The epidemiology of trauma and post-traumatic stress disorder in a representative cohort of young people in England and Wales

Abstract

Background: Despite the emphasis placed on childhood trauma in psychiatry, comparatively little is known about the epidemiology of trauma and trauma-related psychopathology in young people. We therefore aimed to evaluate the prevalence, clinical features, and risk factors associated with trauma exposure and post-traumatic stress disorder (PTSD) in young people.

Methods: We carried out a comprehensive epidemiological study based on participants from the Environmental Risk Longitudinal Twin Study, a population-representative birth-cohort of 2232 children born in England and Wales in 1994-95. At the follow-up home visit at age 18 years, participants were assessed with structured interviews for trauma exposure, PTSD, other psychopathology, risk events, functional impairment, and service use. Risk factors for PTSD were measured prospectively over four previous assessments between age 5 and 12 years. The key outcomes were the prevalence, clinical features, and risk factors associated with trauma exposure and PTSD. We also derived and tested the internal validity of a PTSD risk calculator.

Findings: We found that 642 (31·1%) of 2064 participants reported trauma exposure and 160 (7·8%) of 2063 experienced PTSD by age 18 years. Trauma-exposed participants had high rates of psychopathology (187 [29·2%] of 641 for major depressive episode, 146 [22·9%] of 638 for conduct disorder, and 102 [15·9%] of 641 for alcohol dependence), risk events (160 [25·0%] of 641 for self-harm, 53 [8·3%] of 640 for suicide attempt, and 42 [6·6%] of 640 for violent offence), and functional impairment. Participants with lifetime PTSD had even higher rates of psychopathology (87 [54·7%] of 159 for major depressive episode, 43 [27·0%] of 159 for conduct disorder, and 41 [25·6%] of 160 for alcohol dependence), risk events (78 [48·8%] of 160 for self-harm, 32 [20·1%] of 159 for suicide attempt, and 19 [11·9%] of 159 for violent offence), and functional impairment. However, only 33 (20·6%) of 160 participants with PTSD received help from mental health professionals. The PTSD risk calculator had an internally validated area under the receiver operating characteristic curve of 0·74, indicating adequate discrimination of trauma-exposed participants with and without PTSD, and internally validated calibration-in-the-large of -0·10 and calibration slope of 0·90, indicating adequate calibration.

Interpretation: Trauma exposure and PTSD are associated with complex psychiatric presentations, high risk, and significant impairment in young people. Improved screening, reduced barriers to care provision, and comprehensive clinical assessment are needed to ensure that trauma-exposed young people and those with PTSD receive appropriate treatment.

Funding: The Medical Research Council, the National Institute of Child Health and Development, the Jacobs Foundation, the Nuffield Foundation, the National Society for Prevention of Cruelty to Children, the Economic and Social Research Council, the National Institute for Health Research, MQ, and Canadian Institutes for Advanced Research.

Figure 1

Prevalence of index trauma categories in trauma-exposed participants and in participants with PTSD The index traumas categorised as interpersonal assault or threat involved actions of another person intentionally causing or threatening death, injury, or sexual violation (eg, maltreatment by adults, bullying by peers), as opposed to index traumas categorised as accident or illness. Either trauma category might be directly experienced by the participant (ie, direct) or witnessed only (ie, witnessed). Other trauma categories were network trauma (ie, a traumatic event affecting someone in the participant's network that they learned details of, but did not directly experience or witness) or other trauma (ie, any other trauma that did not fall into the other categories). PTSD=post-traumatic stress disorder.

Figure 2

Risk of clinical features in the overall sample (n=2066) and the trauma-exposed participants (n=642) (A) Unadjusted ORs in the overall sample for mental health conditions. (B) Unadjusted ORs in the overall sample for risk events and functional impairment. (C) Unadjusted ORs in the trauma-exposed participants for mental health conditions. (D) Unadjusted ORs in the trauma-exposed participants for risk events and functional impairment. Filled circles signify significance in which p<0·05. Unfilled circles signify no significance in which p≥0·05. These results are detailed in the appendix (p 9). ADHD=attention-deficit hyperactivity disorder. CD=conduct disorder. GAD=generalised anxiety disorder. MDE=major depressive episode. NEET=not in education, employment, or training. ORs=odds ratios. PTSD=post-traumatic stress disorder. *Any mental health condition listed in (A) or (C).

Figure 3

PTSD risk calculator prediction performance (A) Frequency distribution of predicted probabilities of PTSD in trauma-exposed participants without and with PTSD. (B) Calibration curve of the observed probabilities of lifetime PTSD in relation to the predicted probabilities of lifetime PTSD. The PTSD risk calculator was derived using the multivariate logistic regression model predicting lifetime PTSD in trauma-exposed participants. PTSD=post-traumatic stress disorder.