The relationship between childhood trauma and schizotypy and the pathways underlying this association Tjasa Velikonja Submitted for the award of PhD in Mental Health Sciences University College London June 2014 ‘I, Tjasa Velikonja confirm that the work presented in this thesis is my own. Where information has been derived from other sources, I confirm that this has been indicated in the thesis.' ii | Abstract There is a growing body of literature demonstrating an association between childhood trauma and schizotypy (e.g. Afifi et al. 2011;Myin-Germeys et al. 2011). However, more research is required to build on methodological limitations of previous studies, explore the possible differential effects of specific trauma types and expand the focus from a single contributor (e.g. psychological, biological) by considering the additive/interactive contributors to schizotypal symptomatology. The aim of the thesis was to explore the relationship between a range of childhood traumatic experiences and schizotypy whilst also incorporating several social, psychological and genetic factors underlying these relationships. Participants were recruited as a part of a cross-sectional case-control study conducted in the London Boroughs of Lambeth and Southwark. The thesis covers a subsample of controls (N=212), healthy volunteers, aged 18-64 and residents in the same geographical area. Data were gathered using an in-depth standardised interview regarding childhood abuse (Childhood Experience of Care and Abuse - CECA) and The Structured Interview for Schizotypy – Revised (SIS-R) measuring a range of schizotypal symptoms and signs. The study found a linear association of total trauma and schizotypy (adj. β=.88, p=0.004), with the strongest associations observed for psychological (adj. OR=4.85, p=0.039) and physical abuse (adj. OR=3.56, p=0.003). These particular types of trauma had an especially robust effect on positive schizotypal traits (psychological: adj. OR=3.79, p=0.013; physical abuse: adj. OR=2.32, p=0.042), which are attenuated forms of positive symptoms of schizophrenia (e.g. hallucinations, delusions). Negative beliefs about self/others and depression were the main mediators of these associations. A strong relationship was found for genetic risk of psychosis and increased schizotypy (adj. β=3.41, p=0.015). Other moderators of the childhood trauma - schizotypy association were intrusive life events (adj. β=4.20, p=0.045). This study provides further insights into the association between childhood trauma and schizotypy and gives clues to pathways underlying this association. iii | List of Contents INTRODUCTION 15 CHAPTER 1 - LITERATURE REVIEW: SCHIZOTYPY 15 1.1 What are schizotypy and schizotypal personality disorder? 17 1.1.1 Definitions of the concepts 18 1.1.2 The multifactorial nature of schizotypy 27 1.1.3 The heterogeneity of schizotypy measures 29 1.1.4 Socio-demographic characteristics and schizotypy 44 1.1.5 Schizotypy and the development of psychosis / The continuum model 46 1.2 Heritability of schizotypy 53 CHAPTER 2 - LITERATURE REVIEW: CHILDHOOD TRAUMA AND SCHIZOTYPY 63 2.1 Childhood trauma and schizotypy association 65 2.1.1 Introduction to the childhood trauma and schizotypy association 67 2.1.2 Empirical literature search strategy 71 2.1.3 Results 73 2.1.4 Discussion 83 2.1.5 Methodological issues 89 2.2 Childhood trauma in relation to psychotic-like symptoms 100 2.2.1 Introduction and search strategy 101 2.2.2 Results 104 2.2.3 Discussion 115 2.3 Possible pathways underlying the childhood trauma – schizotypy association 123 2.3.1 Hypothesised theoretical models 126 2.3.2 The effects of adult traumatic experiences /Life events 133 2.3.3 The role of dissociation and PTSD 139 2.3.4 The effect of cannabis use 145 CHAPTER 3 - METHODOLOGY 155 3.1 Aims and Hypotheses 157 3.1.1 Aims of the thesis 158 3.1.2 Hypotheses 158 3.2 Study design 160 3.2.1 Experimental design 161 3.2.2 Sample size calculation 163 3.3 Sample/Data collection 165 3.4 Main assessment tools 173 3.4.1 Structured Interview for Schizotypy Revised (SIS-R) 176 3.4.2 Community Assessment of Psychic Experiences (CAPE) 190 3.4.3 Childhood Experience of Care and Abuse (CECA) 192 3.4.4 Bullying Questionnaire 203 3.4.5 Life Events and Difficulties Schedule (LEDS) 203 3.4.6 Brief Core Schema Scales (BCSS) 205 3.4.7 Hamilton Rating Scale for Depression (HRSD) 206 3.4.8 Family Interview for Genetic Studies (FIGS) 207 3.4.9 Cannabis Experience Questionnaire (CEQ) 208 CHAPTER 4 - RECRUITED SAMPLE AND DATA MANAGEMENT 210 4.1 Recruited sample 212 4.1.1 Socio-demographic characteristics of the final sample 215 iv | 4.2 Data analyses 218 4.2.1 Childhood trauma and schizotypy association 219 4.2.2 Childhood trauma and life events interaction and schizotypy 226 4.2.3 Childhood trauma and familial risk interaction and schizotypy 229 4.2.4 Childhood trauma and schizotypy and the effects of cannabis 231 4.2.5 Possible underlying mechanisms supporting childhood trauma – schizotypy association/ mediation effects 233 CHAPTER 5 - RESULTS 237 5.1 Is there support for the childhood trauma and schizotypy association? 247 5.2 Life events (and interaction with childhood trauma) and schizotypy association 263 5.3 Familial risk for psychosis (and interaction with childhood trauma) and schizotypy association 267 5.4 Cannabis use (and interaction with childhood trauma) and schizotypy association 272 5.5 Mediation effects / Pathways from childhood trauma to schizotypy 277 CHAPTER 6 - DISCUSSION 291 6.1 Overview of the main findings 293 6.2 Study findings and link to existing literature 300 6.3 Study limitations 325 CHAPTER 7 - FINAL REMARKS 341 7.1 Brief summary of the findings (Integrated model) 343 7.2 Clinical Implications 346 7.3 Future directions 354 REFERENCES 358 APPENDICES 443 Appendix I Scoring guide for the quality assessment of the empirical research papers 445 Appendix II Table Suppl.1: continued summary of studies on childhood adversity in relation to psychosis-like experiences 447 Appendix III Participant Information Sheet and Consent Form 451 Appendix IV Psychosis Screening Questionnaire 456 Appendix V MRS Sociodemographic Schedule 457 Appendix VI Wechsler Adult Intelligence Scale (3rd ed.) (WAIS-III, abbreviated) 465 Appendix VII Structured Interview for Schizotypy-Revised (SIS-R) 468 Appendix VIII Community Assessment of Psychic Experiences (CAPE) 487 Appendix IX Childhood Experience of Care and Abuse (CECA) ‘Interview Version’ and Bullying Questionnaire 490 Appendix X Life Events and Difficulties Schedule (LEDS) 497 Appendix XI Brief Core Schema Scales (BCSS) 501 Appendix XII Hamilton Rating Scale for Depression (HRSD) 502 Appendix XIII Family Interview for Genetic Studies (FIGS) 508 Appendix XIV The Cannabis Experiences Questionnaire (CEQ) 516 Appendix XV Table Suppl.2: Socio-demographics of Southwark and Lambeth Boroughs 523 Appendix XVI Additional Result Tables 524 v | List of Tables Table 1: Measures of general schizotypy 35 Table 2: Measures of attenuated psychotic symptoms 40 Table 3: Summary of studies on childhood trauma and schizotypal traits 74 Table 4: Summary of studies on childhood trauma in relation to psychosis-like experiences (PSE) 105 Table 5: Ethnic distribution for England and London Boroughs of Lambeth and Southwark 163 Table 6: Gender and ethnicity of the recruited sample according to the three different sources of recruitment 168 Table 7: Ethnic groups (and gender) of the total sample (comparison with ethnic distribution in Lambeth and Southwark Boroughs) 213 Table 8: Gender and ethnic distribution of the thesis sample 214 Table 9: Ethnicity comparison of those with completed SIS-R (‘included’) and those without SIS-R (‘excluded’) 215 Table 10: Source of recruitment comparison of those with completed SIS-R (‘included’) and those and those without SIS-R (‘excluded’) 215 Table 11: Complete socio-demographic characteristics of the thesis sample 217 Table 12: Frequencies and gender comparison in reported total trauma and all distinct trauma types 239 Table 13: Frequencies and gender comparison of psychological and sexual trauma (including all levels of severity) 240 Table 14: Frequencies for multi-victimisation and between gender comparison 240 Table 15: Frequencies and comparison between ethnic groups in reported total trauma and all distinct trauma types 240 Table 16: Mean schizotypy scores (SIS-R) and attenuated psychotic symptoms scores (CAPE) - between gender comparison 244 Table 17: Frequencies of schizotypal traits in a total sample and between gender comparisons 245 Table 18: Comparison between ethnic groups on mean schizotypy (SIS-R) and attenuated psychotic symptoms (CAPE) scores 246 Table 19: Associations between total trauma (and all distinct trauma types) and total schizotypy 248 Table 20: Association between separation from a parent and parental death experiences and total schizotypy 249 Table 21: Association between trauma types and top 20% and top 10% of schizotypy scorers 251 Table 22: Association between trauma types and top 20% and top 10% of scorers on positive schizotypy 252 Table 23: Association between trauma types and top 20% and top 10% of scorers on negative/disorganised schizotypy 253 Table 24: Association between number of types of traumatic experiences and total, positive and negative/disorganised schizotypy 255 Table 25: Association between trauma types (different levels of frequency and severity) and total schizotypy 255 Table 26: Association between age of trauma occurrence and total, positive and negative/disorganised schizotypy dimensions 257 Table 27: Gender comparison of association between trauma types and total schizotypy 258 Table 28: Associations between total trauma (and all distinct trauma types) and attenuated/psychotic- like symptoms total (on CAPE measure) 261 Table 29: Associations between total trauma (and all distinct trauma types) and CAPE positive and CAPE negative dimensions 262 Table 30: Frequency of recent life events and between gender comparison 264 Table 31: Frequency of recent life events among individuals with and without childhood trauma 264 Table 32: The interaction effects of the recent life events and difficulties and childhood trauma on total schizotypy score 265 vi | Table 33: The interaction effects of recent life events and difficulties and childhood trauma on positive and negative/disorganised schizotypy 266 Table 34: Frequency of familial risk of psychosis (‘psychosis/narrow’ and ‘any mental illness/broad’ definition) 268 Table 35: The interaction effects of familial risk for psychosis and childhood trauma on total schizotypy 270 Table 36: The interaction effects of familial risk for psychosis and childhood trauma on positive and negative/disorganised schizotypy 270 Table 37: The interaction effects of familial risk for psychosis (including any mental illness of a first degree family member) and childhood trauma types on total schizotypy score 271 Table 38: Frequency of lifetime cannabis use 273 Table 39: Frequency of current cannabis use and lifetime dependency 273 Table 40: Frequency of lifetime cannabis use among those with/without childhood trauma 274 Table 41: Types of cannabis among those with/without childhood trauma 274 Table 42: The interaction effects of cannabis use and childhood trauma on schizotypy 275 Table 43: The interaction effects of the age of cannabis use and childhood trauma on schizotypy 275 Table 44: The association between the frequency of cannabis use and total and positive and negative/disorganised schizotypy 276 Table 45: The comparison between the associations of type of cannabis used and schizotypy 276 Table 46: The frequency of lifetime dissociation symptoms among those with/without childhood trauma 278 Table 47: The frequency of recent dissociation symptoms among those with/without childhood trauma 278 Table 48: The association between recent dissociation/derealisation symptoms and total, positive and negative schizotypy 278 Table 49: Mean scores on negative beliefs about self/others and the depression scale 279 Table 50: Mean scores on negative beliefs about self/others and the depression scale among those with/without childhood trauma 279 Table 51: The effects of the negative beliefs/ depression scores and childhood trauma on schizotypy scores 280 Table 52: The effects of the negative beliefs/ depression scores and childhood trauma on positive and negative/disorganised schizotypy 280 Table 53: Associations between different types of childhood trauma and schizotypy total scores, split into total, direct and indirect effects/pathways via possible mediators 282 Table 54: Main study findings according to each of the hypotheses 296 Table 55: The prevalence of trauma in the present study compared to UK study (NSPCC) 302 vii | List of Figures Figure 1: Summary of phases preceding schizophrenia 27 Figure 2: Flowchart of studies included in the literature review - 1 72 Figure 3: Flowchart of studies included in the literature review - 2 104 Figure 4: Diagrammatic presentation of factors predicting psychosis-proneness 125 Figure 5: Hypothesised sociodevelopmental and neurodevelopmental pathways to psychosis 132 Figure 6: The total sample size needed for increase/decrease in odds ratio (at constant power) – for logistic regression analysis 164 Figure 7: The total sample size needed for different statistical power (if OR=2) – for logistic regression analysis 164 Figure 8: The total sample size needed for different effect sizes – for linear regression analysis 164 Figure 9: Flowchart presentation of the recruitment process 169 Figure 10: Overview of the measures used for analysis for Hypothesis 1, including the variables extracted from the measures and types of analyses used 219 Figure 11: Overview of the measures used for analysis for Hypothesis 2, including the variables extracted from the measures and type of analysis used 227 Figure 12: Overview of the measures used for analysis for Hypothesis 4, including the variables extracted from the measures and type of analysis used 229 Figure 13: Overview of the measures used for analysis for Hypothesis 5, including the variables extracted from the measures and type of analysis used 231 Figure 14: Overview of the measures used for exploratory mediation analyses (Hypotheses 3 & 5), including the variables extracted from the measures and type of analysis used 233 Figure 15: Path decomposition into direct and indirect effects 235 Figure 16: Frequencies of schizotypy total scores (SIS-R) - normal distribution of SIS-R total scores 242 Figure 17: Mediation (in percentage) of the effect of household discord on schizotypy 288 Figure 18: Mediation (in percentage) of the effect of psychological abuse on schizotypy 288 Figure 19: Mediation (in percentage) of the effect of physical abuse on schizotypy 288 Figure 20: Mediation (in percentage) of the effect of sexual abuse on schizotypy 289 Figure 21: Mediation (in percentage) of the effect of bullying on schizotypy 289 Figure 22: Conceptual path diagram showing significant associations between childhood trauma and schizotypy, direct and indirect paths 290 Figure 23: Integrated model - hypothesised diagram showing pathways between childhood trauma and schizotypy according to the present findings; direct and indirect paths 345 viii | Acknowledgments I would like to express my deepest appreciation to all my supervisors, Prof. Sonia Johnson, Dr. Oliver Mason and Dr. Helen L. Fisher for their continuous guidance, encouragement and unwavering patience over the past three years. Your immense knowledge and advice on both research and my career have really been invaluable. Special thanks go to Dr. Craig Morgan from the Institute of Psychiatry for giving me the opportunity to be a part of such an exciting study and I greatly appreciate all the hard work of staff working on the EU-GEI project. I feel very humbled that I worked with such an inspiring research team. In particular, I would like to acknowledge Kathryn Hubbard whose enthusiasm and relentless support made the participants’ recruitment really pleasurable. I will also always be indebted to all individuals who participated in the study and were instrumental for its successful completion. Finally, I would like to thank my beloved boyfriend for his endless understanding and emotional backing and all my family and friends for believing in me. Without you I truly would not have been able to complete this challenging but fulfilling journey. ix | Introduction As psychotic disorders, such as schizophrenia, place a major burden on the individual, family and society, it is important to find ways to identify as early as possible who is at risk for such disorders. One way to do this would be to develop a better understanding of the aetiology of subclinical manifestations of psychosis such as schizotypy. Therefore, the importance of schizotypy research lies in the potential to identify the fundamental features of liability to psychosis and could have substantial implications for prevention, clinical assessment and treatment formulation. Literature suggests that the psychosis phenotype is expressed at subclinical levels (e.g. van Os et al. 2009). The term ‘psychosis continuum’ denotes the gradual transition from subclinical attenuated psychotic experiences or schizotypal traits to clinically relevant psychotic disorders as opposed to a sharp categorical division between disorder ‘absent’ or ‘present’. This proposition is based on the pioneering work of Rado (1960) followed by Meehl’s (1962) conceptualisation of schizotypy. According to their views, schizotypy represents a fundamental liability to schizophrenia and underlies a range of clinical manifestations ranging between healthy variation and severe mental illness. The multidimensionality of the schizotypy construct resembles schizophrenia symptomatology (e.g. Lenzenweger et al. 1991) and can be clustered as three dimensions: a positive cluster (e.g. unusual perceptions, magical thinking, ideas of reference), a negative cluster (e.g. restricted affect, social isolation) and a disorganized cluster (e.g. odd behaviour, odd speech) (Claridge et al. 1996;Vollema and Vandenbosch 1995). Similarly, the ‘schizophrenia prodrome’ (period preceding the diagnosis of schizophrenia) also reflects attenuated schizophrenia symptoms, making these constructs not easily distinguishable (Bedwell and Donnelly 2005). 1 |