|Home | About | Journals | Submit | Contact Us | Français|
The impact of political violence on individuals presenting with an episode of first episode psychosis has not been examined. Individuals were assessed for exposure to political violence in Northern Ireland (the “Troubles”) by asking for a response to 2 questions: one asked about the impact of violence “on your area”; the second about the impact of violence “on you or your family’s life.” The participants were separated into 2 groups (high and low impact) for each question. Symptom profiles and rates of substance misuse were compared across the groups at baseline and at 3-year follow up. Of the 178 individuals included in the study 66 (37.1%) reported a high impact of the “Troubles” on their life and 81 (45.5%) a high impact of the “Troubles” on their area. There were no significant differences in symptom profile or rates of substance misuse between high and low groups at presentation. At 3-year follow-up high impact of the “Troubles” on life was associated with higher Positive and Negative Symptom Scale (PANSS) Total (P = .01), PANSS-Positive (P < .05), and PANSS-General (P < .01) scores and lower global assessment of functioning disability (P < .05) scores, after adjusting for confounding factors. Impact of the “Troubles” on area was not associated with differences in symptom outcomes. This finding adds to the evidence that outcomes in psychosis are significantly impacted by environmental factors and suggests that greater attention should be paid to therapeutic strategies designed to address the impact of trauma.
Since political violence erupted in Northern Ireland in 1968 the population has experienced continuous civil disorder and political violence known colloquially as the “Troubles.” To date over 3400 people have been killed and at least 40000 people have been injured. Since the first paramilitary ceasefires in 1994 there has been a marked reduction in the level of violence, though it has continued at a lower level.1
The effects of violence on the mental health of the population have been under-researched2 but there are number of important studies in the literature. A large community survey found that 21.3% of respondents reported that the “Troubles” had either “quite a bit” or “a lot” of impact on their lives or on the lives of their families and that psychological morbidity was more likely if a respondent had been directly exposed to “Troubles” related violence.3 More recent work continues to demonstrate that exposure to conflict related trauma results in increased rates of mood, anxiety, and substance use disorders.4,5
It is still widely held that environmental factors do not play an important causal role in schizophrenia and other psychotic illnesses despite the growing literature suggesting otherwise.6–8 A high proportion of patients with a psychotic diagnosis have been exposed to traumatic experiences including sexual abuse and physical abuse9,10 but despite this, little is known about the impact of trauma on the clinical course of illness in such patients.
We have previously found rates of childhood trauma to be higher in patients with a diagnosis of schizophrenia, as compared to patients with a non-psychotic psychiatric diagnosis, in Northern Ireland11 though we did not assess exposure to “Troubles” related trauma as a separate and distinct category. In a second study we demonstrated that exposure to traumatic events related to the “Troubles” is associated with increased levels of depression, anxiety, dissociative experiences, and hospital admissions in patients with chronic schizophrenia.12 There have been no previous studies specifically studying the impact of the “Troubles” on a population with first episode psychosis though based on previous research it can reasonable be postulated that stress is involved in the maintenance and exacerbation of psychosis.13 There is a developed literature regarding the impact of the experience of refugee status (and refugees have in many instances been exposed to political violence) on individuals, including the impact on psychotic symptoms.14–16 Recent review articles have sought to drawn together the unfolding evidence for the interplay between biological disposition and environmental insults.7,8
This study aims to examine the effect of such exposure on outcomes for first episode psychosis patients. We hypothesized that people with a first episode psychosis and a history of direct exposure to the “Troubles” would have a worse outcome as assessed on the Positive and Negative Syndrome Scale (PANSS),17 the Beck Depression Inventory (BDI),18 and the Global Assessment of Functioning (GAF)19 scale compared with people with a first episode psychosis and no such history.
A key finding of the landmark “Cost of the Troubles Study” suggests that geographical location (ie, where a person lives) is the best predictor of the effect of the “Troubles” on an individual.20 We thus sought to examine the effects on outcomes of the geographical location of each subject, ie, the extent to which the “Troubles” had impacted on the local area in which the person lives.
The Northern Ireland First Episode Psychosis Study (NIFEPS) is a large, population-based study designed to identify all persons aged 18–64 presenting to psychiatric services with first-in-lifetime psychotic symptoms over a 2-year period.21,22 Ethical approval was obtained from the Research Ethics Committee of Queen’s University Belfast.
The study population comprises that served by 2 geographically contiguous Health and Social Care (HSC) Trusts in the eastern part of Northern Ireland. Belfast HSC Trust serves the Belfast urban area (population 343879), and the Northern HSC Trust serves a more rural population in the north-east (population 426965). The total study population (770844) thus amounts to 46% of the population of Northern Ireland (population 1.69 million; all populations derived from NI Census 2001).
We identified potential cases from NHS mental health services in the study area including patients from inpatient and outpatient services, community mental health teams, crisis intervention teams and psychiatric liaison services. Inclusion criteria broadly followed the form of the WHO 10-Country Study.23 Subjects had to be aged 18–64 at the time of presentation to services and had to have been resident in the study area for at least 6 out of the previous 12 months. Those with a clear organic cause for their psychotic symptoms, those with a steroid induced psychosis, and those experiencing short-lived psychotic symptoms solely on the basis of uncomplicated alcohol or drug intoxication or withdrawal were excluded. There were no individuals who had developed a psychotic illness subsequent to a head injury caused by political violence. Previous psychiatric records were examined for any documented past history of psychosis.
A “leakage” study was carried out to identify patients who were not initially notified to researchers by clinical teams. Computerized information systems of the 2 Trusts were interrogated and all patients (both inpatient and outpatient) who had been given a diagnosis of psychosis were identified. Case notes of any potential subjects were then screened to see if they met criteria for the study. A number of suitable patients were identified by this process (n = 10) and invited to participate in the study (8 agreed to do so).
All patients who passed screening were invited to participate in the study as soon as they were deemed capable of providing written informed consent. A standard form was used to collect sociodemographic and illness-related characteristics. A checklist for substance misuse was completed. Clinical assessment was carried out using the Schedules for Clinical Assessment in Neuropsychiatry (SCAN) version 1.1.24 Researchers were trained in the use of the SCAN at a World Health Organisation (WHO)-validated training centre. A clinical consensus group meeting was held regularly and an ICD-1025 diagnosis was agreed for each subject on the basis of the SCAN and information from the case notes.
The “Troubles” questions3 were administered to 178 patients in total. Each was asked 2 questions: one about the impact of violence “on your area”; the second about the impact “on you or your family’s life.” (Specifically each subject was asked: “Thinking about the whole period since 1969, how much violence would you say there has been in this area because of the Troubles?” and “How much have the Troubles affected your own life and the lives of your immediate family?). Four ordered response categories ranging from “not very much at all” to “a lot” were possible. Responses to each question (impact on life and impact on area) were dichotomized into 2 groups: High (“quite a bit” and “a lot”) and low (“not very much at all” and “just a bit”). The PANSS, BDI, and GAF scales were completed during face to face assessments at baseline and 3-year follow up. Of the original group of 178 patients 122 were re-assessed at follow up. A number of patients were known to have left the study area (n = 11), some could not be traced/contacted (n = 23), and some refused consent for a follow up interview (n = 22). There were no significant differences on demographic variables or diagnostic case-mix between patients who were assessed at 3-year follow up and those who were not.
Substance misuse was assessed by the researchers and defined as the presence of DSM-IV substance abuse or dependence in the preceding year (assessed at 3 time points: baseline, 1 and 3 years). Information from face to face assessment was supplemented by review of case notes. Poly-substance misuse was defined as meeting said criteria for more than one substance. Methodology has been described previously.22
An income deprivation score (derived from the Northern Ireland Multiple Deprivation Measure-the percentage of people in the area who are income deprived) was determined for the “Output Area” (small geographical area containing approximately 350 people) where each individual was living at time of initial assessment.26
Urban/rural status for the “output area” in which each individual was living was determined from data available from the Northern Ireland Statistical Research Agency. Urban or rural determination is a function of population size, population density, and service provision. It is estimated that approximately 65% of the population of Northern Ireland live in urban areas.27
Responses to each of the 2 questions on the “Troubles” survey were analyzed separately. Differences between the high and low groups for each question were determined for age, gender, diagnostic category, income deprivation score, proportion living in urban areas, and baseline symptom scores using ANOVA and chi-squared analysis. Differences between symptom scores at 3 years were determined using univariate analysis of variance. Where significant differences in symptom scores at 3 years were observed further linear regression analysis was performed to adjust for possible confounding factors (age, gender, diagnostic group, income deprivation, and urban living). These were considered if found to vary significantly between “Troubles” groups (P < 0.05) at baseline and retained in the final regression analysis if observed to influence outcome. Baseline symptom scores were included in the linear regression analysis. Differences between rates of substance misuse and poly-substance misuse were determined using Pearson chi-squared analysis. Where significant differences were observed, logistic regression analysis was performed to adjust for confounding factors as described above. All data were analyzed in SPSS Version 19.
In response to the question asking about the impact of the “Troubles” on one’s life 57 individuals (32%) responded “not very much at all,” 55 (31%) “just a bit,” 35 (20%) “quite a bit,” and 31 (17%) “a lot.” In response to the question asking about the impact of the “Troubles” on one’s local area, 51 (29%) participants reported “not very much at all,” 46 (26%) “just a bit,” 46 (26%) “quite a bit,” and 35 (20%) reported “a lot.” The numbers of patients allocated to the high and low groups for each question are displayed in table 1.
For both questions there was no significant difference in age, gender, and diagnostic category between the high and low groups at initial assessment. The majority of study participants lived in an urban area (84.5%). Living in an urban area (or indeed a rural area) did not make it significantly more likely that subjects scored in the high group for impact of Troubles on life or area.
Income deprivation scores were significantly higher in the high impact group for both questions. There were no significant differences in PANSS scores and BDI scores between the groups at initial assessment (table 1). Median duration of untreated psychosis (DUP) of the sample was 6 months (interquartile range 1–12 months). There was no significant difference in DUP between the high and low groups for either question.
The proportions of participants in the high and low impact groups of each “Troubles” question did not differ significantly between those lost to follow up and those included in the 3-year analysis.
At 3 years the high impact on life group had significantly higher mean PANSS-Total, PANSS-Positive, PANSS-General scores, and BDI scores. Three-year GAF-symptoms and GAF-disability scores were significantly lower (table 2).
In a regression analyses adjusting for income deprivation and baseline symptom scores, the high impact on life group continued to have higher mean scores on PANSS-Total (beta = .23, P = .01; Adjusted R 2 = .19), PANSS-Positive (beta = .22, P < .05; Adjusted R 2 = .18), and PANSS-General (beta = .27, P < .01; Adjusted R 2 = .17) but not on the BDI at 3-year follow up. High impact of troubles on life predicted lower scores on GAF-disability at 3 years after adjusting for income deprivation and baseline symptoms (beta = −.19, P < .05; Adjusted R 2 = .14) but did not predict GAF-symptoms.
Information on substance misuse was available on 123 individuals (of the original group of 178) over the 3-year study period. Fifty-seven (46.3%) met criteria for substance misuse and 22 (17.9%) met criteria for poly-substance misuse on at least one assessment (meeting such criteria at the time of any assessment was considered to be clinically significant). The rate of substance misuse in the high impact on life group was significantly higher than in the low impact group (56.9% vs 38.9%, χ2 = 3.88, P = .05). Rates of poly-substance misuse were also significantly higher in the high impact group (27.5% vs 11.1%, χ2 = 5.43, P = .02). Income deprivation scores and urban/rural status did not significantly predict levels of substance misuse.
There were no significant differences in 3-year outcomes between the high and low impact on area groups on the PANSS, GAF, or BDI (table 3).
The rate of substance misuse in the high impact on area group was significantly higher than in the low impact group (57.4% vs 35.5%, χ2 = 5.93, P = .01). Rates of poly-substance misuse were also higher in the high impact group (23.0% vs 12.9%) but this was not statistically significant. Higher levels of income deprivation, or urban living, were not associated with higher levels of substance misuse.
The proportion of the subjects in this study who have been directly impacted by the “Troubles” is approximately twice that seen in the general population in a study employing identical methodology. As already noted, in the earlier community general population study 21.3% reported that the “Troubles” had either “quite a bit” or “a lot” of impact on their lives or on the lives of their families and 25.1% had “quite a bit” or “a lot” of impact on their area.3 The comparable figures for our sample are 37.1% and 45.5%, respectively.
The finding of a high rate of exposure to traumas related to the conflict in Northern Ireland in patients with psychosis is novel. The study group was drawn from all first inceptions of psychosis in an area covering approximately half of the Northern Ireland population and as such is a representative sample. This finding suggests 2 possibilities: individuals who are prone to psychotic illness are also more likely to be independently exposed to political violence; or that political violence directly impacts on the onset and/or course of psychotic illness.
We have previously shown an effect on symptoms,12 brain structure,28 and neuropsychology,29,30 of exposure to trauma in individuals with a diagnosis of schizophrenia. Similarly we have demonstrated a negative impact of exposure to trauma on outcomes in bipolar disorder.31 (These studies did not focus on “Troubles” related trauma but on trauma of all types).
In general previous research suggests that psychological morbidity is more likely if a respondent had been individually and directly exposed to the “Troubles” (eg, witnessing a shooting or bombing incident in close proximity). Our findings are consistent with this. Living in areas affected by the “Troubles” has no effect on outcome at 3 years (except that rates of substance misuse are significantly higher), but those who report a direct impact of the “Troubles” on their lives have significantly worse outcomes at 3 years. Furthermore, our study found that exposure to the “Troubles” was associated with increased rates of positive psychotic symptoms which suggests that such exposure has a direct adverse effect on the course of psychotic illness itself.
It might have been expected that a detrimental effect of exposure to the “Troubles” would have been apparent at baseline and it is difficult to explain why this was not the case. A number of studies have examined the effect of life events on the aetiology or course of schizophrenia. A landmark study found a significantly raised rate of life events in the 3 weeks prior to the onset or relapse of schizophrenic illness.32 Some subsequent studies have confirmed these findings33–35 but not all have done so.36 It is possible that the acute psychotic state may have masked any differences, which then become apparent during the recovery phase. Or alternatively it may be that exposure to political violence does not act as a causal factor in the inception of psychosis, but contributes negatively to the course and outcome of psychotic illness.
Living in areas affected by the “Troubles” and reporting a direct effect of the “Troubles” on life are both associated with significantly increased rates of substance misuse in our study. It could be argued that this could explain the poor outcome seen at 3 years, especially given that substance misuse has already been shown to predict poorer outcome in this sample22 but regression analyses controlling for substance misuse as an independent variable demonstrated that belonging to the high impact on life group continued to predict significantly higher PANSS-Positive, PANSS-General, and PANSS-Total scores.
These findings are consistent with the recent suggestion that psychotic syndromes can be understood as disorders of adaption to the social context.8 Although heritability is of some importance in the aetiology of psychotic illness, exposure to adverse environmental factors may have an impact on the developing “social brain” during sensitive periods. Political violence is possibly one such important environmental factor. Other authors have also explored this area and suggested that developmental alterations secondary to genetic variants, combined with childhood adversity, sensitize the dopamine system and result in excessive presynaptic synthesis and release.7 Social adversity biases cognitive schema that individuals use to interpret experience and on-going or future stress results in dysregulated dopamine release which leads to the misattribution of salience to stimuli. Paranoid delusions and hallucinations thus occur, and these are stressful experiences in themselves. In time, chronic stress “hardwires” psychotic symptoms. The experience of the Northern Ireland “Troubles”, continuing as it has over decades, is just such a form of chronic stress.
Two methodological concerns have been identified in studies of this nature: over-reporting and under-reporting of traumatic events. Over-reporting may be related to suggestion or possible secondary gain. Under-reporting may be related to feelings of guilt, shame, fear, a wish to protect the perpetrator, or fear of disbelief or rejection.37 In severe mental illness accurate recall of events may be complicated by delusions or hallucinations, pharmacotherapy, or severe substance abuse. The literature has consistently demonstrated that people with depressed mood over-report negative events from their lives and it is possible that the high depression scores observed in this study may have affected the reliability of reports.38
However, studies have shown that while there may be some concern about the use of self-report measures in people with schizophrenia and other psychotic illnesses, reports of trauma history are generally reliable.39 Furthermore, when reports are unreliable they are mostly the result of under-reporting rather than over-reporting.40 Retrospective reports of traumatic episodes by patients with a psychotic diagnosis can now be accepted as reliable.41
This result of this study represents an important finding as there are no previous studies which examine the effects on outcome of civil conflict on individuals with a psychotic illness, to the best of our knowledge. This is relevant to researchers and clinicians in any part of the world where conflict is prevalent but it also tells us more about the effects of trauma on individuals with a diagnosis of a psychotic illness in general.
Some authors have suggested that patients with schizophrenia may have less effective strategies for coping with stress42 and may have less social support available to them to buffer the effects of stress.43 Given this evidence, and the findings of this study, there may be benefits in developing intervention strategies (informed by treatments developed for PTSD) for patients with schizophrenia and other psychotic illnesses who have been traumatized by the “Troubles.”10 Recent published studies have demonstrated the efficacy of trauma-focused CBT for individuals who have been exposed to the “Troubles.”44 Research to establish if such approaches are useful for individuals with a psychotic illness ought to be initiated. The starting point must be a wider acceptance of the need to routinely enquire after traumatic experiences in clinical practice, something which is often neglected.45
The Research and Development Division (Public Health Agency), the Northern Ireland Health and Personal Social Services.
We thank the medical and nursing staff of the Trusts involved for referring patients and IT personnel at the Trusts for help with the leakage analysis. We also thank Dr Ruth Barr, Dr Gillian Smith, and Dr Adele McKinney for their assistance. The authors have declared that there are no conflicts of interest in relation to the subject of this study.