The fifth edition of the DSM-5 defined trauma as ‘an exposure to actual or threatened death, serious injury or sexual violence’. 1 The experience of trauma has been a known risk factor for the development of psychotic disorders. Traumatic experiences can subject individuals to a substantial amount of stress, which can predispose individuals to maladaptive cognition and perception that are predictive of psychosis. Reference Redman, Corcoran, Kimhy and Malaspina2,Reference Spauwen, Krabbendam, Lieb, Wittchen and Os3
A significant proportion of youths with psychotic symptoms have reported experiencing a history of trauma. Reference Stanton, Denietolis, Goodwin and Dvir4 Additionally, the lifetime prevalence of post-traumatic stress disorder (PTSD) among individuals with a psychotic disorder is three times higher than that observed in the general population. Reference Hardy and Mueser5 As early adulthood is the common onset for schizophrenia-related disorders, the prodromal stage can span across adolescent years. Reference Häfner6 Adolescence and young adulthood mark significant milestones for self, social, cognitive and biological development. Reference Stott, Allison, Felter and Beames7 During this developmental period, exposure to trauma can have a detrimental effect on the psychosocial mechanisms that influence the transition from psychotic-like experiences (PLEs) to psychotic disorders. Reference Chan8 Identifying the factors that mediate the relationship between trauma and psychosis could therefore be particularly relevant to building resilience in youths.
Trauma can compromise a person’s sense of agency (SoA) by instilling helplessness and a lack of control over adversities. Reference Ataria9,Reference Killian, Van der Riet, Hough, O’Neill and Zondi10 SoA is the perceived control over one’s actions and consequences. Reference Synofzik, Vosgerau and Newen11 According to the model of SoA, agency is established by tracking sensorimotor contingencies (i.e. feelings of agency) and causal attribution (i.e. judgements of agency). Reference Synofzik, Vosgerau and Newen11 It is a fundamental part of consciousness that distinguishes between self-initiated and other-generated outcomes. Reference van der Weiden, Prikken and van Haren12 Low SoA or perceived control as a result of trauma can encourage maladaptive coping mechanisms associated with cognitive and affective vulnerabilities to develop psychotic symptoms. Reference Bak, Krabbendam, Janssen, de Graaf, Vollebergh and van Os13 For example, individuals may suppress perceptual sensations to escape emotional and physical pain that they perceive as uncontrollable. Reference van Huijstee and Vermetten14 When attenuated sensorimotor monitoring is coupled with external attribution, self-generated actions could be undermined. Reference van der Weiden, Prikken and van Haren12 The loss of agency can lead to feelings of low self-efficacy, low self-esteem, shame, loss of status and grief. Reference Kays Ebrahim, Fouche and Walker-Williams15 Individuals who believe their situations are unchangeable would subsequently be less motivated to engage in interventions or activities that challenge their dysfunctional beliefs. Reference Chang, Kwong, Hui, Chan, Lee and Chen16,Reference Pietluch17
SoA disturbance has been implicated in the psychopathology of multiple mental health conditions, including psychotic disorders. Reference Moore18 Aberrant hierarchical predictive processes can direct an inappropriate amount of attention to irrelevant details in the early stages of psychosis, which increases the chances of misattributing sensory outcomes to external origins. Reference Krugwasser, Stern, Faivre, Harel and Salomon19 Indeed, both schizophrenic and putative prodromal patients have evinced lower accuracy in distinguishing self- or other-generated stimuli. Reference Rossetti, Mariano, Maravita, Paulesu and Zapparoli20 This can be presented as reduced metacognitive capacities in early stages of psychosis, where awareness of one’s own thinking and performance is attenuated. Reference Krugwasser, Stern, Faivre, Harel and Salomon19,Reference Kozáková, Bakštein, Havlíček, Bečev, Knytl and Zaytseva21,Reference Nelson, Fornito, Harrison, Yücel, Sass and Yung22 When coupled with the impact of trauma on self-cognition, SoA disturbance could be one of the pathways that explain vulnerability to psychotic experiences after trauma.
In addition to SoA disturbance, PTSD symptoms can develop when the stress response system that was activated to temporarily protect an individual from trauma persists. Reference Marshall and Garakani23 PTSD symptoms include intrusive memories, avoidance, negative changes in thinking and mood, and changes in physical and emotional reactions. 1 A heightened stress regulation system can increase sensitivity to future stressors, which can predispose individuals to psychotic disorders through increased activity of the dopaminergic pathways. Reference Giannopoulou, Georgiades, Stefanou, Spandidos and Rizos24,Reference Holtzman, Shapiro, Trotman and Walker25 PTSD also implicates disturbance in the cognitive and somatic domains of the sense of self, which is featured in psychotic pathology. Reference Nelson, Fornito, Harrison, Yücel, Sass and Yung22,Reference Lanius, Terpou and McKinnon26 Existing studies found that PTSD symptoms were consistently associated with affective symptoms in psychosis, but the association was less consistent with schizophrenia-specific symptoms in patients. Reference Berry, Ford, Jellicoe-Jones and Haddock27,Reference Resnick, Bond and Mueser28 It is possible that the impact of PTSD symptoms was relatively weaker on fully manifested psychotic disorders; whether PTSD symptoms contribute to PLEs is yet to be investigated extensively. The effects of post-traumatic alteration on the pathophysiological systems could be particularly detrimental to youths whose emotional and behavioural regulations are still in development. Reference Villalta, Smith, Hickin and Stringaris29 Timely intervention could buffer the impact of trauma on mental health deterioration in youths. Reference Kerbage, Bazzi, El Hage, Corruble and Purper-Ouakil30 This warrants an exploration of the role of PTSD symptoms in the transition from PLEs to diagnosable psychosis.
Given that existing studies imply that traumatic experiences increase the risk of psychosis through disrupting a sense of self, SoA and PTSD symptoms could potentially explain PLE outcomes after trauma. Thus far, the associations between SoA and PLEs have mostly been established using traditional experimental paradigms; the results of the self-reported agency were mixed due to the lack of self-measure tools for a general SoA across situations. Reference Moore18 More recently, the Sense of Agency Scale (SoAS) introduced a two-factor model that measures subjective positive and negative agency. Reference Tapal, Oren, Dar and Eitam31 Positive SoA indexes locus of control, general efficacy and physical efficacy, which are associated with positive mental health outcomes. Reference Tapal, Oren, Dar and Eitam31 Conversely, negative SoA indexes a sense of helplessness and unpredictability rather than an absence of SoA. Reference Tapal, Oren, Dar and Eitam31 Self-reported SoA could provide a valuable indication of subjective perception as well as an interpretation of agency in PLEs.
Altogether, the current study aims to investigate the mediating effects of self-perceived SoA and PTSD symptoms on trauma and PLEs in youths. We hypothesised that, as previous studies suggest, H1) both positive and negative SoA would mediate the positive relationship between traumatic events and PLEs in all sampled community youths. In a subgroup of community youths who experienced trauma, we also expected that H2) both positive and negative SoA, and H3) PTSD symptoms would mediate the positive relationship between traumatic events and PLEs. Given that there may be age and gender differences in the type of trauma and post-traumatic psychopathology experienced, analyses will account for these two variables. Reference Olff32
Method
Participants
Participants aged 12 to 25 were recruited as part of the Hong Kong Youth Epidemiological Study Online (YES Online) project. YES Online provides a public platform (https://www.hkyes.hku.hk/) that generates a customised report that compares the respondent’s survey results to a wider epidemiological sample in Hong Kong. It is an ongoing longitudinal assessment of mental health in youths since May 2022. Youths who completed the surveys are invited to complete the Composite International Diagnostic Interview and cognitive tests as an extension of the baseline assessment. A follow-up online survey will be sent to participants who completed the baseline survey at 3 months, 6 months and 1 year.
At the time of writing, 601 individuals completed the baseline survey online, and 520 of them have completed the extended interviews. The current study analysed cross-sectional data from youths who completed the brief version of the Prodromal Questionnaire (PQ-B), SoAS and screening for potentially traumatic events (life event checklist (LEC)) at baseline. Based on these criteria, the data of 517 youths, recruited from May 2022 to May 2024, are identified and included in the analysis. Among the included youths, a subgroup of 283 youths (54.7%) who endorsed one or more items in the LEC completed an additional screening for PTSD symptoms using the Trauma Screening Questionnaire (TSQ). Written informed consents were received from all participants as well as their parents or guardians for participants aged below 18. Participants received monetary compensation of $500 Hong Kong Dollars for completing the surveys.
Measures
PLEs and related distress
The PQ-B is a 21-item psychosis risk screening tool that assesses the presence of positive symptoms and the level of distress associated with each symptom. Reference Loewy, Pearson, Vinogradov, Bearden and Cannon33 Participants were asked to indicate on a yes–no dichotomous scale the symptoms that they had experienced in the past month. For each endorsed item, they had to rate the level of distress caused by the symptoms on a scale of 1 = strongly disagree to 5 = strongly agree. A total symptom score of 7 and a distress score of 24 were the suggested cut-off points for Chinese help-seeking out-patients. Reference Xu, Zhang, Zheng, Li, Tang and Luo34 Cronbach’s α for symptoms and distress indicated good internal consistency for the full sample (α = 0.879 and α = 0.908) and the subgroup (α = 0.858 and α = 0.887).
Sense of agency
SoAS Reference Tapal, Oren, Dar and Eitam31 was used to index the participants’ perceived control over their minds, bodies and the immediate environment. The 7-point scale (1 = extremely disagree; 7 = extremely agree) includes 13 items that measure a sense of positive (items 1, 4, 8, 9, 12 and 13) and negative agency (items 2, 3, 5, 6, 7, 10 and 11). A higher score indicates a stronger sense of positive or negative agency. Both subscales showed good internal consistency in the full sample (α = 0.811 and 0.828) and the subgroup (α = 0.809 and α = 0.825).
Potential traumatic life events
The 17-item checklist for DSM-5 (LEC-5) Reference Gray, Litz, Hsu and Lombardo35 covers natural disasters, physical assaults, sexual assaults, accidents, life-threatening events, deaths and other significant stressful events. Participants were asked to indicate on a dichotomous yes–no scale all events that they have personally experienced, witnessed or learned about happening to a close friend or family member. The LEC total score represents the number of potentially traumatic life events an individual has been exposed to, but not the frequency of the events.
Symptoms of PTSD
To measure the impact of potentially traumatic events, participants who endorsed one or more items on the LEC would complete the TSQ. Reference Brewin, Rose, Andrews, Green, Tata and McEvedy36 This is a self-report tool for indicating PTSD symptoms in the past week. The TSQ comprises 10 items that cover symptoms of intrusive thoughts or memories, flashbacks, nightmares and hypervigilance. Each endorsed item represents a symptom that appeared at a frequency of at least twice a week. The score ranges from 0 to 10, where a higher score indicates higher symptom severity. Cronbach’s α of TSQ suggested good internal consistency in the subgroup (α = 0.877).
Statistical analysis
Data from 517 youths were analysed after individuals with missing data were excluded. The profiles of the included and excluded groups were compared and are available in Supplementary Table 1 available at https://doi.org/10.1192/bjo.2025.10877. The excluded group had significantly higher PLE and related distress, but lower positive SoA and negative SoA. Specifically, these differences were observed between participants with missing and valid LEC data (Supplementary Table 2). Little’s Missing Completely At Random test was significant, suggesting that the values were not missing at random, X 2 (14, N = 601) = 48.36, p < 0.001. Hence, imputation was not recommended as it might affect the results. Reference Mathur37
The hypothesised mediating variables are positive SoA, negative SoA and PTSD symptoms. The number of traumatic life events, indexed by LEC, is the independent variable and the dependent variables are PLEs and related distress. The PROCESS version 4.3 macro was installed on IBM SPSS Statistics version 27.0 for macOS (https://www.ibm.com/products/spss) to conduct a series of regressions to test the mediation effect. Reference Hayes38 Age and gender were added to the linear regression models conducted between the independent variable, the mediators and the dependent variables to account for their confounding effects.
Model 4 in PROCESS was utilised to assess the effects of the mediators on the relationship between the independent and dependent variables for the full sample and subgroup independently. The model calculates a simple mediation model comprising the direct effects of independent variables to dependent variables independent variables to mediators (path a) and mediators to dependent variables (path b), as well as an indirect effect by multiplying the coefficients of path a and b. Indirect effects are only reported for hypothesised mediators that are significantly associated with the independent variable and outcome. A variable is considered to be a significant mediator if the bias-corrected 95% confidence interval around the indirect effect from 5000 bootstrap resamples excludes zero.
Results
The full sample of 517 youths had a mean age of 20.22 (s.d. = 2.45) and 372 (72.0%) were female. A trauma subgroup of 283 youths (54.73%) who reported one or more traumatic life events on LEC had a mean age of 20.28 (s.d. = 2.53) and 205 (72.4%) were female. Descriptive statistics of the independent variable, outcome variables and mediators from the full sample as well as the trauma subgroup are presented in Table 1. The percentage of youths in the full sample who scored above the PQ-B cut-off point (12.4%) was slightly lower than the 16% in a Spanish community youth study. Reference Fonseca-Pedrero, Gooding, Ortuño-Sierra and Paino39 The percentage above the cut-off in the trauma subgroup, however, almost doubled (31.4%).
Table 1 Descriptive statistics of the full sample and trauma subgroup
LEC, life event checklist; PLEs, psychotic-like experiences; PQ-B, Prodromal Questionnaire-Brief version; SoA, sense of agency; PTSD, post-traumatic stress disorder; TSQ, Trauma Screening Questionnaire.
The study aims to test whether positive and negative SoA in the full sample (H1) and the trauma subgroup (H2), as well as PTSD symptoms in the trauma subgroup (H3), mediate the effect of traumatic life events on PLEs and related distress. Preliminary correlation revealed that all main variables were significantly correlated except for traumatic life events and negative SoA in the full sample (see Supplementary Table 3). There were also no significant correlations between traumatic life events and both SoA subscales in the trauma subgroup (see Supplementary Table 4). A series of regression analyses was used to assess the hypotheses, all controlling for age and gender.
Regression results support the preliminary correlation findings. Traumatic life events negatively predicted positive SoA (B = −0.329, 95% CI (−0.601, −0.057)) but were not significantly associated with negative SoA (B = 0.163, 95% CI (−0.172, 0.528)) in the full sample (Table 2). In Table 3, positive SoA negatively predicted PLEs (B = −0.125, 95% CI (−0.187, −0.063)) and distress (B = −0.529, 95% CI (−0.765, −0.292)). The indirect effect of traumatic life events on PLEs (B = 0.041, 95% CI (0.003, 0.035)) and distress (B = 0.174, 95% CI (0.004, 0.038)) via positive SoA was found to be statistically significant (see Fig. 1). Separately, negative SoA was significantly predictive of PLEs (B = −0.189, 95% CI (−0.238, −0.140)) and distress (B = −0.639, 95% CI (−0.828, −0.450)). These results partially supported Hypothesis 1, whereby positive SoA but not negative SoA mediated PLE outcomes in the full sample.
Fig. 1 The relationship between the number of traumatic life events and PLEs and related distress as mediated by positive SoA in the full sample. Statistics in bold are standardised coefficients. Values in brackets are the standardised coefficients for the number of traumatic life events when the mediator has not been entered (i.e. total effect). LEC, life event checklist; PLE, psychotic-like experience; PQ-B, Prodromal Questionnaire-Brief version; SoA, sense of agency. *p < 0.05, ***p < 0.001.
Table 2 Association between the number of traumatic life events and positive sense of agency (SoA), negative SoA and post-traumatic stress disorder (PTSD) symptoms in the full sample (N = 517) and trauma subgroup (N = 283)
B, unstandardised regression coefficient; β, standardised regression coefficient; LEC, life event checklist.
*p < 0.05, ***p < 0.001.
Table 3 PLEs and related distress as predicted by the number of traumatic life events and positive sense of agency (SoA), negative SoA and post-traumatic stress disorder (PTSD) symptoms in the full sample (N = 517) and trauma subgroup (N = 283)
B, unstandardised regression coefficient; β, standardised regression coefficient; LEC, life event checklist; PLEs, psychotic-like experiences.
*p < 0.05, **p < 0.01, ***p < 0.001.
Table 2 shows that in the trauma subgroup, traumatic life events were not significantly associated with positive SoA (B = −0.245, 95% CI (−0.634, 0.143)) or negative SoA (B = 0.238, 95% CI (−0.229, 0.705)). This contrasted the second hypothesis that SoA would be one of the mediators between traumatic life events and PLEs. In Table 3, positive SoA in the trauma subgroup was predictive of PLEs (B = −0.177, 95% CI (−0.260, −0.093)) and distress (B = −0.751, 95% CI (−1.067, −0.436)). Negative SoA in the trauma subgroup negatively predicted PLEs (B = −0.227, 95% CI (−0.294, −0.161)) and distress (B = −0.884, 95% CI (−1.136, −0.632)). Hypothesis 2 was not supported as positive and negative SoA was not associated with the number of traumatic life events.
Traumatic life events in the trauma subgroup positively predicted PTSD symptoms (B = 0.464, 95% CI (0.254, 0.673)) (Table 2), and PTSD symptoms positively predicted PLEs (B = 0.532, 95% CI (0.386, 0.678)) and distress (B = 2.148, 95% CI (1.596, 2.700)) (Table 3). The findings supported Hypothesis 3 as the confidence intervals of the indirect effect of traumatic life events on PLEs (B = 0.247, 95% CI (0.047, 0.152)) and distress (B = 0.996, 95% CI (0.052, 0.162)) also do not include 0, thus the mediation effect of PTSD symptoms was established (see Fig. 2).
Fig. 2 The relationship between the number of traumatic life events and PLEs and related distress as mediated by PTSD symptoms in the trauma subgroup. Statistics in bold are standardised coefficients. Values in brackets are the standardised coefficients for the number of traumatic life events when the mediator has not been entered (i.e. total effect). LEC, life event checklist; PLE, psychotic-like experience; PQ-B, Prodromal Questionnaire-Brief version; PTSD, post-traumatic stress disorder; TSQ, Trauma Screening Questionnaire. *p < 0.05, **p < 0.01, ***p < 0.001.
Discussion
The current study investigates SoA and PTSD symptoms as mediators of the relationship between potentially traumatic life events and PLEs. Our findings indicated that positive SoA significantly mediates the effect of traumatic life events on PLEs and related distress. However, negative SoA was not significantly associated with traumatic life events in the full sample. In the trauma subgroup, a significant mediation effect was found for PTSD symptoms only; neither positive nor negative SoA was significantly associated with traumatic life events in the trauma subgroup. The results are suggestive of a population-specific mediation effect whereby positive SoA may underlie the relationship between PLEs and the number of traumatic life events, including the absence thereof, in all community youths. Among trauma-exposed youths, the relationship was better explained by PTSD symptoms and not by SoA.
The current findings aligned with previous research, which found altered SoA Reference Hauser, Knoblich, Repp, Lautenschlager, Gallinat and Heinz40,Reference Hauser, Moore, de Millas, Gallinat, Heinz and Haggard41 and traumatic experience Reference Ataria9,Reference Killian, Van der Riet, Hough, O’Neill and Zondi10 to be independently associated with an increased risk of psychosis in the community sample. Our results further suggested that traumatic life events can increase PLEs and related distress by reducing perceived agency in youths to take control over their lives. Notably, only positive SoA was significantly associated with the number of traumatic life events in the full sample. It was postulated that positive SoA captures the feeling of being in control, whereas negative SoA represents the feeling of helplessness. Reference Tapal, Oren, Dar and Eitam31 Sense of control has been consistently linked with trauma and post-traumatic psychogenesis. Reference Bak, Krabbendam, Janssen, de Graaf, Vollebergh and van Os13,Reference Livanou, Baoşğlu, Marks, Silva, Noshirvani and Lovell42,Reference Raz, Shadach and Levy43 While helplessness is also associated with adverse events, findings were comparatively mixed. Reference Palker-Corell and Marcus44,Reference Simmen-Janevska, Brandstätter and Maercker45
Contrary to expectation, negative SoA was negatively associated with PLEs and related distress in this sample. One explanation is that the negative agency reduces exaggerated self-attributions that underlie positive symptoms. Reference Hauser, Knoblich, Repp, Lautenschlager, Gallinat and Heinz40 Some researchers also argue that it could be the maladaptive appraisal of negative SoA rather than the perceived agency itself that increases the risk of psychosis in community samples. Reference Baker, Earle, Medford, Sierra, Towell and David46–Reference Hunter, Phillips, Chalder, Sierra and David48 Nevertheless, the observed deviations from the hypothesis require further investigation into SoA as a multifaceted construct that extends to the interpretation of one’s experience.
Results from the subgroup of youths who all reported at least one traumatic life event revealed PTSD as the sole significant mediator. Neither positive nor negative SoA was associated with the number of traumatic life events and hence were not significant mediators in this subgroup. The different mediators identified between the full sample and the trauma subgroup might suggest that SoA can only account for how the presence and absence of trauma affect prodromal risk. This means SoA could account for the increase in PLE symptoms between youths with and without trauma. Specifically, SoA impairments seem to be consistent across the types of trauma due to the commonality of loathing self-initiated actions in trauma, however adaptive. Reference Webb and Widseth49 The effect of trauma exposure on SoA could therefore be shared by the subgroup and remained unchanged after the first traumatic event, hence the lack of significant mediation.
When examining a subgroup of only youths with trauma, PTSD symptoms might better explain the effect of multiple traumatic events beyond the initial exposure. PTSD symptoms such as dissociation and negative self-attribution could be more sensitive to the variations of trauma experience. Reference Guina, Nahhas, Sutton and Farnsworth50,Reference Hagenaars, Fisch and van Minnen51 Furthermore, researchers proposed that SoA can be distinguished from the sense of body ownership, where more severe trauma asserts greater damage to the latter and leads to greater deterioration. Reference Ataria9,Reference Pyasik, Burin and Pia52 The sense of body ownership could arguably be a better indicator for variations in traumatic experiences, where the length of long-term trauma, sexual trauma and combat trauma were predictors of PTSD severity and distorted subjective bodily experience. Reference Guina, Nahhas, Sutton and Farnsworth50,Reference Laricchiuta, Garofalo and Mazzeschi53,Reference Tsur, Defrin, Lahav and Solomon54 As patients with schizophrenia tend to underreport trauma and PTSD symptoms, these findings have important implications for the screening of trauma history in clinical prevention. Reference Lommen and Restifo55 Together, our findings support the notion that variations in traumatic life events, especially experiencing multiple types of trauma, increase the risk of PTSD and psychotic disorders. Reference Evans, Reid, Preston, Palmier-Claus and Sellwood56 Future research can examine the shared mechanisms of PTSD and psychosis to investigate the post-traumatic transition from PLEs to diagnosable psychotic disorders extensively.
While the findings of this study provide valuable insights into the role of SoA and PTSD symptoms in PLEs, some limitations may influence the interpretation and generalisability of the results. Firstly, online data collection limits the opportunity to clarify survey answers with the respondents, particularly the reporting of traumatic life events, where information may be more likely to be withheld, as demonstrated in the missing data analysis. The evaluation of trauma history could benefit from a more detailed interview as opposed to self-identifying events. In addition, online sampling may be biased towards youths who were motivated and capable of completing the surveys without the assistance of researchers. It can be cognitively taxing for distressed youths to finish a series of self-report surveys, leading to a potentially higher number of missing responses. This is evident in the significantly higher mean scores in PLE and lower SoA measures in the excluded group (see Supplementary Table 1). The exclusion of individuals with missing data in the analysis may therefore fail to represent youths with a higher risk of psychosis and SoA.
Moreover, the participants in this study were mostly female, which may introduce a gender bias in the generalisability of the results. This gender proportion is consistent with what was observed in other research studies that collected online data. Reference Wu, Zhao and Fils-Aime57 For this reason, the objective of this study is only to test the significance of a mediation effect rather than to examine prevalences. We acknowledge the possibility of other variables that could influence the risk of psychosis, such as family history and socioeconomic status. However, introducing more confounders to the study may further limit the generalisability of this online sample. Reference Kahlert, Gribsholt, Gammelager, Dekkers and Luta58 Future investigations may consider testing the replicability of the current findings in an epidemiological or a representative cohort with a secondary analysis of other confounders.
The current findings have clinical implications for preventing post-traumatic psychopathological development. A study found that self-reported control over the future was associated with better post-trauma adjustment, rather than the control over the trauma itself. Reference Frazier, Steward and Mortensen59 Establishing self-agency, such as through reconstruction of trauma and self-narratives, could build resilience against distress. Reference Berán and Unoka60 Moreover, directing focus on the sensations of body movements may help individuals to integrate the sensorimotor representations of trauma with episodic memories. Reference Adrien, Bosc, Peccia Galletto, Diot, Claverie and Reggente61 This could prevent the development of PTSD by facilitating positive affective and body responses to reduce stress activation. Screening for a sense of agency and maladaptive cognition could be useful to identify youths with a higher risk of poor mental health outcomes. Altogether, SoA could be an important construct for future prevention interventions to target.
Aligned with previous research, the current study reveals positive SoA as a mediator of trauma on PLEs in Hong Kong youths. In the trauma subgroup, the mediation effect of SoA diminished, and results indicate PTSD symptoms as the mediator for trauma variation on PLEs. Our findings suggest that subjective SoA could be a multifaceted construct, where a subscale related to reduced locus of control could better account for the impact of trauma on PLEs than a subscale related to feelings of helplessness. The severity of PTSD symptoms could be a better account of trauma variation on psychotic symptoms than SoA in trauma-exposed youths. In conclusion, SoA and PTSD symptoms could be potentially important constructs in the pathway between trauma and psychotic transition.
Supplementary material
The supplementary material is available online at https://doi.org/10.1192/bjo.2025.10877
Data availability
Data available on request from the corresponding author, M.M.S.
Acknowledgements
We would like to thank all research participants and colleagues who contributed to the current study.
Author contributions
All authors have approved the final version of the manuscript. M.M.S., Y.-n.S. and C.L.M.H. conceived the study. M.M.S. analysed the data and wrote the first draft with support from Y.-n.S., S.M.Y.W., S.K.W.C., E.H.M.L., E.Y.H.C. and C.L.M.H. All authors provided critical feedback and revisions.
Funding
This study did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors.
Declaration of interest
None.
Ethical standards
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2013. All procedures involving human subjects/patients were approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (UW 19-017).
Open access
eLetters
No eLetters have been published for this article.