Introduction
Beyond its considerable economic burden and association with a substantially reduced quality of life (Leichsenring et al. Reference Leichsenring, Bertsch, Domes, Huber, Nolte, Philipsen and Herpertz2024; Wibbelink et al. Reference Wibbelink, Arntz, Kamphuis, Groot, Sinnaeve and Evers2025), borderline personality disorder (BPD) affects a significant portion of the population. While its prevalence is estimated at 1.6–5.9% in the general public, it is far more common in treatment settings, with rates of 9.3–10% among outpatients and between 15% and 25% on inpatient psychiatric units, even reaching 50% in some reports (Kulacaoglu et al. Reference Kulacaoglu, Asoglu and Kose2018; Leichsenring et al. Reference Leichsenring, Leibing, Kruse, New and Leweke2011; Stepp et al. Reference Stepp, Lazarus and Byrd2016).
Given that multiple specialized psychotherapies have proven effective in treating BPD (Leichsenring et al. Reference Leichsenring, Heim, Leweke, Spitzer, Steinert and Kernberg2023) and numerous studies have evaluated psychopharmacological treatment (Gartlehner et al. Reference Gartlehner, Crotty, Kennedy, Edlund, Ali, Siddiqui, Fortman, Wines, Persad and Viswanathan2021; Stoffers-Winterling et al. Reference Stoffers-Winterling, Storebø, Völlm, Mattivi, Nielsen, Kielsholm, Faltinsen, Simonsen and Lieb2022; Völlm and Cerci Reference Völlm and Cerci2025), diagnostic accuracy remains a matter of critical importance. At the same time, the diagnosis of BPD sits at the center of a broader debate concerning categorical and dimensional approaches to personality disorder (PD) classification (Aftab et al. Reference Aftab, Banicki, Ruffalo and Frances2024; Paris Reference Paris, Lejuez and Gratz2020). This debate has become especially important in light of the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, Text Revision (DSM-5-TR; American Psychiatric Association 2022) and the International Classification of Diseases, 11th Revision (ICD-11; World Health Organization 2019), both of which reflect attempts, in different ways, to move beyond a purely categorical model of personality pathology (Monaghan and Bizumic Reference Monaghan and Bizumic2023; Paris Reference Paris, Lejuez and Gratz2020).
Dimensional approaches are therefore not merely a future possibility, but already form part of contemporary diagnostic practice. ICD-11 represents a particularly important shift in this regard: rather than retaining a full set of distinct categorical PD diagnoses, it begins with the diagnosis of PD, specifies severity, allows the description of trait domain qualifiers, and includes a borderline pattern specifier (Reed et al. Reference Reed, First, Kogan, Hyman, Gureje, Gaebel, Maj, Stein, Maercker, Tyrer, Claudino, Garralda, Salvador‐Carulla, Ray, Saunders, Dua, Poznyak, Medina‐Mora, Pike, Ayuso‐Mateos, Kanba, Keeley, Khoury, Krasnov, Kulygina, Lovell, de Jesus Mari, Maruta, Matsumoto, Rebello, Roberts, Robles, Sharan, Zhao, Jablensky, Udomratn, Rahimi‐Movaghar, Rydelius, Bährer‐Kohler, Watts and Saxena2019; Tyrer et al. Reference Tyrer, Mulder, Kim and Crawford2019; World Health Organization 2019). DSM-5-TR, by contrast, retains the familiar categorical PD diagnoses in Section II while also including the Alternative Model for Personality Disorders (AMPD) in Section III, which integrates impairments in personality functioning with maladaptive trait dimensions and retained PD types, including BPD (American Psychiatric Association 2022; Sharp et al. Reference Sharp, Clark, Balzen, Widiger, Stepp, Zimmerman and Krueger2025). Thus, both ICD-11 and DSM-5 AMPD already include hybrid/dimensionalized approaches rather than simple continuations of traditional categorical diagnosis. The ICD-11 borderline pattern specifier is important for the present argument because it occupies a distinctive position within an otherwise dimensionalized classification of PD. In contrast to anankastia, which is retained as a trait-domain qualifier rather than as a named PD pattern, the borderline pattern remains the only substantial named PD pattern carried forward from the previous categorical system. This makes BPD a particularly informative case for examining when named PD patterns may retain clinical value within dimensionalized assessment.
The clinical question is therefore no longer whether dimensional assessment should be incorporated into the evaluation of personality pathology. The more precise question is how a clinically and empirically supported construct such as BPD should be recognized, sequenced, and dimensionally characterized in future diagnostic and clinical practice. This issue is particularly important because BPD remains one of the most extensively studied PDs, with substantial evidence regarding prevalence, genetics, longitudinal course, treatment response, risk, functional impairment, and clinical utility (Gunderson Reference Gunderson2010; Leichsenring et al. Reference Leichsenring, Heim, Leweke, Spitzer, Steinert and Kernberg2023, Reference Leichsenring, Bertsch, Domes, Huber, Nolte, Philipsen and Herpertz2024; Paris Reference Paris, Lejuez and Gratz2020; Weinberg Reference Weinberg, Lejuez and Gratz2020; Zanarini et al. Reference Zanarini, Frankenburg, Hennen, Reich and Silk2005, Reference Zanarini, Frankenburg, Hein, Glass and Fitzmaurice2024).
BPD is therefore a clinically important case for examining the broader categorical-dimensional tension in psychiatric diagnosis. This focus does not imply that other PD patterns are clinically irrelevant, but reflects the particular clinical, empirical, and classificatory importance of BPD within contemporary diagnostic debates. Categorical diagnosis provides a recognizable clinical anchor, supports communication among clinicians, facilitates continuity with the existing research literature, and can guide access to specialized treatment (First et al. Reference First, Rebello, Keeley, Bhargava, Dai, Kulygina, Matsumoto, Robles, Stona and Reed2018; Gunderson et al. Reference Gunderson, Herpertz, Skodol, Torgersen and Zanarini2018; Leichsenring et al. Reference Leichsenring, Heim, Leweke, Spitzer, Steinert and Kernberg2023). It can also point to a discrete number of natural forms by which disorders of personality coalesce. Dimensional characterization, however, adds information about severity, personality functioning, trait configuration, functional impairment, and clinical heterogeneity, while outcome/risk assessment clarifies prognostic and risk-related factors relevant to likely course, level of care, and clinical management (American Psychiatric Association 2022; Hopwood et al. Reference Hopwood, Kotov, Krueger, Watson, Widiger, Althoff, Ansell, Bach, Michael Bagby, Blais, Bornovalova, Chmielewski, Cicero, Conway, De Clercq, De Fruyt, Docherty, Eaton, Edens, Forbes, Forbush, Hengartner, Ivanova, Leising, John Livesley, Lukowitsky, Lynam, Markon, Miller, Morey, Mullins‐Sweatt, Hans Ormel, Patrick, Pincus, Ruggero, Samuel, Sellbom, Slade, Tackett, Thomas, Trull, Vachon, Waldman, Waszczuk, Waugh, Wright, Yalch, Zald and Zimmermann2018; Natoli et al. Reference Natoli, Murdock, Merguie and Hopwood2025; Sharp et al. Reference Sharp, Clark, Balzen, Widiger, Stepp, Zimmerman and Krueger2025; World Health Organization 2019). These approaches should not necessarily be understood as mutually exclusive. Rather, they perform different functions and may be most useful when integrated in a clear assessment sequence.
This question is also relevant to current discussions about the future direction of DSM. Recent proposals concerning future DSM revisions emphasize the need to preserve clinical utility while incorporating dimensionality, severity, functioning, quality of life, contextual factors, and transdiagnostic features more explicitly into psychiatric assessment (Drexler et al. Reference Drexler, Alpert, Benton, Fung, Gogtay, Malaspina, O’Keefe, Oquendo Mía, Wainberg, Yonkers, Yousif and Clarke2026; Öngür et al. Reference Öngür, Abi-Dargham, Clarke, Compton, Cuthbert, Fung, Gogtay, Kas, Kumar, Malaspina, O’Keefe, Oquendo, Wainberg, Yonkers, Yousif and Alpert2026; Oquendo et al. Reference Oquendo, Clarke, Kas and Wainberg2026). These discussions do not suggest BPD becomes the basis for a broader DSM-6 proposal. However, they provide a useful context for considering how a well-established categorical construct can be retained while dimensional and cross-cutting domains are assessed more systematically.
In this article, we examine BPD as a clinically important case through which to explore the tension between categorical and dimensional approaches to PD diagnosis. We first review the advantages and limitations of categorical and dimensional models, with direct attention to ICD-11 and DSM-5 AMPD. We then consider the reliability, validity, clinical utility, and treatment relevance of the BPD construct. Finally, we outline a clinically oriented sequencing framework in which clinicians first establish the presence of PD, then assess whether a specific PD pattern, such as BPD or the borderline pattern, is relevant to formulation, prognosis, and treatment planning, before proceeding to dimensional characterization and outcome/risk assessment. This framework is not proposed as a replacement for ICD-11 or DSM-5 AMPD, nor as a diagnostic algorithm, but as a clinical sequencing approach for coordinating general PD diagnosis, named PD patterns, dimensional characterization, and risk/prognostic assessment. In the present article, this question is examined primarily through BPD, where the preservation of the borderline diagnosis remains clinically consequential. Its broader purpose is to support the retention of named categorical PD diagnoses when they remain clinically and empirically meaningful, with BPD serving as the central case examined in this article.
Dimensional and categorical approaches: clinical contributions and limitations
Having situated BPD within contemporary hybrid and dimensionalized systems, it is necessary to clarify what categorical and dimensional approaches each contribute clinically. Dimensional approaches to personality pathology have a long history, and numerous dimensional frameworks have been applied to PD classification and assessment. Alongside formal diagnostic models such as ICD-11 and the DSM-5 AMPD, these include the Five Factor Model (FFM), HEXACO, the Schedule for Nonadaptive and Adaptive Personality (SNAP), the Dimensional Assessment of Personality Pathology (DAPP), and Cloninger’s Temperament and Character Inventory (TCI), among others (Feher and Vernon Reference Feher and Vernon2021). This lack of consensus can complicate the clinical implementation of dimensional approaches and underscores the continued value of maintaining categorical anchors – particularly for BPD – while integrating both perspectives clinically.
Despite these implementation challenges, dimensional models show substantial utility: much of the prognostic value in PD data stems from dimensional rather than categorical classifications (Hopwood et al. Reference Hopwood, Kotov, Krueger, Watson, Widiger, Althoff, Ansell, Bach, Michael Bagby, Blais, Bornovalova, Chmielewski, Cicero, Conway, De Clercq, De Fruyt, Docherty, Eaton, Edens, Forbes, Forbush, Hengartner, Ivanova, Leising, John Livesley, Lukowitsky, Lynam, Markon, Miller, Morey, Mullins‐Sweatt, Hans Ormel, Patrick, Pincus, Ruggero, Samuel, Sellbom, Slade, Tackett, Thomas, Trull, Vachon, Waldman, Waszczuk, Waugh, Wright, Yalch, Zald and Zimmermann2018). Dimensional severity provides a clinically useful marker for describing clinical status and guiding prognosis (Natoli et al. Reference Natoli, Murdock, Merguie and Hopwood2025), and recent findings continue to support the predictive validity of dimensional frameworks (Stinnett Reference Stinnett2025).
This dimensional utility is also evident in studies of BPD. Studies in both adult and adolescent samples demonstrate substantial convergence between categorical BPD diagnosis and dimensional models, including the hybrid DSM-5 AMPD. Depending on the method and sample, dimensional indicators explain between 33% and 74% of the variance in BPD features – a broad range suggesting that, although these models capture essential aspects of the disorder and provide considerable explanatory power, the overlap is substantial but not complete. The remaining variance, and the variability across studies, indicate that dimensional representations, while informative, do not simply reproduce the categorical diagnosis and may provide partly different clinical information (Anderson et al. Reference Anderson, Snider, Sellbom, Krueger and Hopwood2014; Barkauskienė et al. Reference Barkauskienė, Gaudiešiūtė, Adler, Gervinskaitė-Paulaitienė, Laurinavičius and Skabeikytė-Norkienė2022; Few et al. Reference Few, Miller, Rothbaum, Meller, Maples, Terry, Collins and MacKillop2013; Hopwood et al. Reference Hopwood, Kotov, Krueger, Watson, Widiger, Althoff, Ansell, Bach, Michael Bagby, Blais, Bornovalova, Chmielewski, Cicero, Conway, De Clercq, De Fruyt, Docherty, Eaton, Edens, Forbes, Forbush, Hengartner, Ivanova, Leising, John Livesley, Lukowitsky, Lynam, Markon, Miller, Morey, Mullins‐Sweatt, Hans Ormel, Patrick, Pincus, Ruggero, Samuel, Sellbom, Slade, Tackett, Thomas, Trull, Vachon, Waldman, Waszczuk, Waugh, Wright, Yalch, Zald and Zimmermann2018; Sellbom et al. Reference Sellbom, Sansone, Songer and Anderson2014; Weekers et al. Reference Weekers, Hutsebaut, Zimmermann and Kamphuis2022; Yam and Simms Reference Yam and Simms2014). For a comprehensive overview of studies up to 2022 see Vanwoerden and Stepp (Reference Vanwoerden and Stepp2022), while for a review of the reliability and validity of a dimensional model of PD – the DSM-5 AMPD, with BPD as the most extensively studied case – see Sharp et al. (Reference Sharp, Clark, Balzen, Widiger, Stepp, Zimmerman and Krueger2025), and for a position strongly advocating its clinical adoption after a decade of empirical research, see Bach and Tracy (Reference Bach and Tracy2022). Direct comparisons of Section II and AMPD criteria likewise indicate moderate to strong agreement, further suggesting that dimensional and categorical models demonstrate substantial alignment while continuing to perform partly distinct clinical and classificatory functions (Clark et al. Reference Clark, Ro, Nuzum, Vanderbleek and Allen2024; Mulay et al. Reference Mulay, Waugh, Fillauer, Bender, Bram, Cain, Caligor, Forbes, Goodrich, Kamphuis, Keeley, Krueger, Kurtz, Jacobsson, Lewis, Rossi, Ridenour, Roche, Sellbom, Sharp and Skodol2019; Sharp et al. Reference Sharp, Clark, Balzen, Widiger, Stepp, Zimmerman and Krueger2025).
The value of dimensional assessment is partly explained by the limitations of categorical systems (Hopwood et al. Reference Hopwood, Kotov, Krueger, Watson, Widiger, Althoff, Ansell, Bach, Michael Bagby, Blais, Bornovalova, Chmielewski, Cicero, Conway, De Clercq, De Fruyt, Docherty, Eaton, Edens, Forbes, Forbush, Hengartner, Ivanova, Leising, John Livesley, Lukowitsky, Lynam, Markon, Miller, Morey, Mullins‐Sweatt, Hans Ormel, Patrick, Pincus, Ruggero, Samuel, Sellbom, Slade, Tackett, Thomas, Trull, Vachon, Waldman, Waszczuk, Waugh, Wright, Yalch, Zald and Zimmermann2018; Tyrer and Mulder Reference Tyrer and Mulder2024), most notably the tendency to yield multiple overlapping PD diagnoses, inflating comorbidity rates and obscuring boundaries between disorders (American Psychiatric Association 2022; Bernstein et al. Reference Bernstein, Iscan and Maser2007; Kendell and Jablensky Reference Kendell and Jablensky2003; Paris Reference Paris, Lejuez and Gratz2020; Sharp et al. Reference Sharp, Clark, Balzen, Widiger, Stepp, Zimmerman and Krueger2025). This diagnostic overlap can complicate treatment planning and access to care, motivating dimensional and hybrid models that aim to reduce artifactual comorbidity (Aftab et al. Reference Aftab, Banicki, Ruffalo and Frances2024; Clark et al. Reference Clark, Ro, Nuzum, Vanderbleek and Allen2024; Krueger and Markon Reference Krueger and Markon2014; Reed et al. Reference Reed, First, Kogan, Hyman, Gureje, Gaebel, Maj, Stein, Maercker, Tyrer, Claudino, Garralda, Salvador‐Carulla, Ray, Saunders, Dua, Poznyak, Medina‐Mora, Pike, Ayuso‐Mateos, Kanba, Keeley, Khoury, Krasnov, Kulygina, Lovell, de Jesus Mari, Maruta, Matsumoto, Rebello, Roberts, Robles, Sharan, Zhao, Jablensky, Udomratn, Rahimi‐Movaghar, Rydelius, Bährer‐Kohler, Watts and Saxena2019; Tyrer et al. Reference Tyrer, Mulder, Kim and Crawford2019; World Health Organization 2019).
Dimensional trait models may also refine differential diagnosis. In a matched case-control study, DSM-5 AMPD traits reliably differentiated diagnostic groups: emotional lability consistently emerged as the most distinctive marker of BPD, with risk taking and suspiciousness distinguishing BPD from other PDs, while depressivity and suspiciousness differentiated BPD from healthy controls (Bach et al. Reference Bach, Sellbom, Bo and Simonsen2016). These findings highlight the incremental value of trait-based assessment, complementing rather than replacing the categorical framework. Anderson et al. (Reference Anderson, Sellbom, Sansone and Songer2016) reported that both DSM-5 Section II and DSM-5 AMPD operationalizations of BPD showed strong associations with external correlates, with Section II more strongly linked to behavioral outcomes such as self-harm, but AMPD traits often demonstrating comparable or superior predictive validity across other domains.
Although many dimensional instruments are self-report based, dimensional characterization can also draw on multiple sources of information, including structured clinical interview, clinician-rated assessment, clinical observation, collateral information, developmental history, longitudinal course, and functional assessment (Krueger and Markon Reference Krueger and Markon2014). Integrating these perspectives can yield the most comprehensive clinical picture, reinforcing the value of combining categorical recognition with dimensional characterization.
If dimensional assessment adds important nuance, categorical diagnoses continue to offer the advantage of a clear and rapid summary of a patient’s presentation, risks, and clinical needs – an aspect particularly valuable in busy health care settings and for training the wider clinical workforce (American Psychiatric Association 2022; First et al. Reference First, Rebello, Keeley, Bhargava, Dai, Kulygina, Matsumoto, Robles, Stona and Reed2018; Kendell and Jablensky Reference Kendell and Jablensky2003). A diagnostic label can validate suffering and provide access to structured treatments, yet the term “borderline” remains heavily stigmatized and may lead clinicians to dismiss patients as manipulative or undeserving of care (American Psychiatric Association 2022; Campbell et al. Reference Campbell, Clarke, Massey and Lakeman2020; Gunderson et al. Reference Gunderson, Herpertz, Skodol, Torgersen and Zanarini2018; Herpertz et al. Reference Herpertz, Huprich, Bohus, Chanen, Goodman, Mehlum, Moran, Newton-Howes, Scott and Sharp2017; Leichsenring et al. Reference Leichsenring, Heim, Leweke, Spitzer, Steinert and Kernberg2023; Reference Newhill and RuffaloNewhill and Ruffalo in press; Sheehan et al. Reference Sheehan, Nieweglowski and Corrigan2016; Stiles et al. Reference Stiles, Batchelor, Gumley and Gajwani2023). Dimensional assessments are often considered less stigmatizing because they describe traits along a continuum and highlight changeability, rather than applying a fixed diagnostic label. However, patients may perceive these formulations as too abstract or minimizing, and clinicians may continue to associate certain traits with BPD, allowing stigma to remain (Stiles et al. Reference Stiles, Batchelor, Gumley and Gajwani2023).
The clinical value of categorical recognition is therefore not limited to diagnostic validity. It also lies in the practical functions that a named pattern can perform in care. A categorical diagnosis can facilitate communication among clinicians by rapidly conveying a familiar configuration of risks, relational patterns, treatment needs, and likely clinical challenges (First et al. Reference First, Rebello, Keeley, Bhargava, Dai, Kulygina, Matsumoto, Robles, Stona and Reed2018; Kendell and Jablensky Reference Kendell and Jablensky2003). It can also support communication with patients by offering a recognizable explanation for recurrent difficulties without reducing the person to the diagnosis (Campbell et al. Reference Campbell, Clarke, Massey and Lakeman2020; Sheehan et al. Reference Sheehan, Nieweglowski and Corrigan2016; Stiles et al. Reference Stiles, Batchelor, Gumley and Gajwani2023). For BPD in particular, categorical recognition may identify overlapping clinical features – such as affective instability, identity disturbance, interpersonal hypersensitivity, abandonment fears, impulsivity, self-harm, and recurrent crisis presentations – that may be distributed across several dimensional domains when described trait by trait (Gunderson et al. Reference Gunderson, Herpertz, Skodol, Torgersen and Zanarini2018; Leichsenring et al. Reference Leichsenring, Heim, Leweke, Spitzer, Steinert and Kernberg2023; Triantafyllou et al. Reference Triantafyllou, Konstantakopoulos, Stefanatou, Giannouli and Malogiannis2025). Most importantly, categorical recognition can guide the selection of evidence-based treatment modalities, since treatments such as Dialectical Behavior Therapy (DBT), Mentalization-Based Treatment (MBT), Good Psychiatric Management (GPM), and Transference-Focused Psychotherapy (TFP) were developed, tested, and disseminated in relation to the BPD construct (Cailhol et al. Reference Cailhol, St-Amour, Désilets, Larivière, Mills and Klein2025; Leichsenring et al. Reference Leichsenring, Heim, Leweke, Spitzer, Steinert and Kernberg2023, Reference Leichsenring, Bertsch, Domes, Huber, Nolte, Philipsen and Herpertz2024). Dimensional characterization may then refine treatment focus, intensity, prognosis, and risk management, while categorical recognition continues to provide a clinically useful organizing framework when a named pattern remains relevant to formulation and care (Gunderson et al. Reference Gunderson, Herpertz, Skodol, Torgersen and Zanarini2018; Hopwood et al. Reference Hopwood, Kotov, Krueger, Watson, Widiger, Althoff, Ansell, Bach, Michael Bagby, Blais, Bornovalova, Chmielewski, Cicero, Conway, De Clercq, De Fruyt, Docherty, Eaton, Edens, Forbes, Forbush, Hengartner, Ivanova, Leising, John Livesley, Lukowitsky, Lynam, Markon, Miller, Morey, Mullins‐Sweatt, Hans Ormel, Patrick, Pincus, Ruggero, Samuel, Sellbom, Slade, Tackett, Thomas, Trull, Vachon, Waldman, Waszczuk, Waugh, Wright, Yalch, Zald and Zimmermann2018; Natoli et al. Reference Natoli, Murdock, Merguie and Hopwood2025; Widiger and Mullins-Sweatt Reference Widiger and Mullins-Sweatt2010).
The partial convergence of categorical and dimensional approaches is also supported by psychometric evidence. For example, BPD has demonstrated strong psychometric properties, with confirmatory factor analysis supporting a unidimensional structure (Sanislow et al. Reference Sanislow, Grilo, Morey, Bender, Skodol, Gunderson, Shea, Stout, Zanarini and McGlashan2002), individual criteria showing adequate reliability and validity (Sanislow et al. Reference Sanislow, Grilo, Morey, Bender, Skodol, Gunderson, Shea, Stout, Zanarini and McGlashan2002), and the Revised Diagnostic Interview for Borderlines (DIB-R) yielding good interrater reliability with κ values in the moderate to high range (Zanarini et al. Reference Zanarini, Frankenburg and Vujanovic2002). Longitudinal research has further shown that BPD is a stable diagnosis rather than a transient condition (Gunderson et al. Reference Gunderson, Stout, McGlashan, Shea, Morey, Grilo, Zanarini, Yen, Markowitz, Sanislow, Ansell, Pinto and Skodol2011; Zanarini et al. Reference Zanarini, Frankenburg, Hennen, Reich and Silk2005, Reference Zanarini, Frankenburg, Hein, Glass and Fitzmaurice2024), and DSM-5 field trials showed overall good test-retest reliability for BPD (pooled intraclass κ = 0.54, 95% CI: 0.43–0.66; Narrow et al. Reference Narrow, Clarke, Kuramoto, Kraemer, Kupfer, Greiner and Regier2013).
Similarly, DSM-5 AMPD scales have shown good reliability and adequate precision, with Cronbach’s alpha coefficients ranging from 0.72 to 0.96 across scales (Krueger and Markon Reference Krueger and Markon2014). In addition, this model demonstrated substantial shared variance with other established measures, including the Personality Diagnostic Questionnaire-4 (PDQ-4; median correlation = 0.61), DSM-IV criterion counts for the six PD diagnoses in the DSM-5 AMPD (e.g., r = 0.80 for BPD), and expert ratings based on a semi-structured clinical interview (Krueger and Markon Reference Krueger and Markon2014).
Despite this convergence, important challenges remain. Dimensional models improve statistical sensitivity and capture subthreshold features, but the lack of consensus across competing frameworks can complicate comparability and implementation (Feher and Vernon Reference Feher and Vernon2021; Krueger and Markon Reference Krueger and Markon2014). Importantly, much of the existing research on borderline pathology has been conducted under the categorical label BPD, and even dimensional research often situates its findings in relation to this established construct (Gunderson et al. Reference Gunderson, Herpertz, Skodol, Torgersen and Zanarini2018; Leichsenring et al. Reference Leichsenring, Heim, Leweke, Spitzer, Steinert and Kernberg2023; Triantafyllou et al. Reference Triantafyllou, Konstantakopoulos, Stefanatou, Giannouli and Malogiannis2025). The fact that BPD traits tend to “hang together” supports its coherence as a syndrome and justifies its continued use as a research and clinical anchor (Gunderson et al. Reference Gunderson, Herpertz, Skodol, Torgersen and Zanarini2018). It is also worth noting that many nonpsychiatric medical diseases exist on spectra with fuzzy boundaries, but we retain categorical diagnoses to denote these conditions (Guze Reference Guze1992; Roth and Kroll Reference Roth and Kroll1986).
In addition to clinical and research advantages, categorical diagnoses retain value in broader systems and communication contexts. Categorical diagnoses provide a common framework for insurance reimbursement, disability assessments, and access to specialized services (First et al. Reference First, Rebello, Keeley, Bhargava, Dai, Kulygina, Matsumoto, Robles, Stona and Reed2018), facilitate clear interdisciplinary communication across psychiatry, psychology, nursing, and social work (Monaghan and Bizumic Reference Monaghan and Bizumic2023), and support cross-cultural comparability in diverse healthcare settings (Reed Reference Reed and Tyrer2023). At the same time, dimensional models offer broader coverage and more reliable, nuanced descriptions of pathology (Verheul Reference Verheul2005), are often rated by clinicians as easier to use and more effective for communication and treatment planning (Widiger and Mullins-Sweatt Reference Widiger and Mullins-Sweatt2010), and can be organized hierarchically to guide tailored therapeutic interventions (Natoli et al. Reference Natoli, Murdock, Merguie and Hopwood2025). Taken together, these complementary strengths underscore the value of coordinating categorical recognition with dimensional characterization.
In sum, categorical and dimensional approaches each provide clinically relevant information, but neither is sufficient on its own. For BPD, the central question is whether it remains sufficiently reliable, valid, and clinically useful to function as an organizing construct within a dimensionalized assessment framework. This question is addressed in the following section.
Meeting the criteria: establishing the validity of BPD
A well-researched psychiatric diagnosis is validated through several key scientific criteria. Foundational to this is reliability, ensuring different clinicians can consistently arrive at the same diagnosis (American Psychiatric Association 2022). Building on this, the diagnosis must demonstrate multifaceted validity, including its ability to accurately represent a distinct clinical entity (construct validity), predict the illness’s course (predictive validity), and be clearly distinguishable from other disorders (discriminant validity), a framework established decades ago (Robins and Guze Reference Robins and Guze1970; Sharp et al. Reference Sharp, Clark, Balzen, Widiger, Stepp, Zimmerman and Krueger2025; Weinberg Reference Weinberg, Lejuez and Gratz2020). A mature diagnosis is further supported by a growing understanding of its etiology and pathophysiology, incorporating consistent findings from genetics, neurobiology, and environmental risk factor research (Cuthbert and Insel Reference Cuthbert and Insel2013; Kendler Reference Kendler2016). Finally, its clinical utility and validity are supported by the existence of specific, evidence-based treatments shown to be effective for the disorder, indicating that the diagnostic construct can meaningfully guide therapeutic intervention (Leichsenring et al. Reference Leichsenring, Heim, Leweke, Spitzer, Steinert and Kernberg2023).
Modern research supports BPD as a valid clinical diagnosis (Ruffalo Reference Ruffalo2026) with a predictable, though dynamic, life course. Longitudinal studies, as summarized by Temes and Zanarini (Reference Temes and Zanarini2018), show that the diagnosis is stable and not just a transient phase. While the symptoms are most acute in young adulthood, the prognosis is far more optimistic than once believed (Zanarini et al. Reference Zanarini, Frankenburg, Hein, Glass and Fitzmaurice2024). As noted above, the psychometric literature also supports BPD as a reliably identifiable and coherent construct.
Current evidence supports the view that BPD arises from complex gene-environment interactions, with genetic, epigenetic, neurobiological, neuroendocrine, neurochemical, inflammatory, oxidative stress, environmental, and social factors playing a role in its etiology and pathophysiology (Dammann et al. Reference Dammann, Teschler, Haag, Altmüller, Tuczek and Dammann2011; Gescher et al. Reference Gescher, Schanze, Vavra, Wolff, Zimmer-Bensch, Zenker, Frodl and Schmahl2024; Gunderson et al. Reference Gunderson, Herpertz, Skodol, Torgersen and Zanarini2018; Leichsenring et al. Reference Leichsenring, Heim, Leweke, Spitzer, Steinert and Kernberg2023, Reference Leichsenring, Bertsch, Domes, Huber, Nolte, Philipsen and Herpertz2024; for a detailed review, see Radoi and Ruffalo Reference Radoi and Ruffalo2025).
The clinical utility and validity of the BPD diagnosis are supported by the existence of effective, specialized treatments. Psychotherapies such as DBT, MBT, GPM, and TFP have been proven effective in numerous studies (Cailhol et al. Reference Cailhol, St-Amour, Désilets, Larivière, Mills and Klein2025; Leichsenring et al. Reference Leichsenring, Heim, Leweke, Spitzer, Steinert and Kernberg2023, Reference Leichsenring, Bertsch, Domes, Huber, Nolte, Philipsen and Herpertz2024). These therapies target clinically central features of BPD, including emotional dysregulation and interpersonal difficulties, and are associated with meaningful and lasting clinical improvement. While no pharmacological treatment is approved for BPD, virtually all classes of psychiatric medications have been studied, including first- and second-generation antipsychotics, anticonvulsants (mood stabilizers), antidepressants, benzodiazepines, naltrexone, memantine, ketamine, and clonidine (Cailhol et al. Reference Cailhol, St-Amour, Désilets, Larivière, Mills and Klein2025; Gartlehner et al. Reference Gartlehner, Crotty, Kennedy, Edlund, Ali, Siddiqui, Fortman, Wines, Persad and Viswanathan2021; Stoffers-Winterling et al. Reference Stoffers-Winterling, Storebø, Völlm, Mattivi, Nielsen, Kielsholm, Faltinsen, Simonsen and Lieb2022; Völlm and Cerci Reference Völlm and Cerci2025), and polypharmacy is common in BPD, highlighting a need for better psychopharmacological treatment that can be used in busy settings (Tennant et al. Reference Tennant, Frampton, Mulder and Beaglehole2023; Zanarini et al. Reference Zanarini, Frankenburg, Reich, Harned and Fitzmaurice2015).
BPD is also distinguishable from other psychiatric diagnoses based on its core psychopathology - impairments in self-functioning (identity) and interpersonal functioning (Gunderson et al. Reference Gunderson, Herpertz, Skodol, Torgersen and Zanarini2018) - including mood disorders (major depressive disorder and bipolar disorder, particularly type II), complex post-traumatic stress disorder, anxiety disorders, and dissociative disorders, although diagnostic difficulties often persist (Brand and Lanius Reference Brand and Lanius2014; Cloitre et al. Reference Cloitre, Stolbach, Herman, Kolk, Pynoos, Wang and Petkova2009; D’Agostino et al. Reference D’Agostino, Gagliardi, Pagani and Monti2025; Gunderson et al. Reference Gunderson, Herpertz, Skodol, Torgersen and Zanarini2018; Leichsenring et al. Reference Leichsenring, Bertsch, Domes, Huber, Nolte, Philipsen and Herpertz2024; Mendez-Miller et al. Reference Mendez-Miller, Naccarato and Radico2022; Newhill and Ruffalo Reference Newhill and Ruffaloin press; Ogasawara et al. Reference Ogasawara, Nakamura, Kimura, Aleksic and Ozaki2018; Sar et al. Reference Sar, Alioğlu and Akyuz2017). Importantly, mood changes in BPD are distinguishable from those in bipolar spectrum illness on the basis of their reactivity to interpersonal events and relational stressors.
Taken together, these findings demonstrate that BPD fulfills the established scientific criteria for a valid psychiatric diagnosis: it is reliably identified, conceptually coherent, prognostically meaningful, etiologically grounded, clinically useful, and distinguishable from related conditions. This supports its continued use as an organizing categorical construct within broader dimensionalized approaches to PD assessment.
A clinical sequencing model for PD assessment: the case of BPD
Existing hybrid and dimensional systems, including the DSM-5 AMPD and ICD-11, already provide ways of combining categorical diagnosis with dimensional characterization. The issue, therefore, is not whether such integration is possible, but how these elements may be clinically coordinated in practice. This question is especially important in relation to BPD, where the clinical stakes of preserving or losing a named categorical pattern are particularly salient.
On this basis, we propose a clinical sequencing model for PD assessment, developed and illustrated through BPD, within existing hybrid and dimensional frameworks. The model (Figure 1) should be understood as a clinical sequencing model, not as a diagnostic algorithm or a replacement for ICD-11, DSM-5 AMPD, or any formal diagnostic system. Its aim is more limited: to clarify how categorical and dimensional information may be brought into assessment so that neither the general diagnosis of PD, the possible significance of a named PD pattern, nor the individual dimensional profile is lost. In the present article, this question is examined primarily through BPD or the borderline pattern.
Clinical sequencing model for personality disorder assessment.
Note: DSM-5-TR AMPD: Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, Text Revision, Alternative Model for Personality Disorders; ICD 11: International Classification of Diseases, 11th Revision.

The first step is to establish whether personality pathology reaches the threshold for a general PD diagnosis. At this stage, the clinician determines whether disturbances in self-functioning, interpersonal functioning, affect regulation, impulse control, or related domains are enduring, clinically significant, and associated with impairment. This step does not yet require the identification of a specific named PD pattern. Rather, it establishes whether personality pathology is present and whether it provides a clinically meaningful framework for understanding the patient’s difficulties.
The second step is to assess whether a specific PD pattern is meaningful for formulation, communication, prognosis, and treatment planning, with particular attention in this article to BPD or the borderline pattern. This step does not imply that every patient with PD must receive a named categorical diagnosis. Rather, it asks whether the presentation is organized around a recognizable pattern strongly enough for that pattern to guide formulation, communication, prognosis, and treatment planning (Gunderson Reference Gunderson2010; Gunderson et al. Reference Gunderson, Herpertz, Skodol, Torgersen and Zanarini2018; Triantafyllou et al. Reference Triantafyllou, Konstantakopoulos, Stefanatou, Giannouli and Malogiannis2025). In this sense, BPD is not treated as a complete description of the person, but as a clinically meaningful pattern that may organize important aspects of assessment, risk recognition, and therapeutic planning.
The third step is dimensional characterization. This is not a repetition of the preceding pattern-recognition step, even though both steps may draw on overlapping clinical material. Step 2 asks whether there is a meaningful named pattern; Step 3 asks how the patient’s personality pathology is configured. Dimensional characterization can specify severity, trait configuration, personality functioning, functional impairment, and heterogeneity, thereby refining rather than replacing categorical recognition when a named PD pattern, such as BPD, remains relevant to formulation and care (Hopwood et al. Reference Hopwood, Kotov, Krueger, Watson, Widiger, Althoff, Ansell, Bach, Michael Bagby, Blais, Bornovalova, Chmielewski, Cicero, Conway, De Clercq, De Fruyt, Docherty, Eaton, Edens, Forbes, Forbush, Hengartner, Ivanova, Leising, John Livesley, Lukowitsky, Lynam, Markon, Miller, Morey, Mullins‐Sweatt, Hans Ormel, Patrick, Pincus, Ruggero, Samuel, Sellbom, Slade, Tackett, Thomas, Trull, Vachon, Waldman, Waszczuk, Waugh, Wright, Yalch, Zald and Zimmermann2018; Milinkovic and Tiliopoulos Reference Milinkovic and Tiliopoulos2020; Natoli et al. Reference Natoli, Murdock, Merguie and Hopwood2025; Widiger and Mullins-Sweatt Reference Widiger and Mullins-Sweatt2010). This characterization may draw on established dimensional systems such as the FFM (Widiger and Costa Jr. Reference Widiger and Costa2013), the DSM-5 AMPD (American Psychiatric Association 2022; Krueger and Markon Reference Krueger and Markon2014), or the ICD-11 model (Tyrer et al. Reference Tyrer, Mulder, Kim and Crawford2019; World Health Organization 2019). This sequence is broadly consistent with the assessment logic proposed by Widiger and Lowe (Reference Widiger and Lowe2007), but it begins rather than ends with the detailed “personality matching” that a dimensional profile provides. In this way, the model does not require categorical diagnosis where it is not clinically useful, but also does not abandon categorical recognition where a named pattern remains clinically meaningful (Clark et al. Reference Clark, Ro, Nuzum, Vanderbleek and Allen2024; Krueger and Markon Reference Krueger and Markon2014; Widiger and Mullins-Sweatt Reference Widiger and Mullins-Sweatt2010).
The rationale for placing named pattern recognition before more fine-grained dimensional characterization is primarily clinical and pragmatic. In many clinical settings, the earliest assessment questions need to provide high immediate utility with relatively low assessment burden: whether personality pathology reaches the level of disorder, whether the presentation corresponds to a clinically meaningful named pattern, and how this pattern should inform communication, risk recognition, prognosis, and treatment planning. Dimensional characterization then adds a more individualized description of severity, trait configuration, personality functioning, impairment, and heterogeneity. The proposed sequence therefore moves from broad clinical threshold, to clinically recognizable pattern, to progressively more detailed dimensional formulation. This ordering does not make dimensional assessment secondary in importance, but treats it as a necessary refinement of clinical recognition rather than a substitute for it when a named PD pattern remains meaningful.
The fourth step is outcome/risk assessment. In this framework, outcome assessment refers primarily to the evaluation of risk and prognostic factors that shape likely course, level of care, safety, and clinical management, rather than to longitudinal monitoring of treatment response. These include domains such as, but not limited to, suicide risk, quality of life, functional impairment, physical health, substance use or other high-risk behaviors, treatment adherence, and relevant social or occupational factors (Baptista et al. Reference Baptista, Chambon, Hoertel, Olfson, Blanco, Cohen and Jacquet2023; Gunderson et al. Reference Gunderson, Herpertz, Skodol, Torgersen and Zanarini2018; Reed et al. Reference Reed, First, Kogan, Hyman, Gureje, Gaebel, Maj, Stein, Maercker, Tyrer, Claudino, Garralda, Salvador‐Carulla, Ray, Saunders, Dua, Poznyak, Medina‐Mora, Pike, Ayuso‐Mateos, Kanba, Keeley, Khoury, Krasnov, Kulygina, Lovell, de Jesus Mari, Maruta, Matsumoto, Rebello, Roberts, Robles, Sharan, Zhao, Jablensky, Udomratn, Rahimi‐Movaghar, Rydelius, Bährer‐Kohler, Watts and Saxena2019; World Health Organization 2019; Zanarini et al. Reference Zanarini, Frankenburg, Hennen, Reich and Silk2005, Reference Zanarini, Frankenburg, Hein, Glass and Fitzmaurice2024). These domains are not specific to BPD and are part of broader psychiatric assessment; however, in BPD they are especially important because they influence prognosis, risk management, treatment planning, and intensity of care.
Although the sequencing model is formulated for PD assessment more broadly, the present article develops it primarily through BPD. Other named PD patterns, such as antisocial, schizotypal, obsessive-compulsive/anankastic, avoidant, or narcissistic patterns, may also be clinically meaningful in particular cases. They could therefore also be considered at Step 2 of the sequence. BPD remains the focal case because the preservation of the borderline pattern is the central question under examination. Other PD patterns are therefore acknowledged as potential applications of the same sequencing logic, while their detailed clinical utility, overlap, and diagnosis-specific sequencing would require separate analyses.
This clinical sequencing model is consistent with ICD-11 and DSM-5 AMPD in placing general personality pathology at the beginning of assessment, whether this is framed as general PD, severity of personality disturbance, impairment in personality functioning, or related dimensional indicators. It also shares with both systems the view that PD assessment should move beyond a purely categorical model by incorporating dimensional characterization, while still allowing clinically meaningful named disorders to retain a role, as reflected in the ICD-11 borderline pattern specifier and the retained PD types of DSM-5 AMPD. However, the model places greater emphasis on preserving named PD patterns when they remain clinically meaningful, rather than moving directly from general personality pathology to a purely dimensional profile. In this sense, it argues against a fully dimensional approach that would make named PD diagnoses clinically peripheral or unnecessary. Its contribution is not to propose a new diagnostic architecture, but to make explicit the clinical order of questions: whether a named PD pattern is clinically meaningful (whether a specific PD is present), with particular attention here to BPD or the borderline pattern; how the presentation is dimensionally configured; and which risk and prognostic factors shape likely course, level of care, and clinical management (American Psychiatric Association 2022; World Health Organization 2019).
This sequencing clarifies how categorical recognition, dimensional characterization, and outcome/risk assessment can be coordinated when a named PD pattern is clinically meaningful. In this sense, the model does not defend categorical diagnosis as a stand-alone endpoint, but supports the retention of named PD patterns as clinically useful organizing constructs within a broader dimensionalized assessment process. This allows BPD to remain a focal organizing construct without allowing categorical diagnosis to become the endpoint of clinical formulation.
Conclusion
The present review examined BPD as a focal case for addressing the broader categorical-dimensional problem in PD assessment. Categorical diagnosis continues to provide practical clarity for decision-making, communication across disciplines, comparability of findings in research, and access to specialized treatments, while also offering patients a recognizable framework that validates suffering and guides prognosis. At the same time, dimensional assessments contribute predictive validity, capture gradations of severity and subthreshold features, enhance psychometric precision, and provide more nuanced descriptions of individual functioning, thereby supporting tailored interventions and a more refined understanding of treatment trajectories and clinical heterogeneity. While each approach demonstrates clear strengths as well as limitations, their coordination appears most clinically useful when organized in a clear sequence rather than treated as competing alternatives.
The broader implications of this approach extend beyond the comparison of models. Existing systems such as ICD-11 and DSM-5 AMPD already provide hybrid or dimensionalized frameworks. The present article therefore does not propose a new diagnostic architecture, but instead clarifies how categorical and dimensional information may be coordinated in clinical reasoning. In this framework, clinicians first establish the presence of general PD, then consider whether a named PD pattern is clinically meaningful, before proceeding to dimensional characterization and outcome/risk assessment. For research, this approach preserves continuity with the large body of categorical research on BPD while allowing more precise investigation of dimensional traits, severity, functioning, risk, and prognosis. By linking categorical anchors with dimensional characterization, the model supports the retention of named PDs when they remain clinically and empirically meaningful.
Future research could evaluate clinical sequencing approaches across diverse clinical settings, including their impact on diagnostic communication, treatment planning, risk assessment, patient understanding, and health-system implementation. Cross-national studies may also help clarify comparability across DSM and ICD systems, while psychometric research can further refine the relationship between named categorical patterns, dimensional severity, trait configuration, personality functioning, and outcome/risk domains. However, dimensional progress does not necessarily require the abandonment of named PD diagnoses. Future work should clarify when dimensional characterization is sufficient, when named categorical patterns retain clinical utility, and how these forms of information can be coordinated most effectively in practice.
Several limitations should be acknowledged. First, this article offers a narrative and clinically oriented synthesis rather than a systematic review or meta-analysis. Although the manuscript is based on a broad and careful search of the relevant literature, its conclusions necessarily depend on the scope, selection, and interpretation of the studies discussed. Second, the proposed clinical sequencing framework is conceptual and requires empirical evaluation in clinical settings, including assessment of reliability, feasibility, clinician uptake, patient experience, and effects on treatment planning and risk management. Third, although the framework is formulated for PD assessment more broadly, it is developed primarily through BPD; its application to other named PD patterns would require further diagnosis-specific analysis. Fourth, the model does not examine subthreshold personality difficulties in detail, although these may be clinically relevant in dimensionalized systems such as ICD-11 and may require dimensional characterization even when the threshold for a formal PD diagnosis is not met. Finally, the balance between categorical recognition and dimensional characterization may vary across diagnostic systems, services, cultural contexts, and levels of clinical expertise.
In conclusion, BPD remains a clinically and empirically meaningful focal case for examining how named categorical diagnoses can be retained within dimensionalized PD assessment. A clinical sequencing approach allows BPD to remain an organizing construct when clinically meaningful, while ensuring that categorical diagnosis is followed by dimensional characterization and outcome/risk assessment rather than treated as the endpoint of formulation.
Acknowledgments
None.
Funding statement
This research received no specific grant from any funding agency, commercial, or not-for-profit sectors.
Competing interests
The authors declare none.
Ethical standards
This article is a conceptual analysis and contains no original research involving human participants; therefore, ethical approval and informed consent were not required.