; τ = .90: β = 11.13, 95% CI [2.39, 19.87]) — a pattern that persisted after covariate adjustment and was not replicated in other maltreatment classes, indicating a synergistic rather than additive concentration of risk.

Conclusions and Implications: The dramatically amplified somatic burden observed among neglected adolescents engaging in violent delinquency at the upper symptom quantiles points toward a mechanism in which relational deprivation — the defining feature of neglect — impairs emotional regulation and socialization, channeling unresolved attachment needs into violent externalization while generating chronic physiological stress that becomes somatically embedded. This convergence of behavioral and somatic distress is consistent with alexithymia-linked pathways and autonomic dysregulation, and its concentration at the upper symptom distribution means it is precisely the population least likely to be detected by conventional mean-based screening. For practitioners operating within multi-tiered systems of support, these findings reframe somatic complaints in adolescents with neglect histories and violent behavior not as separate clinical concerns but as co-constitutive signals of the same underlying adversity burden — warranting integrated, trauma-informed assessment rather than siloed referral. Embedding somatization inquiry within behavioral risk screening, and calibrating intervention thresholds to distributional vulnerability rather than average risk, may be among the more tractable levers available to practitioners seeking to reach the most burdened youth before symptoms become entrenched.