Applying neural network algorithms to ascertain reported experiences of violence in routine mental healthcare records and distributions of reports by diagnosis.

Introduction: Experiences of violence are important risk factors for worse outcome in people with mental health conditions; however, they are not routinely collected be mental health services, so their ascertainment depends on extraction from text fields with natural language processing (NLP) algorithms. Methods: Applying previously developed neural network algorithms to routine mental healthcare records, we sought to describe the distribution of recorded violence victimisation by demographic and diagnostic characteristics. We ascertained recorded violence victimisation from the records of 60,021 patients receiving care from a large south London NHS mental healthcare provider during 2019. Descriptive and regression analyses were conducted to investigate variation by age, sex, ethnic group, and diagnostic category (ICD-10 F chapter sub-headings plus post-traumatic stress disorder (PTSD) as a specific condition). Results: Patients with a mood disorder (adjusted odds ratio 1.63, 1.55-1.72), personality disorder (4.03, 3.65-4.45), schizophrenia spectrum disorder (1.84, 1.74-1.95) or PTSD (2.36, 2.08-2.69) had a significantly increased likelihood of victimisation compared to those with other mental health diagnoses. Additionally, patients from minority ethnic groups (1.10 (1.02-1.20) for Black, 1.40 (1.31-1.49) for Asian compared to White groups) had significantly higher likelihood of recorded violence victimisation. Males were significantly less likely to have reported recorded violence victimisation (0.44, 0.42-0.45) than females. Discussion: We thus demonstrate the successful deployment of machine learning based NLP algorithms to ascertain important entities for outcome prediction in mental healthcare. The observed distributions highlight which sex, ethnicity and diagnostic groups had more records of violence victimisation. Further development of these algorithms could usefully capture broader experiences, such as differentiating more efficiently between witnessed, perpetrated and experienced violence and broader violence experiences like emotional abuse.

Associations between post-traumatic stress disorders and psychotic symptom severity in adult survivors of developmental trauma: a multisite cross-sectional study in the UK and South Korea.

BACKGROUND: Childhood maltreatment is a risk factor for the development of post-traumatic stress disorders and psychosis. However, the association between post-traumatic stress disorder (PTSD), including complex PTSD, and psychotic symptoms is unknown. We investigated whether the presence of PTSD and complex PTSD was associated with psychotic symptom severity within survivors of developmental trauma. METHODS: As part of the Investigating Mechanisms underlying Psychosis Associated with Childhood Trauma (IMPACT) study, from Aug 20, 2020, to Jan 24, 2021, and from Sept 9, 2022, to Feb 21, 2023, using study advertisement on online platforms we recruited adult (≥18 years) participants who had experienced developmental trauma without a psychiatric diagnosis in the UK and South Korea. We measured whether participants met diagnostic thresholds for PTSD and complex PTSD using the self-reported International Trauma Questionnaire, and psychotic symptoms using the self-reported Community Assessment of Psychic Experiences. We used linear regression, adjusting for sociodemographic variables such as age, sex, ethnicity, educational attainment, and socioeconomic status, to examine whether there was an association between PTSD and complex PTSD and psychotic symptoms. The study is registered in the UK (University College London Research Ethics Committee [14317/001] and the National Health Service Research Ethics Committee [22/YH/0096]) and South Korea (Institutional Review Board of Seoul National University Bundang Hospital [B-2011-648-306]), and is ongoing. FINDINGS: Of the 2675 participants who took part in the study, 1273 had experienced developmental trauma and were included in the study in the UK (n=475) and South Korea (n=798), comprising 422 (33%) men and 851 (67%) women with a mean age of 26·9 years (SD 6, range 18-40), mostly of White British (n=328) or South Korean (n=798) ethnicity. We found no significant association between PTSD and psychotic symptom severity (total severity ß=-2·40 [SE 3·28], p=0·47), compared with participants who did not meet PTSD or complex PTSD caseness. We found a significant relationship between complex PTSD and psychotic symptom severity (total severity ß=22·62 [SE 1·65], p<0·0001), including for positive (ß=12·07 [SE 0·99], p<0·0001) and negative symptoms (ß=10·5 [SE 0·95], p<0·0001), compared with participants who did not meet PTSD or complex PTSD caseness. INTERPRETATION: Health systems must assess individuals with previous developmental trauma for complex PTSD and treat those affected. These individuals should also be assessed for psychotic symptoms, and if necessary, preventative measures should be taken to reduce risk of conversion. Further work should assess whether treating complex PTSD modifies the risk of conversion to psychosis. FUNDING: UKRI Future Leaders Fellowship, British Medical Association Margaret Temple Award for Schizophrenia Research, and the National Research Foundation of Korea-Korea Government.

Multiplicity in the experience of voice-hearing: A phenomenological inquiry.

Although it is recognized that voice-hearers often report a large number and variety of voices there have been few investigations of this multiplicity. Understanding the phenomenology of voice-hearing can provide a firm foundation for theorizing about its causes. In this international online survey of voice-hearers, details were elicited of the content of up to five utterances associated with up to five voices from each respondent. The contents were independently rated and associated with characteristics of each voice such as its perceived age, whether it had changed over time, and whether it was of a familiar person. We investigated predictors (e.g., diagnoses, voice gender, age first heard) of utterance negativity, length, and whether voices referred to themselves. The average number of voices reported was approximately four. The majority were perceived as male and had negative content. Child-aged voices were significantly less negative than all other voices except those perceived as being elderly. Multi-level analyses indicated that there was significant variability at the level of different utterances within voices but variability was more prominent at the level of different voices within an individual. The data were inconsistent with general cognitive models for hearing voices such as the misattribution of inner speech and were more congruent with a dissociation model of voice-hearing. Our findings support approaches based on subtype or dimensional methods of classifying voices, and additionally indicate that research and clinical assessment may benefit from more systematic assessment of multiplicity.