Introduction
Adolescence represents a pivotal period in psychological development, characterized by significant shifts in identity formation, emotional regulation, and interpersonal functioning[1]. These transitions, while normative, often blur the boundary between typical developmental variability and emerging psychopathology, complicating diagnostic clarity for mental health professionals[2,3]. Among the most debated diagnostic categories in adolescent psychiatry are personality disorders (PDs). Historically, clinicians have hesitated to apply these diagnoses to youth, largely due to concerns about stigmatization and the belief that personality is not yet sufficiently stable[4]. Nonetheless, an increasing body of empirical evidence indicates that personality dysfunctions emerge early and are associated with significant long-term impairments in emotional, social, and academic domains[5,6,7]. Contrary to earlier assumptions, research indicates that PDs exhibit greater stability in adolescence than previously thought and tend to become less stable in adulthood, challenging traditional views on their chronicity[6]. Neglecting the impact of PDs in adolescence may contribute to maintaining rather than reducing stigma, as early identification and intervention could prevent further deterioration in functioning. Longitudinal studies have demonstrated that adolescent PDs are associated with increased risk for both Axis I and Axis II disorders in adulthood[7,8]. Specifically, PDs in adolescence show significant comorbidity with substance use and conduct disorders, as well as impairments in self and interpersonal functioning, making therapeutic interventions particularly complex; additionally, pathological personality patterns in adolescence are linked to romantic relationship conflicts, peer difficulties, educational and vocational challenges, family discord, and low self-esteem[6,9,10]. From a public health perspective, adolescent PDs represent a substantial burden, as they are associated with elevated mental health service use, high psychological distress[11], and significant challenges in recovery over time. Consequently, a thorough evaluation of personality is crucial for the development of early intervention strategies aimed at mitigating long-term impairment and fostering resilience[12].
Traditional diagnostic frameworks, such as the International Classification of Diseases (ICD) [13] and DSM-5 Section II [14], rely on categorical classification, where disorders are defined as distinct entities based on symptom clusters. This model, while systematic, has been criticized for its limited sensitivity to developmental variation and for oversimplifying the heterogeneity of personality pathology during adolescence [15]. In response, the DSM-5 Section III introduced the Alternative Model for Personality Disorders (AMPD) [16], which adopts a dimensional approach. It evaluates impairments in personality functioning (Criterion A) and maladaptive personality traits (Criterion B). This model is grounded in contemporary personality theory and supported by research emphasizing that personality traits lie on a continuum rather than fitting neatly into categories [17]. Moreover, the dimensional framework aligns with the developmental psychopathology perspective, which conceptualizes personality development as a dynamic interplay between biological, psychological, and environmental factors over time [17]. From a developmental psychopathology perspective, early difficulties in emotion regulation and stress reactivity, disruptions in identity consolidation (e.g., unstable self-representations and self-direction), and recurrent problems in attachment-related interpersonal functioning can operate as transdiagnostic liabilities across adolescent clinical presentations [18]. Rather than mapping neatly onto single categories, these vulnerabilities may contribute to heterogeneous symptom configurations (internalizing, externalizing, and mixed profiles) and to fluctuating severity across contexts during adolescence. The DSM-5 AMPD was designed to capture these core mechanisms dimensionally: Criterion A operationalizes impairments in self and interpersonal functioning (identity/self-direction; empathy/intimacy), while Criterion B characterizes trait-like dysregulation patterns that cut across traditional diagnoses [19]. In this framework, dimensional assessment may help identify clinically meaningful dysfunction even when categorical thresholds are not met, offering a potentially developmentally sensitive bridge between normative variability and emerging personality pathology. This is particularly relevant in youth mental health services, where early detection and formulation can guide preventive and stage-sensitive interventions [20]. Moreover, dimensional models may be especially informative in adolescent clinical populations where comorbidity is the rule rather than the exception and symptom configurations are developmentally fluid. Trait-level assessment can therefore complement categorical diagnosis by capturing transdiagnostic liabilities that cut across traditional categories in youth clinical presentations [21]. Key developmental principles such as equifinality (diverse pathways leading to similar outcomes) and multifinality (shared risk factors leading to divergent outcomes) illustrate the complex etiology of personality disorders [17,22]. These concepts are particularly relevant during adolescence, when the onset of personality dysfunction may reflect a convergence of genetic vulnerability, attachment insecurity, and environmental stressors [22].
Despite these theoretical advances, the clinical community still lacks consensus on which diagnostic approach is most informative during adolescence - particularly when it comes to treatment planning and prognostic accuracy [12,18]. Furthermore, studies examining concordance between dimensional and categorical systems, or exploring gender differences in trait presentation and diagnosis, remain limited [9, 22].
In light of these limitations, we designed a study to directly compare these models in a real-world clinical setting. Thus, this study aims to compare categorical and dimensional assessment models in a clinical sample of adolescents aged 14 to 18. Using a combination of standardized diagnostic tools, we seek to:
- Assess the prevalence and distribution of personality disorder diagnoses across models.
- Explore gender differences in maladaptive traits and diagnostic outcomes.
- Examine the degree of diagnostic convergence across systems.
- Identify which model provides a more comprehensive and clinically useful portrait of adolescent personality dysfunction.
By integrating multiple diagnostic frameworks, this study contributes to a more nuanced understanding of personality pathology in adolescence, with implications for early detection, gender-sensitive assessment, and targeted intervention.
Methods
Procedure
A retrospective analysis was performed on clinical data collected between December 2019 and December 2023 at a specialized community mental health service in Italy ([masked for review]), focused on early intervention for adolescent mental health disorders. All consecutive referrals in the age range 14-18 were screened for eligibility. Inclusion criteria were: (a) availability of SCID-5-AMPD and SCID-5-PD assessments conducted as part of routine clinical evaluation; and (b) completion of PID-5 and SCL-90-R during assessment. Exclusion criteria were: (a) diagnosis of intellectual disability; (b) autism spectrum disorder; (c) major neurological conditions; and (d) early psychotic onset. For SCID-5-PD, the retrospective charts systematically reported only the primary categorical PD diagnosis per participant: therefore, categorical PD comorbidity could not be examined in the present dataset. This process resulted in a final analytic sample of N = 38. ICD-9 diagnoses were assigned by the attending psychiatrist of the service based on clinical evaluation supported by diagnostic manual [23]. ICD-9 was used because ICD is the administrative diagnostic standard within the Italian National Health System, ensuring continuity with historical diagnostic documentation and comparability across different services on the national territory and version 9 was employed at the time of the study and until now at local level (Lazio region). Personality disorder assessments were conducted using standardized structured interviews (SCID-5-PD and SCID-5-AMPD), administered by trained clinicians with formal preparation in structured diagnostic interviewing. This ensured that all diagnostic procedures were performed by appropriately qualified professionals. Assessors followed uniform training and supervision procedures to ensure consistency in the application of diagnostic criteria. Personality disorder diagnoses were assessed according to both Criterion A and Criterion B of the DSM-5 Alternative Model for Personality Disorders (AMPD), considering a diagnosis valid only when both criteria were simultaneously met. Although the DSM-5 dimensional model for personality disorders is a recently developed and alternative model, the clinical service routinely adopts it alongside the categorical DSM-5 system because of its developmental sensitivity and its suitability for assessing emerging personality dysfunction in adolescents. Criterion A evaluates impairments in personality functioning across two domains: Self (identity and self-direction) and Interpersonal (empathy and intimacy). Criterion B involves the presence of pathological personality traits assessed via the Personality Inventory for DSM-5 (PID-5).
Participants
A total of 146 adolescents were admitted for routine clinical assessment. Of these, 59 met the eligibility criteria (age 14-18 years, absence of psychotic onset, adequate Italian language proficiency) and at least partial AMPD dimensional information was available—i.e., data allowing evaluation of Criterion A and/or Criterion B (SCID-5-AMPD and PID-5). From these 59 cases, only adolescents with sufficient information to evaluate both AMPD dimensional criteria (Criterion A and Criterion B) were retained (n = 44); cases with incomplete or missing information for the joint Criterion A+B evaluation were excluded. Among the 44 retained adolescents, six could not be contacted to complete the SCID-5-PD interview. Thus, a final analytic sample consisted of 38 subjects aged between 14 and 18 years were recruited (mean age = 16.1; SD = 1.19), comprising 25 females (65.8%) and 13 males (34.2%). Socioeconomic distribution was as follows: 42.1% (n = 16) from middle-class backgrounds, 28.9% (n = 11) from lower-class, and 28.9% (n = 11) from upper-class families. For age-related comparisons, participants were divided into two groups: 14-15 years and 16-18 years. No significant differences in test scores were found between these age groups. Written informed consent was obtained from parents/legal guardians, and adolescents provided written assent in accordance with local ethical requirements.
Materials
The assessment of personality was conducted utilizing a comprehensive battery of instruments tailored to capture the multifaceted nature of personality disorders. Specifically, we employed the following instruments:
The International Classification of Diseases, 9th Revision (ICD-9)
The International Classification of Diseases (ICD) is a standardized system developed by the World Health Organization (WHO) for classifying and coding diseases and other health-related issues. Its primary purpose is to provide a uniform basis for collecting, analyzing, and communicating health data globally. The ICD is used to classify a wide range of medical conditions, including physical illnesses, mental disorders, injuries, external causes of illness, and other health problems. Its hierarchical structure organizes conditions using alphanumeric codes to identify specific categories of diseases and disorders. This classification system is widely utilized in healthcare, clinical research, epidemiology, and health resource management worldwide, including in Italy [23].The International Classification of Diseases, Ninth Revision (ICD-9), was used because it complies with the Italian National regulations, which formally requires ICD diagnostic codes on clinical records [24; 25].
Symptom Checklist-90 (SCL-90- R)
The SCL-90- R is a widely used self-report questionnaire designed to assess a broad range of psychological symptoms and distress. It comprises 90 items measuring nine primary symptom dimensions, including somatization, obsessive-compulsive symptoms, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety, paranoid ideation, and psychoticism. The SCL-90- R aids in the assessment of general psychopathology and provides valuable insights into symptom severity and psychological distress across various clinical populations. The Italian version demonstrates strong reliability and validity [26].
SCID-5-AMPD (Structured Clinical Interview for DSM-5 Alternative Model for Personality Disorders) - Module I
The SCID-5-AMPD is a structured diagnostic interview developed to assess the alternative model for personality disorders proposed in DSM-5 Section III. It evaluates impairments in personality functioning and the presence of pathological personality traits as outlined in the DSM-5 AMPD. The SCID-5-AMPD shows good inter-rater reliability and convergent validity with other measures of personality pathology [27]. It effectively captures dimensional traits specified in the DSM-5 AMPD, enhancing diagnostic accuracy and treatment planning. In the present study, inter-rater agreement was examined on a subset of SCID-5-AMPD interviews that were independently double-rated by a second trained clinician (n = 10). Cohen’s kappa coefficients indicated a high level of interrater agreement across SCID domains, with k values ranging from 0.72 to 1.00. All coefficients were statistically significant.
PID-5 (Personality Inventory for DSM-5) adolescent version
The PID-5 is a widely used self-report inventory designed to assess the dimensional traits specified in the DSM-5 Section III Personality Disorders. All participants completed the PID-5 Adolescent Form, which is developmentally adapted for youth and validated for assessing personality traits in adolescents. It comprises 220 items, covering five broad trait domains: Negative Affectivity, Detachment, Antagonism, Disinhibition, and Psychoticism. Each domain is further subdivided into 25 specific facets. The PID-5 demonstrates robust psychometric properties, including high internal consistency, test-retest reliability, and convergent validity with other measures of personality pathology. Factor analyses support the five-factor structure, and the inventory shows sensitivity to clinically relevant personality traits and their severity levels. The Italian validation of this instrument was conducted by Fossati and colleagues, ensuring its applicability and reliability within Italian-speaking populations[28].
SCID 5-PD (Structured Clinical Interview for DSM-5 Personality Disorders)
The SCID 5-PD is a semi-structured diagnostic interview widely utilized for the assessment of personality disorders based on DSM-5 criteria. It provides a comprehensive evaluation of personality pathology through a systematic inquiry into diagnostic criteria, symptom severity, and functional impairment. The SCID 5-PD ensures standardized administration and scoring, enhancing reliability and facilitating accurate diagnosis of personality disorders.The SCID 5-PD exhibits good inter-rater reliability and concurrent validity. It has demonstrated high sensitivity and specificity in diagnosing personality disorders, with structured interview format reducing diagnostic variability [29].
Data Analysis
All statistical analyses were performed using Jamovi [30]. Descriptive statistics were computed to summarize demographic variables, psychiatric symptomatology, and personality traits. To investigate gender differences in symptom severity, a Mann-Whitney U test was conducted on the global severity index (GSI) of the SCL-90-R. Similarly, Mann-Whitney U tests were employed to compare male and female participants on personality trait scores from the PID-5, assessing potential differences in individual maladaptive personality traits. Regarding personality functioning, a Mann-Whitney U test was performed on SCID-5 AMPD scores to examine differences between genders and age groups (14-15 years vs. 16-18 years) in global personality functioning and its subdomains (Identity, Self-Direction, Empathy, and Intimacy). For categorical personality disorder diagnoses, chi-square (χ2) tests were conducted to analyze the distribution of DSM-5 (SCID-5 PD) and Alternative Model (SCID-5 AMPD + PID-5) diagnoses across gender groups. No correction for multiple comparisons was applied because the analyses were exploratory and aimed at identifying preliminary patterns rather than testing confirmatory hypotheses. Given the small sample size, applying strict correction procedures would have substantially increased Type II error, making it more difficult to detect clinically meaningful trends. This approach is consistent with methodological recommendations for early-stage clinical research. For Mann-Whitney U tests, effect sizes were expressed as rank-biserial correlations (r), and 95% confidence intervals for r were computed using Fisher’s z transformation. All analyses adhered to conventional statistical thresholds for significance (p < 0.05). No missing data were present for the variables included in the analyses; therefore, no imputation procedures were applied.
Results
The sample comprised 38 adolescents aged between 14 and 18 years (M = 16.1; SD = 1.19), with a predominance of females (N = 25; 65.8%) over males (N = 13; 34.2%). Socioeconomic status was primarily middle (N = 16; 42.1%), followed by lower (N = 11; 28.9%) and upper class (N = 11; 28.9%). No significant age-related differences emerged when participants were stratified into two groups (14-15 years vs. 16-18 years). Comparisons between age groups on personality traits (PID-5) and personality functioning (SCID-5 AMPD) also did not yield significant results. Traits frequently observed in the sample included depressivity, intimacy avoidance, and anxiousness. Overall psychopathological distress on the SCL-90-R Global Severity Index (GSI) was elevated in the sample (M = 1.89, SD = 0.72). Females showed significantly higher GSI scores than males (M(females) = 2.14, SD = 0.53; M(males) = 1.42, SD = 0.81; U = 76.0, p = 0.008, r = 0.53. All subgroup comparisons were conducted as exploratory analyses in this small clinical sample and should be interpreted cautiously. Regarding categorical diagnoses based on ICD-9 codes, the most frequent were Borderline Personality Disorder (ICD 301.83; n = 4, 10.5%) and Adjustment Disorder with Anxiety and Depressed Mood (ICD 309.28; n = 3, 7.9%). Other diagnoses were less common, with two cases each (5.3%) of Atypical Depressive Disorder (ICD code 296.82), Obsessive-Compulsive Disorder (ICD code 300.3), Non-Depressive NCA Disorders (ICD code 311), Unspecified Anxiety Disorder (ICD code 300.00), Eating Disorders Not Otherwise Specified (ICD code 307.5), Unspecified Personality Disorder (ICD code 301.9), and Adjustment Disorder with Depressed Mood (ICD code 309.0).
The dimensional assessment using the PID-5 identified Negative Affectivity as the most impaired domain (M = 1.87, SD = 0.62), followed by Disinhibition (M = 1.62, SD = 0.57), Psychoticism (M = 1.55, SD = 0.70), Detachment (M = 1.48, SD = 0.47), and Antagonism (M = 0.98, SD = 0.59). At the facet level, the most elevated traits were Emotional Lability (M = 2.04, SD = 0.66), Anxiousness (M = 2.03, SD = 0.63), Anhedonia (M = 1.98, SD = 0.78), Distractibility (M = 1.94, SD = 0.67), Depressivity (M = 1.94, SD = 0.69), Eccentricity (M = 1.86, SD = 0.70), and Suspiciousness (M = 1.83, SD = 0.50). In contrast, Grandiosity (M = 0.67, SD = 0.69) and Insensitivity/Callousness (M = 0.90, SD = 0.68) were among the least elevated facets in the sample.Females showed significantly higher Disinhibition scores than males (M(females) = 1.77, SD = 0.58; M(males) = 1.34, SD = 0.42; U = 95.5, p = 0.041, r = 0.41) and higher Psychoticism (Mfemales = 1.71, SD = 0.72; Mmales = 1.25, SD = 0.57; U = 97.0, p = 0.044, r = 0.40). For Negative Affectivity, females tended to score higher than males (M(females) = 2.02, SD = 0.58; M(males) = 1.59, SD = 0.63), although this difference did not reach conventional significance (U = 101.0, p = 0.061, r = 0.38). For the main gender effects, confidence intervals for the rank-biserial correlations confirmed the robustness of the findings: GSI (r = 0.53, 95% CI [0.26, 0.73]), Disinhibition (r = 0.41, 95% CI [0.11, 0.65]), and Psychoticism (r = 0.40, 95% CI [0.10, 0.64]), all of which did not include zero.
Levels of personality functioning assessed with the SCID-5-AMPD indicated moderate impairment in the sample, with mean scores of 2.15 (SD = 0.67) for the global score, 2.55 (SD = 0.70) for Identity, 2.05 (SD = 0.75) for Self-direction, 1.76 (SD = 0.89) for Empathy, and 2.22 (SD = 0.85) for Intimacy. No significant gender differences emerged on the SCID-5-AMPD global score or individual domains (e.g., (M(females) = 2.14, SD = 0.72; M(males) = 2.16, SD = 0.60) global score: U = 160.5, p = 0.963, r = 0.01). Using the AMPD diagnostic criteria (i.e., meeting both Criterion A: two or more domains ≥ 2, and Criterion B: two or more traits ≥ 2), the most frequent diagnosis was Specific Personality Trait Disorder (n = 24, 63.2%). Among specific diagnoses, Borderline Personality Disorder was the most prevalent (n = 6, 15.8%), followed by Schizotypal Personality Disorder (n = 3, 7.9%), Narcissistic Personality Disorder (n = 3, 7.9%), and Avoidant Personality Disorder (n = 2, 5.3%).
For categorical diagnoses according to the DSM-5 (SCID-5 PD), the most frequent personality disorders were Borderline Personality Disorder (N = 10; 26.3%), Avoidant Personality Disorder (N = 8; 21.1%), and Personality Disorder Not Otherwise Specified (N = 7; 18.4%). Less frequent diagnoses included Paranoid Personality Disorder (N = 3; 7.9%), Schizotypal Personality Disorder (N = 3; 7.9%), Narcissistic Personality Disorder (N = 2; 5.3%), Histrionic Personality Disorder (N = 2; 5.3%), Schizoid Personality Disorder (N = 2; 5.3%), and Antisocial Personality Disorder (N = 1; 2.6%). No significant differences were found between males and females or between different age groups.
Finally, an analysis of diagnostic convergence across different classification systems (ICD-9, SCID-5 PD, and PID-5) revealed partial agreement. Among the nine ICD diagnoses indicative of personality disorders, only two cases showed full concordance across all three assessment tools (N = 1 Borderline Personality Disorder; N = 1 Personality Disorder Not Otherwise Specified). Additionally, two other cases exhibited alignment between ICD-9 and SCID-5 PD (N = 1 Histrionic Personality Disorder; N = 1 Borderline Personality Disorder). The remaining cases displayed discrepancies across classification methods.
Discussion
This study aimed to explore the prevalence and characteristics of personality disorders in adolescents by integrating both categorical and dimensional diagnostic models. This multi-method approach yielded a comprehensive view of adolescent personality pathology, demonstrating that a considerable portion of the sample exhibited clinically relevant maladaptive traits and impairments. Notably, Borderline Personality Disorder (BPD) emerged as the most common specific diagnosis, while Specific Personality Trait Disorder accounted for the majority of cases overall.
Prevalence of Specific Personality Trait Disorder
Specific Personality Trait Disorder was the most frequent diagnosis, affecting 63.2% of the sample. This highlights the importance of considering personality dysfunctions that fall outside traditional syndromic categories. Adolescents frequently exhibit clinically significant traits that may not fully meet criteria for a defined personality disorder but still warrant clinical attention. This finding supports the value of the dimensional framework, such as the DSM-5 AMPD and ICD-11, which emphasizes the severity and pervasiveness of maladaptive traits over rigid diagnostic thresholds [15].
Prevalence of Borderline Personality Disorder
Borderline Personality Disorder was the most prevalent categorical diagnosis (15.8%), consistent with existing literature highlighting its early onset and strong presence in adolescent clinical populations [18]. BPD is characterized by emotional dysregulation, unstable interpersonal relationships, and identity disturbance, and its early detection is crucial for mitigating future functional impairments. Studies have shown that adverse childhood experiences, including abuse, neglect, and insecure attachment, contribute significantly to the development of BPD traits[22]. However, in the present study these trauma- and attachment-related mechanisms were not directly assessed and are therefore referenced here as part of the broader theoretical framework rather than as empirically examined correlates. Our findings align with research advocating for early identification and targeted interventions, which have been shown to reduce long-term morbidity and improve psychosocial outcomes in youth diagnosed with BPD[12].
Comparison of Categorical and Dimensional Assessments
The comparison between categorical and dimensional approaches revealed important differences in diagnostic yield and clinical insight. While categorical models offer structured, rule-based diagnosis, they may fail to capture the complexity and variability of adolescent psychopathology. In contrast, dimensional tools such as the PID-5 and SCID-5 AMPD identified broader patterns of dysfunction, particularly in the domains of Negative Affectivity, Disinhibition, and Psychoticism. The dimensional model appeared informative in capturing trait-level vulnerabilities that may be less visible within categorical thresholds and potentially relevant for early clinical formulation in adolescent services. These findings are consistent with literature supporting dimensional models as more sensitive to developmental changes and better suited to reflect the fluid nature of adolescent personality pathology [18]. By identifying trait-level vulnerabilities, clinicians may intervene earlier and with greater precision, even in the absence of full-blown disorders. More specifically, the sensitivity of the dimensional model in our adolescent sample is supported by the elevation of key PID-5 domains (Negative Affectivity, Disinhibition, and Psychoticism) even in cases where categorical DSM-5 personality disorder diagnoses were not applicable. This pattern indicates that dimensional traits can detect subthreshold or emerging personality dysfunction that may not yet meet full categorical criteria, consistent with developmental literature showing that personality traits are detectable earlier and fluctuate during adolescence. These findings align with current evidence suggesting that dimensional assessments capture early affective and behavioral dysregulation more effectively than categorical systems. Clinically, this supports the use of dimensional models in youth to improve early identification of risk profiles and guide intervention planning in a developmentally informed way. In particular, elevations in Negative Affectivity are consistent with evidence that emotional instability represents a core mechanism in borderline phenotypes and related presentations, with relevance for both formulation and early intervention planning [31].
Gender Differences in Personality Traits
Females showed higher scores than males on Disinhibition and Psychoticism, but these findings should be interpreted cautiously given the small clinical sample and the exploratory nature of the analyses. In adolescent clinical settings, apparent sex differences may reflect referral pathways, socialization-related differences in symptom expression and help-seeking, or measurement sensitivity (including potential non-invariance across gender), rather than stable categorical differences in underlying liability. Accordingly, we frame these results as signals for clinical attention during assessment, not as evidence of distinct diagnostic types. Future studies using larger samples and explicit tests of measurement invariance and referral bias are needed before drawing stronger conclusions about gender-related trait profiles in adolescent personality pathology.
Age-Related Stability of Personality Dysfunctions
No significant differences were observed between younger (14-15) and older (16-18) adolescents in personality trait severity or functioning levels. Although this null age-group finding cannot be interpreted as evidence of developmental stability, it is broadly consistent with longitudinal work suggesting that clinically relevant personality pathology may be detectable by mid-adolescence and can show moderate continuity over time [12]. Longitudinal studies, such as the CIC study, report that the peak onset for personality disorder symptoms occurs around age 14 and that stability coefficients for PD symptoms range between 0.42 and 0.65, indicating moderate-to-high continuity over time. Accordingly, our results support the clinical utility of early assessment and monitoring during adolescence, while underscoring the need for longitudinal designs to test trajectories, persistence, and prognostic implications.
Convergence Between Diagnostic Systems
Only partial agreement was observed between the ICD-9, SCID-5 PD, and dimensional models. Few cases showed full diagnostic overlap, highlighting the discordance that can arise from using different classification systems. This discrepancy is well documented in the literature and underscores the limitations of relying on a single model for complex, developmentally fluid disorders such as those seen in adolescence. Integrating multiple diagnostic perspectives can provide a more complete picture of individual functioning and needs, facilitating more personalized and developmentally appropriate care [18].
Clinical implications
Treatment implications of trait-domain elevations.
The pattern observed in this sample, with elevations in Negative Affectivity, Disinhibition, and Psychoticism, has direct implications for early treatment planning. Elevated Negative Affectivity points to a core vulnerability in affect regulation and stress reactivity; in early intervention pathways, this supports prioritizing structured skills-based approaches targeting emotion regulation, distress tolerance, and crisis management, as typically emphasized in Dialectical Behavior Therapy for Adolescents (DBT-A). Elevated Disinhibition suggests risk for behavioral dyscontrol and rapid escalation under interpersonal stress: clinically, this supports interventions that combine behavioral analysis, safety planning, with explicit involvement of caregivers when appropriate. Finally, elevated Psychoticism in adolescence may reflect transient psychotic-like experiences, dissociative or trauma-related phenomena, and heightened interpersonal mistrust: rather than implying categorical psychosis, this pattern supports careful differential assessment and a focus on grounding, reality-testing, and improving reflective functioning. In this respect, Mentalization-Based Treatment for Adolescents (MBT-A) targets instability in self-experience and interpersonal meaning-making, and may improve engagement and epistemic trust in youths with fluctuating affect and interpersonal sensitivity [32].
Selecting levels of care.
From a clinical standpoint, the level of personality functioning scores may assist clinicians in determining the appropriate level of care. Adolescents showing higher impairment in self and interpersonal functioning can be prioritized for more intensive or structured interventions (e.g., day programs, individual psychotherapy with family involvement), while those with milder impairment may be adequately followed in outpatient settings. In parallel, specific PID-5 trait configurations offer concrete treatment targets. Elevated Negative Affectivity may indicate a need for emotion-regulation interventions or referral to services specializing in affective dysregulation, whereas elevations in Disinhibition may signal risk-taking behaviors that warrant behavioral monitoring, motivational interventions, or parent-focused work. The PD-TS (Personality Disorder - Trait Specified) framework also facilitates clearer case formulation by allowing clinicians and families to understand symptoms in terms of trait vulnerabilities rather than categorical labels, thereby improving psychoeducation and engagement in treatment. Finally, the dimensional assessment of both Criterion A and PID-5 traits can guide longitudinal monitoring, allowing services to track change over time, adjust intervention intensity, and identify adolescents at elevated risk for later personality psychopathology.
Limitations
This study presents several limitations that should be considered when interpreting the findings. First, the relatively small sample size, drawn from a single clinical setting, may limit the generalizability of the results to broader adolescent populations, particularly those outside clinical care or from different cultural backgrounds. Future studies should include larger, more diverse samples to enhance external validity. In addition, trauma exposure and attachment-related constructs were not directly assessed in this dataset. Therefore, any references to these mechanisms are intended as theoretical context and should not be interpreted as empirically tested correlates of the observed trait and diagnostic profiles. Second, the retrospective cross-sectional design captures data at a single time point, precluding any conclusions about developmental trajectories or causal relationships between variables. Longitudinal research would allow for the examination of how personality traits and disorders evolve over time and whether specific traits predict persistence, remission, or worsening of symptoms. Third, because the retrospective database captured only a single primary SCID-5-PD diagnosis per participant, we did not assess comorbidity across categorical personality disorder diagnoses; future studies should systematically code multiple concurrent PD diagnoses to quantify comorbidity patterns in adolescent clinical samples. In addition, the reliance on self-report measures (e.g., SCL-90-R, PID-5) introduces potential response biases, including social desirability and symptom exaggeration. Adolescents may underreport or overreport symptoms, consciously or unconsciously, which can affect diagnostic accuracy. In particular, the possibility of symptom magnification - where participants may report greater distress to access support - could lead to inflated estimates of psychopathology. Future studies should incorporate multi-informant approaches (e.g., parent or clinician reports) and structured clinical interviews to increase reliability. Moreover, gender differences were explored, but the study did not deeply examine potential biases in diagnosis or treatment across genders. Gender-related diagnostic disparities, shaped by socialization or clinician perception, could have influenced the observed prevalence patterns. Addressing these factors in future research may improve the equity and accuracy of adolescent mental health assessment. By acknowledging these limitations, we underscore the importance of continued research using multi-method, multi-informant, and longitudinal designs to refine the diagnosis and treatment of personality disorders in adolescence. Such work will contribute to a more accurate and developmentally informed understanding of personality pathology during this critical period.
Conclusions
Understanding personality disorders during adolescence is crucial for shaping effective early interventions and mitigating long-term psychological and social impairment. This study provides an integrated view of adolescent personality pathology by comparing categorical (e.g., SCID-5 PD, ICD-9) and dimensional (e.g., AMPD, PID-5) diagnostic approaches within a clinical population. Our findings reveal a high prevalence of Specific Personality Trait Disorder and Borderline Personality Disorder, underscoring the significance of trait-level impairments even when categorical criteria are not met. The dimensional model - particularly the assessment of domains such as Negative Affectivity, Disinhibition, and Psychoticism - appeared useful in capturing nuanced dysfunctions that may be overlooked in traditional categorical frameworks. Additionally, the observation of gender-specific profiles, with females exhibiting greater emotional dysregulation and perceptual alterations, suggests the potential relevance of tailoring assessments and interventions to developmental and gender-based factors.
Importantly, the partial diagnostic overlap between systems reinforces the idea that no single approach fully captures the complexity of adolescent personality pathology. A combined use of categorical and dimensional tools appears to offer the most comprehensive diagnostic insight, supporting calls for hybrid models in clinical practice. While the study’s cross-sectional nature and limited sample size constrain the generalizability of results, its implications are important in the debate on the assessment of personality pathology in adolescence: early identification through a multi-method assessment can contribute to a more precise evaluation of adolescent’s difficulties and inform treatment planning, preventing the progression of maladaptive trajectories into adulthood. Future research should expand on these findings by employing longitudinal designs, larger samples, and incorporating multi-informant perspectives to further validate assessment strategies. Ultimately, refining how we conceptualize and diagnose personality disorders in adolescence is not merely a technical task - it is a clinical priority to better support youth at risk.
Ethical consideration
The study was approved by the local IRB (prot. 12/2021) and carried out in line with the Declaration of Helsinki and the Ethical guidelines for psychological research of the Italian Psychological Association (AIP). Potential participants and their legal guardians were approached by the researcher within 1 month from their first contact with the Service, informed about the study aims and procedures, the right to withdraw from the study at any time and the lack of compensation. Written informed consent was obtained from study participants and their legal guardians.
Acknowledgements
The authors want to thank all participants who contributed to the project.
Funding
No funding was received for this project.
Conflict of interest statement
None to declare.
Authors’ contributions
L.S., A.F., V.C. substantially contributed to the conception of the work, and G.C., L.T., M.M., R.C., contributes to acquisition of data. P.B., C.S., I.P., N.L.C., I.I.G. investigations and resources. V.C., G.D.C. and G.D. supervised the project. L.T. and R.C. drafted the manuscript, which was then revised critically and approved for publication by all authors.
Data availability
The datasets generated and/or analyzed during the current study are not publicly available due to clinical confidentiality constraints but are available from the corresponding author upon reasonable request.
AI-use statement
AI-assisted tools were used solely for language editing and clarity and did not contribute to study design, data analysis, or interpretation.
Figures and tables
| Variable | Category | Frequency (n) | Percentage (%) |
|---|---|---|---|
| Gender | Female | 25 | 65.8 |
| Male | 13 | 34.2 | |
| Age (years) | M (SD) | 16.1 (1.19) | |
| Social Class | Middle Class | 16 | 42.1 |
| Diagnosis | Female (n, %) | Male (n, %) |
|---|---|---|
| Borderline Personality Disorder | 5 (20.0%) | 1 (7.7%) |
| Trait-Specific Personality Disorder | 13 (52.0%) | 11 (84.6%) |
| Avoidant Personality Disorder | 2 (8.0%) | 0 (0.0%) |
| Schizotypal Personality Disorder | 3 (12.0%) | 0 (0.0%) |
| Narcissistic Personality Disorder | 2 (8.0%) | 1 (7.7%) |
| Diagnosis | Female (n, %) | Male (n, %) |
|---|---|---|
| Borderline Personality Disorder | 7 (28.0%) | 3 (23.1%) |
| Histrionic Personality Disorder | 2 (8.0%) | 0 (0.0%) |
| Avoidant Personality Disorder | 5 (20.0%) | 3 (23.1%) |
| Schizotypal Personality Disorder | 2 (8.0%) | 1 (7.7%) |
| Narcissistic Personality Disorder | 0 (0.0%) | 2 (15.4%) |
| Antisocial Personality Disorder | 0 (0.0%) | 1 (7.7%) |
| Schizoid Personality Disorder | 1 (4.0%) | 1 (7.7%) |
| Paranoid Personality Disorder | 3 (12.0%) | 0 (0.0%) |
| Personality Disorder Not Otherwise Specified (NOS) | 5 (20.0%) | 2 (15.4%) |
| Variable | N | Mean | SD | 95% CI Lower | 95% CI Upper |
|---|---|---|---|---|---|
| SCL-90-R Global Severity Index (GSI) | 38 | 1.89 | 0.72 | 1.65 | 2.13 |
| PID-5 Negative Affectivity | 38 | 1.87 | 0.62 | 1.67 | 2.07 |
| PID-5 Detachment | 38 | 1.48 | 0.47 | 1.33 | 1.63 |
| PID-5 Antagonism | 38 | 0.98 | 0.59 | 0.79 | 1.17 |
| PID-5 Disinhibition | 38 | 1.62 | 0.57 | 1.44 | 1.8 |
| PID-5 Psychoticism | 38 | 1.55 | 0.7 | 1.33 | 1.77 |
| Note. 95% confidence intervals were computed post hoc using standard formulas based on the t distribution. No additional analyses were performed beyond those originally conducted in Jamovi. | |||||
