Personality Functioning, Maladaptive Traits, and Emotion Dysregulation in Severe Psychiatric Inpatients: Associations with Psychological Symptom Domains

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

In general, this manuscript is clinically relevant and well-conducted, with a distinct research question, a sound methodological framework, and high levels of consistency among the study objectives, methods, results, and conclusions. The HiTOP framework's integration of personality functioning, maladaptive personality traits, and emotion dysregulation is a conceptually novel and significant contribution to the literature, particularly from a transdiagnostic perspective. The scientific rigor of the study is enhanced by the utilization of validated psychometric instruments and appropriate statistical analyses, such as hierarchical regression models with bootstrap confidence intervals. However, numerous areas could be enhanced.

1) The introduction would be enhanced by a more explicit emphasis on the study's originality and the inclusion of a few more recent references on the Level of Personality Functioning and HiTOP.

2)The observed associations may be influenced by the diagnostic procedures, the interval between assessments, and the pharmacological treatment of the participants, which is why the Methods section should provide more detail.

3)The Results section is rather dense, but it is comprehensive. The presentation of the regression analyses could be simplified to improve readability, and a graphical summary of the principal findings would also enhance the manuscript.

4)The tables are informative, well-organized, and provide sufficient detail to support the findings. However, Tables 3 and 4 are densely populated with statistical information and could be simplified or reformatted to improve readability. 

5)Given the cross-sectional design, it is advisable to exercise greater caution when expressing statements that suggest explanatory mechanisms or causal relationships in the Discussion and Conclusions.

 

Author Response

Reviewer 1

In general, this manuscript is clinically relevant and well-conducted, with a distinct research question, a sound methodological framework, and high levels of consistency among the study objectives, methods, results, and conclusions. The HiTOP framework's integration of personality functioning, maladaptive personality traits, and emotion dysregulation is a conceptually novel and significant contribution to the literature, particularly from a transdiagnostic perspective. The scientific rigor of the study is enhanced by the utilization of validated psychometric instruments and appropriate statistical analyses, such as hierarchical regression models with bootstrap confidence intervals. However, numerous areas could be enhanced.

• The introduction would be enhanced by a more explicit emphasis on the study's originality and the inclusion of a few more recent references on the Level of Personality Functioning and HiTOP.

We thank the Reviewer for this valuable suggestion. We have revised the final part of the Introduction to state more explicitly the originality of the present study. Specifically, we now clarified that, to our knowledge, this is among the first studies to examine simultaneously the contributions of Level of Personality Functioning (LPF), maladaptive personality traits, alexithymia, and cognitive emotion-regulation strategies to multiple dimensional symptom domains in a transdiagnostic sample of severe psychiatric inpatients. The revised Introduction also explains that this integrative approach extends previous work on HiTOP and LPF by examining the unique and overlapping contributions of these constructs within a diagnostically heterogeneous inpatient population, rather than considering them in isolation. In addition, we have incorporated recent literature addressing the empirical structure and clinical relevance of HiTOP, the associations between broad psychopathology dimensions and functional impairment, the relationship between LPF and maladaptive personality traits, and the transdiagnostic role of emotion-regulation difficulties. These revisions have been made in the final paragraphs of the Introduction, immediately before the hypotheses. We believe that these revisions provide a clearer articulation of the study’s contribution and better situate the work within current dimensional and transdiagnostic models of psychopathology.

• The observed associations may be influenced by the diagnostic procedures, the interval between assessments, and the pharmacological treatment of the participants, which is why the Methods section should provide more detail.

We thank the reviewer for this important suggestion. We have added a dedicated paragraph in the Methods section explicitly describing how diagnoses were assigned by referring services and subsequently confirmed/revised by the ward psychiatrist according to ICD‑9 criteria; (b) the standardized two‑week interval between T0 and T1, maintained across both intensive and residential settings; and (c) the status of pharmacological treatment data. As now clarified, while pharmacotherapy was subject to continuous clinical monitoring (daily in intensive units, weekly in residential facilities) as part of routine care, individual drug classes and dosages were not systematically recorded in the research database and therefore could not be analytically modeled. We also acknowledge that T0 and T1 do not represent perfectly simultaneous clinical snapshots.

• The Results section is rather dense, but it is comprehensive. The presentation of the regression analyses could be simplified to improve readability, and a graphical summary of the principal findings would also enhance the manuscript.

We thank the reviewer for this constructive feedback. We have revised the Results section to improve readability while preserving all key findings. Specifically, we have simplified the presentation by restructuring the regression results into a bulleted summary organized by symptom domain. This reduces narrative density and allows readers to quickly identify the principal predictors for each outcome. Next, we added a graphical summary in the form of Figure 1, a forest plot that displays regression coefficients and their 95% BCa bootstrap confidence intervals across all predictors and symptom domains. This figure visually highlights significant (red) and non-significant (black) associations, providing an immediate overview of the pattern of results. Detailed statistical information remains available in Table A1 (Appendix A) for readers seeking full numerical detail. Finally, we strengthened the narrative integration of the figure by explicitly referencing it in the text and summarizing the two key patterns that emerge from the graphical display: the dissociation between self-functioning impairment (predicting internalizing symptoms) and interpersonal functioning impairment (predicting interpersonal symptoms). We believe these changes substantially improve the readability and accessibility of the Results section while maintaining its comprehensiveness.

• The tables are informative, well-organized, and provide sufficient detail to support the findings. However, Tables 3 and 4 are densely populated with statistical information and could be simplified or reformatted to improve readability. 

In response to this comment, we have revised the presentation of the regression results to improve readability and facilitate the interpretation of the main findings. Specifically, Tables 3 and 4 have been replaced in the main text with forest plots, which provide a more immediate and visually accessible representation of the magnitude and direction of the regression coefficients and their confidence intervals. The original tables have been moved to the Appendix, where they are available for readers who wish to consult the complete numerical results and detailed statistical information. We believe that this revised presentation improves the clarity of the main text while preserving access to the full statistical results.

• Given the cross-sectional design, it is advisable to exercise greater caution when expressing statements that suggest explanatory mechanisms or causal relationships in the Discussion and Conclusions.

We thank the Reviewer for this important comment. We agree that, given the cross-sectional design, the interpretation of the findings should be framed in terms of associations rather than causal or explanatory relationships. We have therefore carefully revised the Discussion and Conclusions to avoid language that could imply causality, mediation, or temporal ordering. We believe these revisions provide a more appropriately cautious interpretation of the findings while preserving their theoretical relevance and consistency with the existing literature.

Reviewer 2 Report

Comments and Suggestions for Authors

Brief summary: This manuscript examines the relationships among personality functioning, maladaptive personality traits, alexithymia, emotion regulation strategies, and psychological symptoms in a transdiagnostic psychiatric inpatient population. The study addresses an important and clinically relevant question and provides novel insights into the potential role of personality functioning within dimensional models of severe psychopathology.
Please see my comments below.

Introduction

• The theoretical rationale linking HiTOP, personality functioning (LPF), and emotion dysregulation is generally well developed. However, the distinction between maladaptive personality traits and personality functioning could be explained more explicitly for readers who are less familiar with the Alternative DSM-5 Model for Personality Disorders.
• While the manuscript discusses psychosis extensively, the sample includes a diagnostically heterogeneous inpatient population. The introduction would benefit from a clearer justification for focusing primarily on psychosis-related literature when the study is not restricted to psychotic disorders.
• The hypotheses are comprehensive, but some appear highly specific relative to the study design and sample size. Consider simplifying the hypotheses and distinguishing primary from exploratory expectations.
• Additional discussion of prior research examining LPF and symptom severity in transdiagnostic inpatient populations would strengthen the rationale and better position the study within the broader literature.
• The proposed role of alexithymia as a mediator or explanatory mechanism is interesting, but the introduction could more clearly describe the theoretical pathways through which alexithymia may connect personality functioning and symptom expression.

Methods

• The inclusion and exclusion criteria are clearly described. However, additional information regarding illness severity, duration of illness, medication status, and current treatment exposure would provide important clinical context for interpreting the findings.
• Given the relatively small sample size and the large number of predictors entered into the regression models, additional discussion of model complexity and potential risks of overfitting would strengthen the methodological transparency.
• The procedures used to support participants with reading difficulties or cognitive challenges should be described in greater detail. Clarification regarding how assessor assistance was standardized would help readers evaluate the potential influence on self-report responses.
• Although the use of bootstrap confidence intervals is commendable, the manuscript would benefit from reporting the number of bootstrap replications used and providing additional justification for the sign-flip criterion.
• The decision to aggregate cognitive emotion regulation strategies into adaptive and maladaptive composite scores should be justified more thoroughly, as important information from individual emotion regulation strategies may have been lost.

Results

• The presentation of the regression analyses is thorough. However, the Results section would benefit from greater emphasis on effect sizes and clinical significance rather than statistical significance alone.
• The correlation matrix demonstrates substantial overlap among personality functioning, alexithymia, and maladaptive traits. Consider reporting additional indicators of shared variance to assist readers in understanding the degree of construct overlap.
• Several findings differ between the standard and hierarchical regression models. A concise summary table highlighting predictors that remained significant after adjustment would improve clarity and facilitate interpretation.
• The finding that alexithymia lost significance after personality variables were introduced is potentially important. Greater attention to the magnitude of these changes would strengthen the interpretation.
• Some sections of the Results move into explanatory interpretation. Consider limiting this section to description of findings and reserving mechanistic explanations for the Discussion.

Discussion

• The discussion appropriately highlights the potential importance of personality functioning in severe psychopathology. However, some interpretations appear stronger than warranted given the cross-sectional nature of the study and should be framed more cautiously.
• The distinction between self-functioning and interpersonal functioning represents one of the manuscript's most interesting contributions. Expanding the discussion of potential clinical applications of these differential findings would enhance the paper's translational value.
• The discussion of alexithymia is thoughtful, but statements suggesting mediation or indirect effects should be interpreted cautiously because formal mediation analyses were not conducted.
• Greater attention should be given to the transdiagnostic composition of the sample. The extent to which the findings may differ across psychotic disorders, mood disorders, and personality disorders deserves further consideration.
• The clinical implications section is promising. However, recommendations regarding treatment approaches, including mentalization-based interventions and emotion regulation interventions, should be presented as hypotheses for future investigation rather than direct implications of the current data.

 

Author Response

Reviewer 2.

Brief summary: This manuscript examines the relationships among personality functioning, maladaptive personality traits, alexithymia, emotion regulation strategies, and psychological symptoms in a transdiagnostic psychiatric inpatient population. The study addresses an important and clinically relevant question and provides novel insights into the potential role of personality functioning within dimensional models of severe psychopathology.
Please see my comments below.

Introduction

• The theoretical rationale linking HiTOP, personality functioning (LPF), and emotion dysregulation is generally well developed. However, the distinction between maladaptive personality traits and personality functioning could be explained more explicitly for readers who are less familiar with the Alternative DSM-5 Model for Personality Disorders.

To improve clarity for readers less familiar with the Alternative DSM-5 Model for Personality Disorders, we have added a brief explanation distinguishing maladaptive personality traits from the Level of Personality Functioning (LPF). Specifically, the revised text clarifies that maladaptive traits represent relatively enduring stylistic tendencies and dispositions through which personality pathology is expressed, whereas LPF captures core impairments in self-functioning. We further clarify that, although these constructs are empirically related, they are not interchangeable: maladaptive traits describe the characteristic style of dysfunction, while LPF primarily reflects the severity of impairment in self and interpersonal functioning.

• While the manuscript discusses psychosis extensively, the sample includes a diagnostically heterogeneous inpatient population. The introduction would benefit from a clearer justification for focusing primarily on psychosis-related literature when the study is not restricted to psychotic disorders.

We agree that the original introduction did not adequately justify the focus on a diagnostically heterogeneous sample, and we have revised the text accordingly. Specifically, we have inserted a new paragraph in the initial part of the introduction that grounds the transdiagnostic relevance of our work in the core HiTOP premise that maladaptive personality traits are dimensional constructs cutting across traditional diagnostic boundaries. In addition, it explains that the traits under study in psychotic samples (detachment, psychoticism, negative affectivity) reach pathological severity across severe mental illness broadly, not exclusively in psychotic disorders. Last, the new paragraph reframes our mixed sample as a strength that permits testing whether trait-clinical relationships established in psychosis generalize across the wider spectrum of inpatient psychopathology. This approach also aligns with the clinical reality of inpatient settings, where comorbidity and non-categorical presentations are common. We believe these revisions provide a clear, theoretically principled justification for our design.

• The hypotheses are comprehensive, but some appear highly specific relative to the study design and sample size. Consider simplifying the hypotheses and distinguishing primary from exploratory expectations.

We agree that some of our original hypotheses were overly specific given the study's design and sample size, and we have revised this section accordingly. Specifically, we have streamlined the predictions into two primary hypotheses (one focused on LPF as an independent predictor of symptom severity, and one on alexithymia and emotion-regulation strategies as incremental predictors) while the more fine-grained expectations (including differential trait effects, self- vs. interpersonal functioning comparisons, and domain-specific strategy associations) are now explicitly designated as exploratory. This structure clarifies which predictions are theoretically central and adequately powered, while transparently framing the more granular analyses as hypothesis-generating. We believe this revision strengthens the conceptual rigor of the manuscript and better aligns the stated expectations with the study's methodological scope.

• Additional discussion of prior research examining LPF and symptom severity in transdiagnostic inpatient populations would strengthen the rationale and better position the study within the broader literature.

We thank the reviewer for this helpful suggestion. In response, we have expanded the Introduction to incorporate prior research examining LPF in transdiagnostic clinical and community populations. Specifically, we now discussed evidence linking impairments in self- and interpersonal functioning to psychosocial and social-occupational dysfunction, internalizing and externalizing symptom dimensions, and broader dimensions of psychopathology. We further clarified the extent to which LPF and maladaptive traits overlap while highlighting evidence that they may account for unique variance across outcomes. Finally, we explicitly identified the limited evidence regarding whether LPF contributes to the explanation of symptom severity beyond maladaptive personality traits in diagnostically heterogeneous inpatient populations, thereby highlighting the rationale and contribution of the present study.

• The proposed role of alexithymia as a mediator or explanatory mechanism is interesting, but the introduction could more clearly describe the theoretical pathways through which alexithymia may connect personality functioning and symptom expression.

We agree that the theoretical pathways linking alexithymia to personality functioning and symptom expression required clearer articulation, and we have substantially expanded this section of the introduction. Specifically, we now explicitly ground the proposed pathway in the DSM-5-TR conceptualization of emotional awareness and management as integral components of self-functioning, ...from which we derive the proposition that LPF deficits and alexithymia may reflect partly overlapping impairments in the capacity to identify, label, and regulate emotional experiences, that is, alexithymia. We then delineate three specific routes through which these difficulties may contribute to general symptom severity across psychopathology domains: (i) difficulty identifying feelings leading to diffuse distress manifesting as somatic, anxiety, or depressive symptoms; (ii) difficulty processing and communicating emotions exacerbating interpersonal sensitivity and social withdrawal; and (iii) impoverished imaginative activity contributing to thought constriction and disorganized experience. We also present the alternative framing that alexithymia and LPF deficits may function as overlapping correlates of a more fundamental self-processing disturbance. Finally, we have added a clarifying statement that the present study examines cross-sectional associations and does not formally test mediation. We believe these revisions provide a theoretically principled and empirically grounded account of the pathways through which alexithymia may relate to the constructs under investigation.

Methods

• The inclusion and exclusion criteria are clearly described. However, additional information regarding illness severity, duration of illness, medication status, and current treatment exposure would provide important clinical context for interpreting the findings.

We thank the reviewer for this valuable suggestion. We agree that additional clinical details would enrich interpretability. However, data on illness severity, duration, age at onset, and standardized medication exposure were not systematically recorded in this observational study: such information is frequently incomplete in acute intensive care settings, where full anamnestic reconstruction may not be possible at admission and awaits subsequent validation by territorial psychiatric services. We have therefore added text to the manuscript explicitly stating that these data were unavailable for inclusion as covariates. We also supplemented the description with clinically relevant contextual information inferred from treatment setting and routine clinical practice at the facility, as detailed in the revised manuscript.

• Given the relatively small sample size and the large number of predictors entered into the regression models, additional discussion of model complexity and potential risks of overfitting would strengthen the methodological transparency.

We thank the reviewer for this valuable observation and for highlighting the importance of transparency regarding model complexity. As recommended, we have added a statement in the Statistical Analysis section explicitly acknowledging that the models are relatively complex for the available sample size, noting that bootstrap estimation was employed to improve parameter stability, and advising that findings should therefore be interpreted cautiously. We trust this addition clarifies the methodological constraints and strengthens transparency regarding the interpretability of our results.

• The procedures used to support participants with reading difficulties or cognitive challenges should be described in greater detail. Clarification regarding how assessor assistance was standardized would help readers evaluate the potential influence on self-report responses.

We thank the reviewer for this helpful clarification. We have expanded the description of assessment support to explicitly state how assistance was standardized. As now detailed in the text, all assessor help was intentionally limited to reading items aloud, clarifying lexical terms, and recording responses verbatim; no item interpretation, rephrasing, or response guidance was provided, to minimize potential influence on participants’ self-report.

• Although the use of bootstrap confidence intervals is commendable, the manuscript would benefit from reporting the number of bootstrap replications used and providing additional justification for the sign-flip criterion.

As recommended, we have now specified the number of bootstrap replications (5,000) in the Statistical Analysis section and expanded the justification for the sign-flip criterion, clarifying that it serves as a conservative robustness indicator to retain only effects whose direction is consistent across the vast majority of resamples. This principle is grounded in the literature on bootstrap-based assessment of coefficient stability. These additions enhance methodological transparency and clarify the rationale underlying our robustness criteria.

• The decision to aggregate cognitive emotion regulation strategies into adaptive and maladaptive composite scores should be justified more thoroughly, as important information from individual emotion regulation strategies may have been lost.

We thank the reviewer for this thoughtful observation and for highlighting the potential value of examining individual emotion regulation strategies. We fully acknowledge that aggregating strategies into composite scores may obscure strategy-specific associations. However, this aggregation was a deliberate methodological choice made to reduce model complexity, minimize multicollinearity among conceptually related strategies, and maintain statistical stability given the available sample size and the broader set of predictors entered into the models. By creating adaptive and maladaptive composites, we obtained a parsimonious summary of general regulatory styles that preserves the core distinction between functionally distinct approaches to emotion regulation, while avoiding further inflation of predictor-to-case ratios and correlated regressors that would have compromised model stability and interpretability. In short, this approach prioritizes reliable estimation of broader associations — a necessary trade-off given the study’s transdiagnostic scope and sample constraints — and we acknowledge that strategy-specific analyses represent an important direction for future research in larger samples.

Results

• The presentation of the regression analyses is thorough. However, the Results section would benefit from greater emphasis on effect sizes and clinical significance rather than statistical significance alone.

We have thoroughly revised the Results section to prioritize effect sizes and clinical significance over statistical significance alone. Every β coefficient is now paired with magnitude labels and concrete clinical interpretation, translating standardized effects into tangible differences in symptom severity. R² and ΔR² values are framed as clinical impact, describing the proportion of symptom severity captured by each set of predictors, rather than reported as bare statistics, and "significant/not significant" phrasing has been replaced with magnitude-focused descriptions of effect strength and attenuation across models. All original data and conclusions were preserved; the revisions simply shifted the narrative emphasis to the magnitude and clinical relevance of the findings.

• The correlation matrix demonstrates substantial overlap among personality functioning, alexithymia, and maladaptive traits. Consider reporting additional indicators of shared variance to assist readers in understanding the degree of construct overlap.

To quantify the degree of construct overlap, we have added Supplementary Table A1, which reports squared correlations representing the proportion of shared variance among key predictors. Additionally, we have supplemented the Statistical Analysis section with Variance Inflation Factors (VIF) values computed across all regression models. These analyses confirmed that while substantial shared variance exists among core constructs, all VIF values were well below commonly accepted thresholds for problematic collinearity. Together, these additions provide readers with a comprehensive, transparent account of construct overlap and its potential impact on regression estimates.

• Several findings differ between the standard and hierarchical regression models. A concise summary table highlighting predictors that remained significant after adjustment would improve clarity and facilitate interpretation.

We thank the reviewer for this excellent suggestion. We fully agree that contrasting results across modeling approaches is central to interpreting the findings, and we have added Table 3 to provide a concise, side-by-side comparison of predictors across both models. This table clearly distinguishes associations that remained robust after full adjustment from those that were explained by demographic and emotion regulation variables.

• The finding that alexithymia lost significance after personality variables were introduced is potentially important. Greater attention to the magnitude of these changes would strengthen the interpretation.

As the reviewer correctly noted, the attenuation of alexithymia’s effect is a theoretically important finding. We have now quantified the magnitude of these changes across all symptom domains, revealing reductions in predictive strength and a near-total loss of statistical significance once personality variables are included. These results are now explicitly presented in the text, addressed in the discussion, and demonstrated that alexithymia’s robust initial associations with symptomatology are substantially accounted for by its overlap with personality functioning, suggesting that alexithymia may operate largely as a manifestation of deeper structural deficits in self and interpersonal functioning, rather than as an entirely independent vulnerability factor.

• Some sections of the Results move into explanatory interpretation. Consider limiting this section to description of findings and reserving mechanistic explanations for the Discussion.

Thank you for this helpful comment. We have thoroughly revised the Results section to ensure that it remains focused on the description of the empirical findings. Specifically, we edited each paragraph ensuring it contained the relevant statistical finding only, while explanatory or mechanistic interpretations have been removed from this section. Phrases such as “suggesting” and “may reflect” as well as related interpretive statements, have been moved to the Discussion, where the findings are considered in their broader theoretical and clinical context.

Discussion

• The discussion appropriately highlights the potential importance of personality functioning in severe psychopathology. However, some interpretations appear stronger than warranted given the cross-sectional nature of the study and should be framed more cautiously.

We thank the Reviewer for this important comment. We agree that, given the cross-sectional design, the interpretation of the findings should be framed in terms of associations rather than causal or explanatory relationships. We have therefore carefully revised the Discussion and Conclusions to avoid language that could imply causality, mediation, or temporal ordering. We believe these revisions provide a more appropriately cautious interpretation of the findings while preserving their theoretical relevance and consistency with the existing literature.

• The distinction between self-functioning and interpersonal functioning represents one of the manuscript's most interesting contributions. Expanding the discussion of potential clinical applications of these differential findings would enhance the paper's translational value.

We thank the reviewer for this encouraging comment and fully agree that the distinction between self- and interpersonal functioning represents one of the manuscript's key contributions. We have expanded the Conclusions section to elaborate on the specific clinical implications arising from these differential associations. In doing so, however, we have proceeded with the utmost caution, as we are aware that our study is still at a preliminary stage and has not directly tested the effectiveness of these recommendations.  We trust these revisions strengthen the translational value of the findings.

• The discussion of alexithymia is thoughtful, but statements suggesting mediation or indirect effects should be interpreted cautiously because formal mediation analyses were not conducted.

We thank the reviewer for constructive observation. We agree that language implying mediation, causal pathways, or established mechanisms must be tempered given the cross-sectional design and absence of formal mediation testing. Accordingly, we have revised the Discussion throughout to replace terminology suggestive of mediation or causal mechanisms with more conservative phrasing (e.g., changing references to 'unmediated' effects to 'independent' associations, and 'mechanisms' to 'explanatory relationships'). We have also added an explicit statement acknowledging that these interpretations are based on patterns of shared variance rather than formal pathway testing, and that definitive demonstration of indirect effects would require dedicated mediation or longitudinal designs.

• Greater attention should be given to the transdiagnostic composition of the sample. The extent to which the findings may differ across psychotic disorders, mood disorders, and personality disorders deserves further consideration.

We thank the reviewer for this important suggestion and agree that the transdiagnostic composition of our sample merits more explicit discussion, both as a strength and as a limitation. We have revised the manuscript in two key places to address this concern. First, in the opening of the Discussion, we now explicitly acknowledge that the sample comprises individuals with psychotic disorders, mood disorders, and personality disorders, and we clarify how the assessed symptom domains are relevant across all three groups. Second, we have substantially expanded the Limitations section to directly address the possibility that findings may differ across diagnostic subgroups. We now provide concrete examples of how specific associations might vary. For instance, that the link between interpersonal functioning and mistrust could be more pronounced in schizophrenia spectrum disorders, while the association between self-functioning and depressive symptoms might be stronger in mood and personality disorders. We explicitly stated that formal diagnostic stratification would be required to identify such disorder-specific nuances, and we framed this as an important direction for future research. We trust these revisions give appropriate attention to the transdiagnostic composition of the sample while acknowledging the need for future diagnostic stratification.

• The clinical implications section is promising. However, recommendations regarding treatment approaches, including mentalization-based interventions and emotion regulation interventions, should be presented as hypotheses for future investigation rather than direct implications of the current data.

We fully agree with the reviewer's important point. Given the cross-sectional, correlational design of our study, the clinical implications we outline are indeed best framed as hypotheses generated by our findings rather than as evidence-based treatment recommendations. We have revised the Conclusions section throughout to reflect this distinction carefully. Specifically, we have replaced more definitive language (e.g., 'interventions may profit from,' 'may represent a particularly suitable pathway') with phrasing that explicitly frames these proposals as testable hypotheses (e.g., 'our findings generate the hypothesis that,' 'emerge as a promising candidate for future investigation,' 'worthy of empirical evaluation'). We have also added a closing sentence that explicitly states these proposed treatment directions are derived from cross-sectional correlational data and should be regarded as hypotheses to be tested in prospective, longitudinal, and intervention research. We believe this revised framing preserves the translational value of our findings, while appropriately acknowledging the limits of what can be concluded from the current study design.

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript under evaluation addresses one of the most methodologically complex debates in contemporary psychopathology, the integration between the Level of Personality Functioning and pathological personality traits within the dimensional framework of the Hierarchical Taxonomy of Psychopathology. The investigation into how these personality components, in conjunction with alexithymia and cognitive emotion regulation strategies, predict clinical symptom severity in a sample of severely ill inpatient psychiatric patients holds undeniable clinical and ecological value. 

Recent scientific literature demonstrates that dimensional diagnosis aims to solve the problem of systematic comorbidity and heterogeneity within traditional diagnostic categories. However, the conceptual and empirical overlap between Criterion A and B remains under intense scrutiny. Some theoretical perspectives suggest that personality functioning is largely redundant when pathological traits are adequately measured. Conversely, proponents of Criterion A autonomy argue that personality functioning reflects an essential general severity factor that transcends the stylistic manifestations of traits, providing incremental predictive validity.

The present study directly addresses this gap, presenting data collected in a subacute and long-term psychiatric rehabilitation setting. The use of robust statistical analyses, such as bootstrapping with BCa confidence intervals and sign-flip rates below 5%, reflects a commendable concern for parameter stability given a moderate sample size. However, the manuscript presents methodological weaknesses, critical statistical omissions, and substantive theoretical inconsistencies that must be addressed before the work can be considered for publication in Psychiatry International.

1.  The manuscript repeatedly states, including in the abstract, that "the impact of alexithymia on most clinical symptoms appeared to be mediated by self-functioning deficits." Statistical mediation models tested on strictly cross-sectional data do not allow for the robust establishment of directionality or temporal precedence. Furthermore, alexithymia and self-functioning share deep conceptual and empirical overlap. The purported mediation may simply reflect the redundancy of highly correlated constructs (r = .72 in Table 2) rather than a true causal mechanism of effect transmission

2. A detailed analysis of the tabulated data reveals severe statistical anomalies caused by multicollinearity or suppression effects that were omitted from the discussion. In Table 4e, adaptive emotion regulation strategies (Adaptive Str.) emerge as a significant positive predictor (beta = 0.23, p = .026), whereas maladaptive strategies (Maladaptive Str.) appear as a significant negative predictor (beta = -0.20, p = .044). However, if I understood correctly, the zero-order correlation matrix (Table 2) shows the exact opposite: adaptive strategies correlate negatively with dysthymia (r = -.12), and maladaptive strategies correlate positively (r = .16). This flip indicates a mathematical suppression effect resulting from the simultaneous entry of multiple correlated predictors into the same regression model. Leaving this inversion without a conceptual explanation yields a dangerous and paradoxical clinical interpretation, implying that using adaptive coping worsens dysthymia while maladaptive coping serves as a protective factor. The same phenomenon occurs with the Antagonism variable when predicting Mistrust symptoms in Tables 3f and 4f

3. The text grounds much of its introduction and theoretical discussion in the utility of personality functioning for understanding psychosis and clinical manifestations within HiTOPs Thought Disorder spectrum. However, the instrument used to measure symptom outcomes, the SCL-27, is a brief screening scale focused primarily on affective, somatic, and anxiety symptoms. It lacks validated scales to measure core positive psychotic symptoms (such as hallucinations, structured delusions, or thought disorganization). Attempting to equate symptoms measured by the SCL-27 to the full psychotic spectrum weakens the theoretical robustness of the conclusions, requiring a significant softening of the diagnostic generalizations made by the authors

4. In step 3 of the hierarchical regression models, including 14 predictor variables against a final sample of N=93 participants yields a Subject-to-Predictor Variable ratio of approximately 6.6:1. This level is significantly below the conservative benchmark of 10:1 or the ideal 15:1 recommended to ensure parameter estimate stability (https://doi.org/10.1016/j.jclinepi.2014.12.014). Although the bootstrap method mitigates this limitation regarding inference, the risk of overfitting and the resulting lack of generalizability of the final models remain high

5. Relying exclusively on self-report measures in a sample of severely ill psychiatric inpatients also represents a critical methodological bias. Acute psychiatric populations frequently exhibit deficits in clinical insight regarding their own condition and adaptive functioning. This can severely compromise response reliability on complex reflective scales such as the PID-5-BF and the LPFS-BF, which presuppose intact self-observation capacities.

My Recommendations: the authors must revise the manuscript by removing all causal and mediational language to reflect the cross-sectional design, while adding an explicit discussion on the heavy conceptual overlap between alexithymia and Criterion A self-functioning that leads to lost predictive power. Additionally, they must explain the statistical suppression effects (flips) observed in dysthymic and mistrust symptoms both mathematically and clinically, cautioning readers that these inversions are statistical artifacts of collinearity rather than evidence that maladaptive coping is protective. The introduction and discussion should also be rescaled to align the SCL-27 outcome measures accurately with HiTOP, clarifying that positive psychotic symptoms were not measured and that findings are limited to features like interpersonal mistrust and social withdrawal. Finally, the authors need to expand the limitations section to candidly address the low subject-to-variable ratio and its associated risk of overfitting, the reliance on self-report measures in severe inpatients who may lack clinical insight, and the cross-sectional constraints that prevent establishing true temporal causality.

 

Author Response

Reviewer 3

The manuscript under evaluation addresses one of the most methodologically complex debates in contemporary psychopathology, the integration between the Level of Personality Functioning and pathological personality traits within the dimensional framework of the Hierarchical Taxonomy of Psychopathology. The investigation into how these personality components, in conjunction with alexithymia and cognitive emotion regulation strategies, predict clinical symptom severity in a sample of severely ill inpatient psychiatric patients holds undeniable clinical and ecological value. 

Recent scientific literature demonstrates that dimensional diagnosis aims to solve the problem of systematic comorbidity and heterogeneity within traditional diagnostic categories. However, the conceptual and empirical overlap between Criterion A and B remains under intense scrutiny. Some theoretical perspectives suggest that personality functioning is largely redundant when pathological traits are adequately measured. Conversely, proponents of Criterion A autonomy argue that personality functioning reflects an essential general severity factor that transcends the stylistic manifestations of traits, providing incremental predictive validity.

The present study directly addresses this gap, presenting data collected in a subacute and long-term psychiatric rehabilitation setting. The use of robust statistical analyses, such as bootstrapping with BCa confidence intervals and sign-flip rates below 5%, reflects a commendable concern for parameter stability given a moderate sample size. However, the manuscript presents methodological weaknesses, critical statistical omissions, and substantive theoretical inconsistencies that must be addressed before the work can be considered for publication in Psychiatry International.

1. The manuscript repeatedly states, including in the abstract, that "the impact of alexithymia on most clinical symptoms appeared to be mediated by self-functioning deficits." Statistical mediation models tested on strictly cross-sectional data do not allow for the robust establishment of directionality or temporal precedence. Furthermore, alexithymia and self-functioning share deep conceptual and empirical overlap. The purported mediation may simply reflect the redundancy of highly correlated constructs (r = .72 in Table 2) rather than a true causal mechanism of effect transmission

We thank the Reviewer for this important comment. We agree that, given the cross-sectional design, the interpretation of the findings should be framed in terms of associations rather than causal or explanatory relationships. We have therefore carefully revised the Discussion and Conclusions to avoid language that could imply causality, mediation, or temporal ordering. In addition we explicitly noted that alexithymia and self-functioning are conceptually overlapping and that their shared variance likely reduces the unique contribution of alexithymia. We believe these revisions provide a more appropriately cautious interpretation of the findings while preserving their theoretical relevance and consistency with the existing literature.

2. A detailed analysis of the tabulated data reveals severe statistical anomalies caused by multicollinearity or suppression effects that were omitted from the discussion. In Table 4e, adaptive emotion regulation strategies (Adaptive Str.) emerge as a significant positive predictor (beta = 0.23, p = .026), whereas maladaptive strategies (Maladaptive Str.) appear as a significant negative predictor (beta = -0.20, p = .044). However, if I understood correctly, the zero-order correlation matrix (Table 2) shows the exact opposite: adaptive strategies correlate negatively with dysthymia (r = -.12), and maladaptive strategies correlate positively (r = .16). This flip indicates a mathematical suppression effect resulting from the simultaneous entry of multiple correlated predictors into the same regression model. Leaving this inversion without a conceptual explanation yields a dangerous and paradoxical clinical interpretation, implying that using adaptive coping worsens dysthymia while maladaptive coping serves as a protective factor. The same phenomenon occurs with the Antagonism variable when predicting Mistrust symptoms in Tables 3f and 4f

We thank the reviewer for this careful observation. The sign inversions noted in former Table 4e and for Antagonism in former Tables 3f/4f are indeed suppression effects, which represents a well-documented, mathematically legitimate outcome, when correlated predictors share overlapping variance with the outcome, not "severe statistical anomalies." Importantly, multicollinearity is not severe: as reported in Appendix A, Table A1, all VIF values were well below the conservative threshold of 5 (maximum VIF = 4.40), confirming multicollinearity was present but remained within acceptable limits. We have clarified in the revised text that these adjusted partial coefficients represent unique conditional contributions when covariates are held constant, and explicitly warned that they must not be interpreted as evidence that adaptive coping is harmful, maladaptive coping beneficial, or antagonism protective. The zero-order correlations (Table 2) remain the appropriate estimate of each predictor's total bivariate relationship with the outcome, while the regression coefficients capture only the partialled, unique component.

3. The text grounds much of its introduction and theoretical discussion in the utility of personality functioning for understanding psychosis and clinical manifestations within HiTOPs Thought Disorder spectrum. However, the instrument used to measure symptom outcomes, the SCL-27, is a brief screening scale focused primarily on affective, somatic, and anxiety symptoms. It lacks validated scales to measure core positive psychotic symptoms (such as hallucinations, structured delusions, or thought disorganization). Attempting to equate symptoms measured by the SCL-27 to the full psychotic spectrum weakens the theoretical robustness of the conclusions, requiring a significant softening of the diagnostic generalizations made by the authors

We appreciate the reviewer's insightful observation regarding the mismatch between the theoretical framing and the SCL-27's measurement scope. In response, we have thoroughly revised the manuscript to clarify this distinction at multiple levels. First, we have added explicit signposting early in the Introduction to acknowledge that while the theoretical context is anchored in the psychosis literature, the SCL-27 captures broader affective, somatic, and interpersonal distress domains rather than core positive psychotic phenomena specifically. Second, we have reworked the introduction that grounds the transdiagnostic relevance of our work in the core HiTOP premise that maladaptive personality traits are dimensional constructs cutting across traditional diagnostic boundaries. In addition, it explains that the traits under study in psychotic samples (detachment, psychoticism, negative affectivity) reach pathological severity across severe mental illness broadly, not exclusively in psychotic disorders. Third, we have refined the Discussion to characterize the SCL-27’s limitations more precisely, explicitly noting that it lacks validated scales for hallucinations, formal thought disorder, and related positive symptoms, and we have softened phrasing such as "symptoms associated with psychosis" to "symptoms frequently observed in psychosis" to avoid overstating the instrument's scope. Fourth, we have ensured throughout that diagnostic generalizations are appropriately tempered, framing findings as transdiagnostic patterns of association between personality functioning and broader psychological symptom domains rather than claims specific to the full psychotic spectrum. These revisions collectively strengthen the theoretical robustness of our conclusions by clearly delineating what our instrument can and cannot support, while preserving the value of examining personality functioning pathways across severe psychopathology.

4. In step 3 of the hierarchical regression models, including 14 predictor variables against a final sample of N=93 participants yields a Subject-to-Predictor Variable ratio of approximately 6.6:1. This level is significantly below the conservative benchmark of 10:1 or the ideal 15:1 recommended to ensure parameter estimate stability (https://doi.org/10.1016/j.jclinepi.2014.12.014). Although the bootstrap method mitigates this limitation regarding inference, the risk of overfitting and the resulting lack of generalizability of the final models remain high

We sincerely thank the reviewer for this valuable observation and for recommending the reference by Austin & Steyerberg (2015). We have carefully studied the paper and have revised the Limitations section accordingly to address this concern transparently, as detailed below. The reviewer correctly noted that our subject-to-predictor ratio of approximately 6.6:1 falls below the traditional conservative benchmarks of 10:1 to 15:1 commonly recommended in the literature. However, the very paper cited by the reviewer provides important empirical evidence that refines these guidelines specifically for linear regression models estimated via ordinary least squares. Austin & Steyerberg (2015) demonstrated through extensive Monte Carlo simulations that regression coefficients, standard errors, and confidence intervals remain largely unbiased with as few as 2 subjects per variable (SPV), with minimal relative bias (<10%) even at SPV ratios far below the traditional 10:1 threshold. Thus, while our ratio of ~6.6:1 is indeed below conservative historical benchmarks, it lies above the empirically derived minimum threshold for accurate coefficient estimation established in the recommended reference. Importantly, Austin & Steyerberg further clarified that the primary risks associated with lower SPV ratios are not biased coefficient estimates, but rather upward bias in the conventional R² statistic and reduced certainty regarding generalizability to independent samples. We have explicitly acknowledged these limitations in the revised text, noting that effect size estimates and model R² should be interpreted with appropriate caution and that future validation in independent samples is desirable. To further mitigate the risk of unstable estimates, we employed bootstrap resampling for statistical inference and applied a strict dual criterion for identifying robust predictors (statistical significance based on 95% bootstrap confidence intervals plus sign-flip rate <5%). Accordingly, we believe the revised Limitations section now provides a balanced, evidence-based assessment (directly grounded in the reference the reviewer recommended) that candidly acknowledges the suboptimal SPV ratio while distinguishing clearly between quantities that remain reliable (coefficient estimates, standard errors, confidence intervals) and those that require caution (conventional R², generalizability).

5. Relying exclusively on self-report measures in a sample of severely ill psychiatric inpatients also represents a critical methodological bias. Acute psychiatric populations frequently exhibit deficits in clinical insight regarding their own condition and adaptive functioning. This can severely compromise response reliability on complex reflective scales such as the PID-5-BF and the LPFS-BF, which presuppose intact self-observation capacities.

We thank the reviewer for raising this important methodological consideration, which we already acknowledged as highly relevant to this population. We have revised the Limitations section to clarify that impaired clinical insight and reduced self-observational capacity in acutely ill psychiatric patients may constrain the reliability of self-report—especially for reflective measures such as the PID‑5‑BF and LPFS‑BF. Importantly, we also clarify that assessments were only administered after the treating psychiatrist confirmed each participant was clinically and cognitively suitable to complete the evaluation; this pre-assessment screening helps exclude periods of acute instability where self-report would be unreliable. As now stated in the text, prior evidence supports the validity of self-report for most personality and symptom domains even in clinical populations [67], and the transdiagnostic composition of the sample reflects real-world clinical presentation. We fully agree that future studies should complement self-report with clinician-rated measures and alternative assessment methods to better address insight-related biases.

 

My Recommendations: the authors must revise the manuscript by removing all causal and mediational language to reflect the cross-sectional design, while adding an explicit discussion on the heavy conceptual overlap between alexithymia and Criterion A self-functioning that leads to lost predictive power. Additionally, they must explain the statistical suppression effects (flips) observed in dysthymic and mistrust symptoms both mathematically and clinically, cautioning readers that these inversions are statistical artifacts of collinearity rather than evidence that maladaptive coping is protective. The introduction and discussion should also be rescaled to align the SCL-27 outcome measures accurately with HiTOP, clarifying that positive psychotic symptoms were not measured and that findings are limited to features like interpersonal mistrust and social withdrawal. Finally, the authors need to expand the limitations section to candidly address the low subject-to-variable ratio and its associated risk of overfitting, the reliance on self-report measures in severe inpatients who may lack clinical insight, and the cross-sectional constraints that prevent establishing true temporal causality.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript has been substantially improved following the previous round of review. The red-marked modifications effectively strengthen the theoretical rationale, enhance methodological transparency, and temper causal claims appropriately. The following minor revisions are requested to further clarify key findings and limitations.

Missing Precision in P-Value Reporting

• Table B1 and B2 (various rows)
• P-values are sometimes reported as "†" or ".092" or ".047" without consistent decimal precision.
• Verify that all p-values in Tables B1–B2 follow this format consistently. 

"Dysthymia" vs. "Dysthymic"

• Line 264: "SCL-27 covers six symptom domains: depressive, dysthymic, vegetative, agoraphobic, social phobia, and mistrust."

• Line 407: "...self-functioning impairment was the sole robust predictor for Social Phobia and Dysthymicsymptoms..."

Author Response

Missing Precision in P-Value Reporting

Table B1 and B2 (various rows)

P-values are sometimes reported as "†" or ".092" or ".047" without consistent decimal precision.

Verify that all p-values in Tables B1–B2 follow this format consistently.

• We thank the reviewer for this careful observation. We conducted a full audit of all 126 p-values reported across Tables B1 and B2 and standardized them to consistent format with no leading zero, three decimal places, and <.001 for values below .001. Specifically, we identified and corrected the following inconsistencies: Seven p-values in Table B1 (Social Phobia Symptoms block) contained stray leading zeros (e.g., 0.038, 0.084, 0.762). These have been reformatted to match the rest of the tables (.038, .084, .762). No other block in either table contained leading zeros. One p-value in Table B1 (Depressive Symptoms, Self Functioning) was reported as .001 but marked with ***, which our note defines as p < .001. Given t = 3.73, the true value is below .001; we have therefore reported it as .000. We also verified that no significance markers (†, *, **, ***) appear in the p-value column itself. These are consistently placed in the t column as intended, and that all bold formatting (indicating bootstrap significance with sign-flip rate < 5%) was preserved. Both tables now follow a uniform p-value format throughout.

"Dysthymia" vs. "Dysthymic"

Line 264: "SCL-27 covers six symptom domains: depressive, dysthymic, vegetative, agoraphobic, social phobia, and mistrust."

Line 407: "...self-functioning impairment was the sole robust predictor for Social Phobia and Dysthymicsymptoms..."

• We thank the reviewer for noting this. We have standardized terminology throughout the manuscript: the SCL-27 subscale is consistently referred to as the Dysthymic symptom domain (adjective form, matching the scale's official subscale label and all table headers). There is no isolated use of "Dysthymia" (noun) elsewhere in the text. The typo with the missing space has been fixed, and the term now reads 'Dysthymic symptoms' throughout the manuscript, consistent with the adjective form used for all other symptom domains.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors responded to all my previous comments.

Author Response

• Thank you very much.

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript is significantly strengthened and meets the scientific standards required for publication. I recommend acceptance in its present form.

Author Response

• Thank you very much