Introduction
Traumatic life experiences increase psychosis risk (Bell, Foulds, Horwood, Mulder, & Boden, Reference Bell, Foulds, Horwood, Mulder and Boden2019; Varese et al., Reference Varese, Smeets, Drukker, Lieverse, Lataster, Viechtbauer and Bentall2012), and symptoms of post-traumatic stress mediate this relationship (Alameda et al., Reference Alameda, Rodriguez, Carr, Aas, Trotta, Marino and Murray2020; Sideli et al., Reference Sideli, Murray, Schimmenti, Corso, La Barbera, Trotta and Fisher2020; Williams, Bucci, Berry, & Varese, Reference Williams, Bucci, Berry and Varese2018). In addition to the high rates of childhood trauma, people with psychosis also have a high risk of re-victimization that compounds the risk of symptoms of post-traumatic stress (De Vries et al., Reference De Vries, Van Busschbach, Van Der Stouwe, Aleman, Van Dijk, Lysaker and Pijnenborg2019). Comorbid post-traumatic stress disorder (PTSD) is therefore unsurprisingly common among people with psychosis, worsening mental health outcomes (DeTore, Gottlieb, & Mueser, Reference DeTore, Gottlieb and Mueser2021). However, it is unclear which post-traumatic sequelae account for the association between trauma and psychosis. The diagnostic conceptualization of PTSD has been subject to much debate, particularly regarding the extent to which it should focus on ‘core’ symptoms (i.e. intrusions, avoidance, hyperarousal) or those commonly associated with the complex trauma histories typically experienced by people with psychosis, conceptualized as disturbances of self-organization (DSOs; emotional dysregulation, interpersonal difficulties, negative self-concept) (Karatzias et al., Reference Karatzias, Cloitre, Maercker, Kazlauskas, Shevlin, Hyland and Brewin2017; Trauelsen et al., Reference Trauelsen, Bendall, Jansen, Nielsen, Pedersen, Trier and Simonsen2015). The newly classified diagnosis of complex PTSD (cPTSD) includes the core symptoms of PTSD as well as DSOs (Maercker et al., Reference Maercker, Brewin, Bryant, Cloitre, Van Ommeren, Jones and Reed2013).
Whilst the conceptualization of cPTSD in the latest International Classification of Disease (ICD-11; World Health Organisation, 2019) is relatively recent, existing findings on related concepts shed light on the role of DSOs in psychosis. Avoidant attachment, emotion dysregulation and negative self-beliefs have all been shown to predict positive symptoms, demonstrating a potential association between cPTSD and psychosis (Bloomfield et al., Reference Bloomfield, Chang, Woodl, Lyons, Cheng, Bauer-Staeb and Lewis2021; Hardy et al., Reference Hardy, Emsley, Freeman, Bebbington, Garety, Kuipers and Fowler2016; Hardy, O'Driscoll, Steel, Van Der Gaag, & Van Den Berg, Reference Hardy, O'Driscoll, Steel, Van Der Gaag and Van Den Berg2021; Sitko, Bentall, Shevlin, O'Sullivan, & Sellwood, Reference Sitko, Bentall, Shevlin, O'Sullivan and Sellwood2014). Epidemiological studies indicate that a sizeable proportion of trauma survivors experience comorbid psychosis and cPTSD symptoms (Frost, Louison Vang, Karatzias, Hyland, & Shevlin, Reference Frost, Louison Vang, Karatzias, Hyland and Shevlin2019). Preliminary studies suggest cPTSD may be more common than PTSD among people with psychosis, and potentially contribute to maintaining positive and affective symptoms (Panayi et al., Reference Panayi, Berry, Sellwood, Campodonico, Bentall and Varese2022). Considering the impact of cPTSD may therefore prove essential to a comprehensive understanding of post-traumatic sequelae in people with psychosis.
One limitation of existing research on core PTSD and DSO difficulties in people with psychosis is the dominant use of cross-sectional designs (Bloomfield et al., Reference Bloomfield, Chang, Woodl, Lyons, Cheng, Bauer-Staeb and Lewis2021), since retrospective designs cannot establish temporal relationships and are subject to recall bias (Blum et al., Reference Blum, Vakhrusheva, Saperstein, Khan, Chang, Hansen and Kimhy2015; Decker, Rosen, Cooney, Schnurr, & Black, Reference Decker, Rosen, Cooney, Schnurr and Black2021). In turn, existing longitudinal studies are typically aimed at detecting developmental change utilizing widely spread assessment points (Ram & Gerstorf, Reference Ram and Gerstorf2009), which may fail to capture the potential confounding effects of contextual factors (McNeish, Stapleton, & Silverman, Reference McNeish, Stapleton and Silverman2017) and/or dynamic fluctuations in observed variables (Wang, Hamaker, & Bergeman, Reference Wang, Hamaker and Bergeman2012). An examination of temporal dynamics using intensive longitudinal methods can address these limitations and shed light on the potential interrelatedness of cPTSD and psychosis in daily life.
Experience sampling methodology (ESM), a structured diary technique wherein participants are prompted multiple times per day to complete ambulatory assessments, can assess these temporal dynamics (Trull & Ebner-Priemer, Reference Trull and Ebner-Priemer2009). ESM has increasingly been applied to study the flow and impact of post-traumatic symptoms in the daily lives of people with PTSD (Chun, Reference Chun2016; Vachon, Viechtbauer, Rintala, & Myin-Germeys, Reference Vachon, Viechtbauer, Rintala and Myin-Germeys2019), as well as psychosis (Bell et al., Reference Bell, Eisner, Allan, Cartner, Torous, Bucci and Thomas2024). Brand et al. (Reference Brand, Bendall, Hardy, Rossell, Meyer and Thomas2020) investigated the daily impact of PTSD in people with psychosis, and did not find a temporal relationship between core PTSD symptoms and auditory hallucinations. The current study extends prior research by including wider post-traumatic sequelae – namely, DSOs – as well as a wider array of positive psychotic symptoms (i.e. paranoia, voices, and visions) to clarify previous cross-sectional relationships between DSOs and psychosis (Panayi et al., Reference Panayi, Berry, Sellwood, Campodonico, Bentall and Varese2022) and aid the identification of further treatment targets of trauma-focused psychological interventions.
To the best of our knowledge, this study is the first to examine the dynamic effects of cPTSD on psychosis symptoms in people meeting ICD-10 criteria for schizophrenia-spectrum disorders and DSM-5 diagnostic criteria for PTSD. We aimed to examine the temporal association between DSOs and positive psychosis symptoms, accounting for concurrent PTSD symptoms. Based on cross-sectional findings that DSO-like difficulties mediate the relationship between childhood trauma and psychosis (Bloomfield et al., Reference Bloomfield, Chang, Woodl, Lyons, Cheng, Bauer-Staeb and Lewis2021; Sideli et al., Reference Sideli, Murray, Schimmenti, Corso, La Barbera, Trotta and Fisher2020; Williams et al., Reference Williams, Bucci, Berry and Varese2018), it was hypothesized that daily increases in DSOs would predict subsequent exacerbations in psychotic experiences, and that this relationship would persist when controlling for core PTSD symptoms.
Method
Study design
This study used an ESM design (Myin-Germeys et al., Reference Myin-Germeys, Kasanova, Vaessen, Vachon, Kirtley, Viechtbauer and Reininghaus2018) involving the repeated assessment of positive psychosis symptoms (auditory and visual hallucinations; paranoia), PTSD (intrusions, avoidance, hyperarousal) and DSOs (emotional dysregulation, interpersonal difficulties, negative self-concept), using a mobile app that prompted participants to rate their experiences up to 10 times per day over 6 consecutive days.
Participants
Participants (N = 153) were a subsample of the Study of Trauma and Recovery (STAR trial) (Peters et al., Reference Peters, Hardy, Dudley, Varese, Greenwood, Steel and Morrison2022) who consented to additional ESM procedures prior to randomisation. Participants met ICD-10 criteria for schizophrenia-spectrum diagnoses (F20–29) ascertained from the ICD-10 checklist by the research team, following clinical notes review and consultation with the care team, as appropriate, and scored ≥2 (‘moderate’ intensity) on the distress item of at least one psychotic symptom rating scale (see section 2.3.1) to ensure presence of at least one distressing positive symptom. Notably, 31% of the sample were recruited from Early Intervention for Psychosis (EIP) services, where United Kingdom (UK) good practice guidelines stipulate against assigning potentially stigmatizing diagnoses such as schizophrenia to those experiencing a first episode of psychosis, and instead routinely apply ICD-10 F28 (other non-organic psychotic disorder) or F29 (unspecified psychotic disorder) categories. The majority (114[75%]) had at least one other diagnosis, in addition to meeting criteria for psychosis and PTSD, with the most common being depression (58[38%]).
Participants also endorsed at least one traumatic life event on the Trauma and Life Events checklist (Carr, Hardy, & Fornells-Ambrojo, Reference Carr, Hardy and Fornells-Ambrojo2018), and met PTSD criteria on the Clinician-Administered PTSD Scale for DSM-5 (Weathers et al., Reference Weathers, Bovin, Lee, Sloan, Schnurr, Kaloupek and Marx2018). Participants were excluded if they were <18, their psychotic or PTSD symptoms were primarily organic in etiology, had a primary substance misuse diagnosis, required an interpreter to engage with the trial, or (within the previous 3 months) had major medication changes or received trauma-focussed therapies.
Descriptive demographic and clinical variables are presented in Table 1, split by ESM participants and non-participants. ESM participants had significantly higher rates of ICD-10 F28 diagnoses, and non-participants significantly higher rates of ICD-10 F20 diagnoses. However, the groups did not differ significantly in severity of core PTSD, DSOs, paranoia, voices, or visions (test statistics presented in online Supplementary Table S1).
Demographic and clinical variables for participants who accepted (n = 153) and declined (n = 152) ESM
ESM, Experience Sampling Methodology; M, Mean; s.d., Standard Deviation; PTSD, Post-traumatic Stress Disorder; DSO, Disturbances of Self-Organisation; ICD-10, International Classification of Disease 10th Ed.; ITQ, International Trauma Questionnaire (Cloitre et al., Reference Cloitre, Shevlin, Brewin, Bisson, Roberts, Maercker and Hyland2018); OCD, Obsessive Compulsive Disorder; ADHD, Attention Deficit and Hyperactivity Disorder; GPTS, Green et al., Paranoid Thoughts Scale (Freeman et al., Reference Freeman, Loe, Kingdon, Startup, Molodynski, Rosebrock and Bird2021); PSYRATS, Psychotic Symptom Rating Scales (Haddock et al., Reference Haddock, McCarron, Tarrier and Faragher1999); TALE, Trauma and Life Events checklist (Carr et al., Reference Carr, Hardy and Fornells-Ambrojo2018).
a Based on baseline PSYRATS data.
b Based on baseline TALE data.
Measures
Baseline measures
Participants completed the following instruments as part of a larger STAR trial baseline assessment battery (Peters et al., Reference Peters, Hardy, Dudley, Varese, Greenwood, Steel and Morrison2022):
-
Trauma and Life Events Checklist (TALE; (Carr et al., Reference Carr, Hardy and Fornells-Ambrojo2018)): 22-item self-report checklist assessing difficult life experiences. Each event is rated according to its occurrence, whether this occurred repeatedly, and its timing (i.e. whether participants were under age 16, 16 or over, or both in instances of repeated events). The number of traumas endorsed may be summed to indicate the number of different types of traumatic experiences. The TALE demonstrates good test-retest reliability and convergent validity with related trauma measures (33).
-
International Trauma Questionnaire (ITQ; (Cloitre et al., Reference Cloitre, Shevlin, Brewin, Bisson, Roberts, Maercker and Hyland2018)): 18-item self-report scale assessing presence and severity of PTSD and DSO symptoms within the past month. Both subscales comprise 3 symptom clusters, themselves composed of 2 items each, and 3 items capturing the functional impact of symptoms. Items are scored on a 5-point Likert scale from 0–4 (‘Not at all’–‘Extremely’). The ITQ diagnostic algorithm identifies a probable diagnosis of PTSD when a participant scores ≥2 on at least one item in each PTSD cluster, plus ≥2 on at least one functional impairment item associated with these symptoms. The cPTSD threshold includes that of PTSD plus a score of ≥2 on at least one item in each DSO cluster and of ≥2 on at least one functional impairment item associated with these symptoms. PTSD and DSO items were totaled to derive continuous severity scores from 0–24 on each subscale, with higher scores indicating higher severity. These were used to examine correlations between ESM and baseline measures of PTSD and DSOs. Both subscales demonstrate high internal consistency (both α's ≥0.79; (Cloitre et al., Reference Cloitre, Shevlin, Brewin, Bisson, Roberts, Maercker and Hyland2018)).
-
Revised-Green et al., Paranoid Thoughts Scale (R-GPTS; (Freeman et al., Reference Freeman, Loe, Kingdon, Startup, Molodynski, Rosebrock and Bird2021)): 18-item self-report scale assessing paranoid ideation in the past month. Composed of two subscales, ideas of reference (8 items) and persecution (10 items), items are scored on a 5-point Likert scale from 0–4 (‘Not at all’–‘Totally’). Higher scores on both the reference (range 0–32) and persecution (range 0–40) subscale suggest higher intensity. Items on the GPTS-persecution subscale were totaled to examine correlations between ESM and baseline paranoia measures. Both subscales demonstrate high internal consistency (all α's >0.9; (Freeman et al., Reference Freeman, Loe, Kingdon, Startup, Molodynski, Rosebrock and Bird2021)).
-
Psychotic Symptom Rating Scales (PSYRATS; (Haddock, McCarron, Tarrier, & Faragher, Reference Haddock, McCarron, Tarrier and Faragher1999)): structured clinical interview measuring positive symptoms of psychosis over the past month across two subscales. One measures frequency, intensity, impact and phenomenology of auditory hallucinations (11 items) and the other measures preoccupation, conviction, and impact of distressing beliefs (6 items). Items are scored on a 5-point Likert scale from 0–4, with anchors varying to suit each item; higher scores on both the voices (range 0–44) and beliefs (range 0–24) subscale suggest higher intensity. The PSYRATS displays good inter-rater and test-retest reliability (Drake, Haddock, Tarrier, Bentall, & Lewis, Reference Drake, Haddock, Tarrier, Bentall and Lewis2007). Average scores on the frequency and duration items of the auditory hallucinations PSYRATS were used to examine associations with the ESM voices item. The STAR trial added an adapted PSYRATs version capturing the presence of hallucinations in other modalities. Items, anchors, and scoring are identical to the voices subscale, but adapted to refer to non-auditory hallucinations (Tsang, Reference Tsang2023). Average scores on frequency and duration of visual hallucinations were used to examine associations with the ESM visions item
ESM measures
The ESM assessment comprised 29 items assessing affective states, contextual information, psychosis, and cPTSD symptoms. Items were scored on a 7-point Likert scale from 1–7 (‘Not at all’–‘Very much so’). Items were based on previous ESM studies of similar populations (Chun, Reference Chun2016; Kimhy et al., Reference Kimhy, Delespaul, Corcoran, Ahn, Yale and Malaspina2006), amended in collaboration with consultants with lived experience of trauma and psychosis.
As illustrated in Fig. 1, some items (psychosis symptoms; negative self-concept [DSO]) were phrased to capture the current moment (Right now…, i.e. ‘momentary’ items), whereas others (core PTSD symptoms; emotional dysregulation [DSO]; interpersonal difficulties [DSO]) referred to the timeframe between moments (Since the last beep…, i.e. ‘interval’ items). We considered proximal models based on data from a single timepoint (time t) to nevertheless be longitudinal when interval items were included, since these analyses tested whether symptoms experienced since the previous moment (interval item scores) predicted symptoms at the current moment (momentary item scores). To further corroborate the temporal ordering of effects, we also assessed distal models using lagged predictor variables to test whether interval items at timepoint t − 1, and momentary items at t − 2, predicted momentary items at time t.
Temporal ordering of ESM items, adapted from (Palmier-Claus, Haddock, & Varese, Reference Palmier-Claus, Haddock and Varese2019). Boxes indicate momentary items; arrows indicate interval items.
Current psychotic symptoms: Two items (‘Right now I feel suspicious’; ‘Right now I believe that some people want to hurt me deliberately’) were used to compute momentary paranoia scores. The mean of the two items was used, unless one was missing, in which case the remaining single item was used. Single items measuring voices (‘Right now I hear a voice or voices that other people cannot hear’) and visions (‘Right now I see things that other people cannot see’) were used to assess momentary hallucinatory experiences.
Core PTSD and DSO symptoms: An interval PTSD score was derived from the mean of five items across the three core PTSD symptom domains (intrusions; avoidance; hyperarousal). These items were anchored to the same traumatic experience(s) as the baseline STAR assessment (i.e. that which participants identified as affecting them the most recently). One item captured intrusive memories (‘Unwanted memories about the experience popped into my mind’), two measured avoidance (‘I avoided thoughts, feelings and physical sensations that remind me of the experience’; ‘I avoided people, places or situations that reminded me of the experience’) and two measured hyperarousal (‘I felt super alert, watchful or on guard’; ‘I felt jumpy or easily startled’). All five items were interval items (‘Since the last beep…’, capturing phenomena of interest that might have occurred in the time interval between the previous and current ESM time point) due to the possible insufficient occurrence of specific PTSD symptoms to enable sampling using momentary items (‘Right now..’).
Six items across the three DSO domains (emotional dysregulation; interpersonal difficulties; negative self-concept) were used to derive an interval DSO score. Two items measured emotional dysregulation (‘I found it hard to control my emotions’; ‘I felt spaced out, numb or emotionally shut down’) and another two measured interpersonal difficulties (‘I felt distant or cut off from people’; ‘I found it easy to stay emotionally close to people’ [reverse-scored]). These items were phrased as interval items (‘Since the last beep…’), as the reflective nature of these constructs makes momentary evaluation difficult. The negative self-concept items, however, were momentary (‘Right now, I feel ashamed’; ‘Right now, I believe I am a good person’ [reverse-scored]). An average of all six items would therefore have included momentary and interval items, which may not confer a valid estimate. To account for this, DSO scores were calculated using the mean of negative self-concept item scores from the preceding moment (t − 1; the so-called ‘lagged’ score) and of the emotional dysregulation and interpersonal difficulty item scores at time t (i.e. assessing the interval between the preceding and current moment), to capture DSO severity up to the current assessment point.
ESM item validation: Multilevel confirmatory factor analysis (mCFA) was conducted to verify the factor structure of multi-item ESM constructs whilst accounting for the hierarchical structure of the data (Forkmann et al., Reference Forkmann, Spangenberg, Rath, Hallensleben, Hegerl, Kersting and Glaesmer2018). A three-factor solution representing paranoia, PTSD and DSOs demonstrated acceptable fit (CFI = 0.91; RMSEA = 0.05) (Forkmann et al., Reference Forkmann, Spangenberg, Rath, Hallensleben, Hegerl, Kersting and Glaesmer2018) that was better (AIC = 203 555.14; BIC = 204 012.59) than a single-factor solution (AIC = 213 402.71; BIC = 213 825.15) at within- and between-participant levels.
Multilevel regressions were used to test whether cross-sectional scores on validated questionnaires at baseline predicted ESM measures of matching constructs. These confirmed that GPTS-persecution scores predicted daily paranoia scores (b = 0.043, p < 0.001, 95% CI 0.028–0.058)<TS: Please check for ‘95% CI’ here and elsewhere>. Additionally, frequency and duration items on the PSYRATS-voices and ‘other modalities’ subscales predicted daily voice and vision frequency among voice hearers and those who experienced visions, respectively (b = 1.174, p < 0.001, 95% CI 0.931–1.417; b = 1.005, p < 0.001, 95% CI 0.566–1.444). Similarly, ITQ-PTSD subscale scores predicted daily PTSD scores (b = 0.035, p = 0.004, 95% CI 0.011–0.059), and ITQ-DSO subscale scores predicted daily DSO scores (b = 0.025, p = 0.012, 95% CI 0.005–0.044).
Owing to the multilevel nature of ESM data, the restrictions placed by Cronbach's alpha on measures of internal consistency are too strict, and Macdonald's omega (ω) is preferred as a more robust estimate (Eisele, Kasanova, & Houben, Reference Eisele, Kasanova, Houben, Myin-Germeys and Kuppens2022), yet the dual-item nature of paranoia in our ESM questionnaire prevented a valid calculation of ω. Cronbach's alpha was therefore used to conservatively estimate reliability in this sample, which was acceptable (α = 0.79). The internal reliability of ESM core PTSD and DSO measures was good (ω = 0.83 for both scales). Internal consistency estimates were not suited to the measurement of voices or visions, as they were measured by single items.
Procedure
Participants were recruited from five National Health Service (NHS) Mental Health Trusts across the UK (NHS research ethics committee IRAS ID: 275697). After providing informed consent, participants referred to the trial completed an eligibility assessment administered by a trained research assistant (RA). Those eligible for the trial completed the standardized questionnaire and clinical interview measures used in this study (ITQ, PSYRATS, GPTS-R). The full STAR trial assessment battery and procedures are listed elsewhere (Peters et al., Reference Peters, Hardy, Dudley, Varese, Greenwood, Steel and Morrison2022). Following baseline assessment, those who consented to hear about additional studies within the trial were invited to take part in the ESM study, which took place prior to randomization to the trial arms.
The ESM study used a mobile app, m-Path (Mestdagh et al., Reference Mestdagh, Verdonck, Piot, Niemeijer, Tuerlinckx, Kuppens and Dejonckheere2023), to deliver ESM questionnaires up to 10 times a day for six days. Smartphones with preloaded SIM cards were provided for participants without access to either a smartphone or sufficient internet connection. To reduce burden, schedules were adapted to suit the waking hours of participants. Participants with an atypical schedule (e.g. an upcoming holiday or invasive medical procedure) were not onboarded to the study until after the event(s) had passed, to ensure study responses reflected a typical week for participants.
After providing informed consent, participants were supported by a RA to download m-Path onto their personal mobile phone or were provided with a phone for the duration of the study. Participants completed a practice ESM assessment with the RA to clarify understanding of the items and resolve any queries. Next, the RA scheduled up to 10 notifications – as many as the participants' waking hours would allow using 90-min increments – per day. Compliance with the study was monitored throughout, and a monitoring phone call made two days following the practice to ensure the app ran smoothly and items made sense. Participants who missed all notifications in a day were contacted to check for any issues with the app. Following completion of all six days, participants were debriefed, provided feedback on the study procedures, and reimbursed.
Statistical analyses
Linear regression models are unsuitable for ESM data, due to its clustered nature (Carter & Emsley, Reference Carter, Emsley, Palmier-Claus, Haddock and Varese2019). Multilevel regressions extend linear models by allowing for variation within and between participants, thereby accounting for ESM clustering (Garson, Reference Garson2013). This study employed two-level (observations nested within participants) as opposed to three-level models (observations nested within days nested within participants), as the latter has been shown to be suboptimal in the presence of autocorrelation (de Haan-Rietdijk, Kuppens, & Hamaker, Reference de Haan-Rietdijk, Kuppens and Hamaker2016), which temporal networks suggest is typical of psychotic experiences (Contreras, Valiente, Heeren, & Bentall, Reference Contreras, Valiente, Heeren and Bentall2020; Jongeneel et al., Reference Jongeneel, Aalbers, Bell, Fried, Delespaul, Riper and van den Berg2020). Multilevel models were estimated using the XTREG module in STATA 14 (Statacorp, 2015). Multilevel models report standardized regression coefficients (b); their relative sizes were compared using Wald tests of coefficient standard errors (Wald, Reference Wald1943). As a dimensional experience, analyses involving paranoia included all participants (N = 153). Analyses involving hallucinations only included participants who heard voices (n = 125) or saw visions (n = 97), respectively.
Proximal analyses tested whether DSO symptoms since or at the previous moment (i.e. time t − 1 for momentary item scores [negative self-concept] and time t for interval item scores [emotional dysregulation and interpersonal difficulties]) predict current psychotic symptom scores at time t, and whether this association persisted when controlling for PTSD symptoms since the previous moment (i.e. time t interval item scores).
Distal analyses tested whether lagged DSO scores (i.e. time t − 2 for momentary items and time t − 1 for interval items) predict current psychotic experiences at time t, and whether this association persisted when controlling for PTSD symptoms at time t − 1 (interval items).
The sampling scheme was quasi-random (i.e. beeps sent at random points within each consecutive 90-min interval), meaning the interval between beeps varied. The maximum interval between beeps during which DSO and PTSD items were measured spanned 3 hours in proximal analyses (e.g. if time t was at the end of the 90 min block, and time t − 1 was at the beginning of the previous block), and 6 hours in distal analyses. The average intervals, however, were 90 and 180 min, respectively. Non-consecutive scores were used where immediately preceding observations were missing.
Results
Descriptive statistics
Of a total 8174 data points, 3.326 (40.69%) were missing. On average, participants were scheduled 54 beeps (maximum = 60) over 6 days and responded to 59.26% of these (M = 32; s.d. = 15). Participants who responded to any notifications were included in the analysis, since multilevel models are robust against unbalanced data, though 2.83% of questionnaires were responded to outside a predetermined 15-min window and were excluded from analyses (Delespaul, Reference Delespaul1995). A multilevel model including consecutive beeps as a predictor of missingness suggested measurements were increasingly more likely to be missed as the assessment period progressed (OR = 1.01, p < 0.001, 95% CI 1.00–1.01).
An experience was considered present if a participant rated ≥2, since 1 was anchored to ‘not at all’. Any paranoia item was endorsed at 4295 (93.03%) timepoints; voices and visions at 3077 (66.65%) and 2580 (55.88%) timepoints, respectively. Any core PTSD symptom was endorsed at 4397 (95.24%) timepoints, and any DSO at 4600 (99.63%) timepoints. Within-participant means and standard deviations of ESM variables are listed in Table 2.
Means (M) and standard deviations (s.d.) of within-participant means and within-participant standard deviations for all ESM variables
Note: All variables were measured using a 7-point Likert scale (range 1–7). M (s.d.) refers to means of each item across participants, and s.d.s of those mean scores between participants. MSD (s.d.) refers to the mean s.d. of each participant's own mean, and the s.d. between participants.
Proximal analyses
Multilevel linear regression models indicated that DSO symptoms since or at the previous moment significantly predicted paranoia, voices, and visions at the current moment. This effect of DSOs persisted for all three outcomes when including core PTSD symptoms since the previous moment, themselves also predicting all three outcomes. Wald tests suggested DSOs were a significantly better predictor of paranoia, but not voices or visions, than core PTSD symptoms. Regression and Wald test statistics are listed in Table 3.
Regression statistics of multilevel models
*Significant at p < 0.05 level; **at p < 0.01 level; ***at p < 0.001 level.
Distal analyses
Multilevel linear regression models assessing whether lagged DSO scores (described above) predicted positive psychosis symptoms at the current moment, and whether this association persisted when controlling for lagged PTSD symptoms, showed the same pattern of findings as the proximal analyses. Wald tests suggested that DSOs were a significantly stronger predictor of paranoia, but not voices or visions, than core PTSD symptoms. Regression and Wald test statistics are listed in Table 3.
Discussion
We aimed to ascertain the relationship between cPTSD and psychosis symptoms in the flow of daily life in psychosis individuals who meet criteria for comorbid PTSD. Consistent with our hypotheses, the key findings indicate that both proximal and distal fluctuations in DSOs predicted momentary measures of paranoia, voices, and visions. These temporal associations persisted when controlling for fluctuations in core PTSD symptoms within the same timeframe. Furthermore, they were stronger than those between core PTSD and psychosis symptoms, which were themselves significant, particularly for paranoia.
To our knowledge, this is the first study demonstrating the impact of cPTSD in the daily life of people with psychosis. Our findings imply that cPTSD may play a key role in maintaining psychotic experiences in daily life. In particular, emotional and interpersonal difficulties, and negative self-concept, seem to have a greater influence on psychosis symptoms than traumatic memory intrusions, hypervigilance, and avoidance. These results align with cross-sectional research demonstrating the impact of complex trauma, PTSD, and prevalence of cPTSD in people with psychosis (Campodonico, Varese, & Berry, Reference Campodonico, Varese and Berry2022; Panayi et al., Reference Panayi, Berry, Sellwood, Campodonico, Bentall and Varese2022). They further extend those of previous ESM studies demonstrating the daily impact of childhood trauma and core PTSD in people with psychosis (Brand et al., Reference Brand, Bendall, Hardy, Rossell, Meyer and Thomas2020; Dokuz, Kani, Uysal, & Kuşcu, Reference Dokuz, Kani, Uysal and Kuşcu2022) to include DSOs. In turn, our findings also align with prior ESM studies investigating the impact of constructs consistent with DSOs; for instance, momentary attachment insecurity (consistent with interpersonal difficulties) and emotional instability (consistent with emotional dysregulation) have been shown to increase subsequent paranoia (Nittel et al., Reference Nittel, Lincoln, Lamster, Leube, Rief, Kircher and Mehl2018; Sitko, Varese, Sellwood, Hammond, & Bentall, Reference Sitko, Varese, Sellwood, Hammond and Bentall2016). The distal analyses carried out in this study confirmed that the short-term effect of DSOs on psychosis revealed by proximal analyses persists, albeit with smaller effect size, for a period of up to 6 hours. Significant relationships over this period suggest symptoms of cPTSD potentially earlier on in a day may affect subsequent psychotic experiences (especially for paranoia, where our confidence intervals are more robust).
The findings are aligned with an affective pathway to psychosis and multifactorial accounts of trauma in psychosis (Hardy, Reference Hardy2017; Morrison, Frame, & Larkin, Reference Morrison, Frame and Larkin2003). Particularly, they highlight how negative self-beliefs, relationship difficulties, and emotion regulation may be important treatment targets alongside core symptoms of PTSD, since the effect of DSOs was comparatively larger than that of core PTSD symptoms for all positive psychosis symptoms (significantly so for paranoia). Novel statistical innovations may be used to extend the present findings by exploring the symptom overlap between cPTSD and psychosis, and by identifying directed paths between these clusters to support the development of trauma-focused psychosis interventions (e.g. using network analysis; Contreras et al., Reference Contreras, Valiente, Heeren and Bentall2020).
There were a number of limitations in this study. Interval ESM items are standard in ESM questionnaire design (Eisele et al., Reference Eisele, Kasanova, Houben, Myin-Germeys and Kuppens2022), but their use to make longitudinal assumptions could be questioned, since retrospective items could be influenced by current states and/or recall bias. However, DSOs by nature are psychologically reflective, such as emotional dysregulation (‘Since the last beep, I found it difficult to control my emotions’) and may occur infrequently enough that momentary items risk experiences being missed, such as interpersonal difficulties not arising due to a participant being alone (Eisele et al., Reference Eisele, Kasanova, Houben, Myin-Germeys and Kuppens2022). Furthermore, the distal analyses findings mirrored those of the proximal analyses, providing strong evidence of a temporal relationship between DSOs and psychotic symptoms.
While a strength of this study was the inclusion of both visions and voices, the reliability of single items to measure them is questionable (Eisele et al., Reference Eisele, Kasanova, Houben, Myin-Germeys and Kuppens2022). Hallucinations are multidimensional experiences, with disparate temporal dynamics for differing dimensions (Bless et al., Reference Bless, Hjelmervik, Torsheim, Gudmundsen, Larøi, Holma and Hugdahl2020). Further, visual and auditory hallucinations can be difficult to distinguish from dissociative post-traumatic flashbacks (Wearne et al., Reference Wearne, Ayalde, Curtis, Gopisetty, Banerjee, Melvill-Smith and Waters2022). Future studies should assess how different aspects of psychotic experiences (e.g. content, appraisal, impact/distress) may be affected by DSOs, and capture these experiences using multiple items to maximize precision.
ESM completion rates in this study may appear low at an average of 59%, indicating potential limitations with the sampling scheme. However, 6 days are considered a standard measurement window in ESM studies and participant-level variables more typically predict non-response in ESM than design characteristics (Rintala, Wampers, Myin-Germeys, & Viechtbauer, Reference Rintala, Wampers, Myin-Germeys and Viechtbauer2020; Vachon et al., Reference Vachon, Viechtbauer, Rintala and Myin-Germeys2019; van Berkel et al., Reference van Berkel, Goncalves, Hosio, Sarsenbayeva, Velloso and Kostakos2020). Indeed, owing to the complexities of research engagement among people with psychosis, data attrition is more common than in control samples. The completion rates in our sample are typical of other ESM datasets in people with psychosis (Bell et al., Reference Bell, Eisner, Allan, Cartner, Torous, Bucci and Thomas2024).
The majority of our sample (79%) met criteria for cPTSD on the ITQ (Cloitre et al., Reference Cloitre, Shevlin, Brewin, Bisson, Roberts, Maercker and Hyland2018). Our findings demonstrating a stronger relationship between DSOs and psychosis symptoms than those with core PTSD symptoms may have been a result of the preponderance of people with cPTSD relative to PTSD. Nevertheless, this imbalance is consistent with other trauma-exposed samples of people with psychosis (Panayi et al., Reference Panayi, Berry, Sellwood, Campodonico, Bentall and Varese2022), and is typical of individuals presenting to mental health services, who overwhelmingly have complex trauma histories (Trauelsen et al., Reference Trauelsen, Bendall, Jansen, Nielsen, Pedersen, Trier and Simonsen2015). Additionally, 30% of our sample presented with personality disorder diagnoses alongside higher rates of ‘other non-organic psychosis’ than ESM non-participants. Since there were no significant differences in symptom severity between groups, our sample is likely representative of people with comorbid psychosis and PTSD symptoms. This includes those in EIP services, who show similar rates of clinically significant personality disorder traits (Archer, Shnyien, Mansfield, & Draycott, Reference Archer, Shnyien, Mansfield and Draycott2023), making our findings relevant to psychological interventions offered for this population.
Limitations notwithstanding, there are several implications of these findings. The potentially maintaining role of DSO symptoms supports current trauma-focused therapy practices involving stabilization to establish emotion regulation and build positive relationships, such as trauma-focused cognitive-behavioral and eye-movement desensitization and reprocessing therapies (Keen, Hunter, & Peters, Reference Keen, Hunter and Peters2017; Peters et al., Reference Peters, Hardy, Dudley, Varese, Greenwood, Steel and Morrison2022). Therapeutic approaches aimed at addressing intra- and interpersonal wellbeing, such as compassion-focused techniques (Millard, Wan, Smith, & Wittkowski, Reference Millard, Wan, Smith and Wittkowski2023), may also be a promising avenue for the ongoing development of trauma-focused psychosis interventions, particularly for those who may find trauma reprocessing intolerable (Lewis, Roberts, Gibson, & Bisson, Reference Lewis, Roberts, Gibson and Bisson2020). Our findings also suggest that the assessment of DSOs may be highly relevant for subgroups of people with psychosis and complex trauma histories. This assessment should be sensitive to the potential difficulties associated with cPTSD, including establishing social support and emotional regulation to manage distress associated with trauma disclosure (UK Psychological Trauma Society, Reference McFetridge, Swan, Heke, Karatzias, Greenberg, Kitchiner and Morley2017). Lastly, the large proportion of participants in our sample endorsing visual hallucinations suggests the assessment of hallucinations in modalities other than auditory could be valuable.
To conclude, this study highlights the profound impact of cPTSD on people with psychosis. Specifically, DSO symptoms (i.e. emotional dysregulation, negative self-concept, and interpersonal difficulties) may maintain psychosis symptoms in the flow of daily life to an even greater extent than core PTSD symptoms. In turn, the exacerbation of distressing psychosis symptoms is in addition to the direct impact of these difficulties on individuals' quality of life and daily functioning. Future research is required to delineate relationships between specific symptoms of psychosis, PTSD and DSOs, as well as mechanisms by which this impact occurs. There are clear clinical implications to this research, namely confirming the need for incorporating therapeutic practices aimed at addressing DSOs in trauma-focused psychosis interventions.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0033291724001934.
Acknowledgements
The authors would like to thank the ESRC and NIHR for funding this study. Thank you also to all participants of the STAR trial for their participation. For the purposes of Open Access, the authors have applied a Creative Commons Attribution (CC BY) license to any Accepted Author Manuscript version arising from this submission.
Funding statement
This work was supported by the Economic and Social Research Council (2488414 to PP) and the National Institute of Health Research (NIHR) (NIHR128623 and NIHR130971). AH was funded in part by the NIHR Maudsley Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King's College London.
Collaborators
Affiliated with South London and Maudsley NHS Foundation Trust
Lucy Ring
Christopher Shoulder
Morwennna Rickard
Firat Denizcaglar
Kavita Bains
Freya Meyer
Leilia Jameel
Majella Byrne
Nadine Keen
Sarah Swan
Rachel Frost
Ze Freeman
Gary Ngai
Jess Bird
Constance Harvey
Edward Bickers
Kaisha Santosa
Samantha Mansell
Siamara Goddard
Molly Bird
Harriet Mellotte
Rhea Kohli
Katja Shultz
Antonia Obele
Elizabeth Leahy
Max Dormitontov
Faiza Abbaz
Ho Chak (Milton) Suen
Affiliated with King's College London
Margaret Heslin
Emma Tassie
Sarah Byford
Elizabeth Kuipers
Laura Potts
Inez Verdaasdonk
Affiliated with Greater Manchester Mental Health NHS Foundation Trust
Tony Morrison
Wendy Jones
Eleanor Longden
Aqsa Choudary
Marina Sandys
Kim Towey-Swift
Elizabeth Murphy
Vicky Brooks
Samantha Bowe
Alice Newton-Braithwaite
Elliot Brewer
Leah Orme
Affiliated with University of Manchester
Tony Morrison
Affiliated with Cumbria, Northumberland, Tyne, and Wear NHS Foundation Trust
Rebecca Miskin
Laura McCartney
Marsha Cochrane
Antonia Newman
Sarah White
Nina Cioroboiu
Louise Prentice
Jane Mitchell
Doug Turkington
Kevin Meares
Libby Oakes
Affiliated with Oxford Health NHS Foundation Trust
Amy Langman
Emily Smyth
Georgina Kirtland
Natalie Barnes
Kate Costello
Collins Larrys
Iulia Grabovschi
Charlotte Sagnay de la Bastida
Affiliated with Sussex Partnership NHS Foundation Trust
Kathryn Greenwood
Cat Sacadura
Sarah Mansfield
Ann Steele
Lauren Edwards
Dan Elton
Angie Culham
Nick Grey
Molly Heeger
Lucy Fisher
Krutika Sharma
Chris Scane
Lisa Wood
Joseph Sherborne
Lauren Mose
Guy Emery
Aparajita Pandey
Affiliated with University of Sussex
Kathryn Greenwood
David Fowler
Competing interests
E. P., A. H., F. V., C. S., and R. U. provide psychological therapies for individuals with psychosis and/or PTSD in NHS settings, and E. P. is the Director of a psychological therapies specialist service for psychosis (PICuP). C. S. has written manuals for psychological therapies for psychosis and psychological formulation for which they receive book royalties (from APPI; Guildford Press; Wiley; Routledge; New Harbinger). E. P., A. H., F. V., and C. S. are employed to provide training and/or receive fees (or generate fees for their clinics or research units) for workshops and presentations on psychological therapies for psychosis and/or PTSD; E. P., A. H., F. V., C. S., and R. E. hold or have held grants to carry out trials of psychological therapy for individuals with psychosis. R. E. is a member of the NIHR HTA Clinical Evaluation and Trials Committee.
Open access
