Controlled Trial of Early Intervention With Children and Adolescents Exposed to Nonrelational Traumatic Events

Background A significant number of children and adolescents have been exposed to potentially traumatic events (PTEs). American and Australian studies report a prevalence of about 20-25% of high magnitude traumatic events among adolescents. A Danish study of four Nordic countries found that 90 % of adolescents with a mean age of 14.5 years had been exposed to at least one low magnitude traumatic event. A potentially traumatic event may be either relational such as sexual or physical abuse or nonrelational such as illness, motor vehicle accident or fire. The incidence of nonrelational trauma is very high in children and adolescents with accidental injuries being the largest cause of morbidity and mortality among children in the United States as well as in Denmark. Although many children and adolescents demonstrate great resilience or appear to be only briefly affected by nonrelational traumatic events a significant minority of young people will develop posttraumatic stress disorder (PTSD), posttraumatic stress symptoms (PTSSs) or other psychological difficulties following exposure to trauma. Accidental injury is the most common type of trauma. Researchers report prevalence rates of PTSD among children and adolescents in road traffic accidents ranging from 6% to 35%. Rates of PTSSs (also termed subclinical, subthreshold or partial PTSD) range from 11% to 50%6-14. These rates are important since recent research has indicated only minor differences in terms of distress and impairment between children meeting full and partial criteria for PTSD. The diagnosis of medical illness such as cancer or diabetes has also been shown to cause PTSD in some children and adolescents with prevalence rates of 5.4% for diabetes and of 10% to 18% for cancer. A recent meta-analysis of prevalence related to both injury and illness reported average prevalence rates of PTSD related to injury ranging from 0% to 37.5%, with a mean of 19.82%, and rates in ill youth ranging from 0% to 32%, with a mean of 12.04%. Preventive interventions for children and adolescents Given the high prevalence rates of exposure and the likewise high prevalence rates of PTSD and PTSSs after PTEs there is a pressing need for preventive interventions that decrease the risk of PTSD. In a recent systematic review by Forman-Hoffman et al. (2013) of interventions for children exposed to nonrelational traumatic events 6 different interventions based exclusively on exposure were identified. These included trauma-focused cognitive-behavioural therapy child and family traumatic stress intervention (CFTSI), 2 different school interventions with elements of CBT, early psychological intervention and a propranolol study. The two latter studies found no improvement in outcomes, the remaining 4 interventions found some improvement with regards to PTSD, severity of PTSD and PTSSs. Another recent meta-analysis by Kramer (2011) reviewing early psychological interventions in children and adolescents after single trauma included seven studies (two of which reappeared in the above mentioned review). These included CFTSI, school-based interventions, early psychological interventions, and two interventions based exclusively on psychoeducational information provision. The authors concluded that intervention effects regarding PTSSs and PTSD were small or non-significant but found promising beneficial tendencies with regards to dissociation, anxiety and, to some extent, arousal. Overall, the empirical literature has not supported the effectiveness of early psychological interventions such as psychological debriefing and single-session psychological intervention on preventing PTSD and PTSSs in children. These results are consistent with a Cochrane review by Rose 2002 of the effects of psychological debriefing for preventing PTSD in adults. However, school-based and family-focused interventions, as well as stepped-procedure and multiple-session psychological interventions have shown promising results. In conclusion, the current outcomes of the empirical research suggest early interventions that include the following elements31: age-appropriate psychoeducation, the provision of individual coping skills, parental involvement, social support, risk-assessment, multiple sessions and trauma exposure. The modest results of early psychological interventions suggest a need for a broader perspective on preventive interventions for children and adolescents exposed to nonrelational trauma. Despite the variety of interventions evaluated by the studies included in the two cited reviews none of the studies focused on the practice of health care professionals. This study is the first to conduct a controlled trial of a standardized preventive intervention conducted by health care professionals. The intervention to be evaluated in this study includes all of the above mentioned empirically supported elements except multiple sessions and trauma exposure. Trauma-informed health care The concept of trauma-informed practice has been advocated by institutions and researchers in diverse areas such as psychology, mental health care, education, child welfare, first responders, juvenile justice, and health care. The main aim of trauma-informed practice is to create systematic approaches within these areas to develop evidence-based services that address the impact of trauma on the children they serve37. This study will provide evidence on the impact of the practice of HCPs on symptoms, coping and social functioning of children, adolescents and their families who have been exposed to nonrelational trauma. It investigates the effects of educating HCPs in trauma-informed practice and the effects of a standardized trauma-informed plan of action implemented in health care departments working with children and adolescents. Design of the trial This study is designed as a controlled trial with one active intervention and one comparison. The study will be conducted across 19 health care departments in 7 hospitals in the Capitol Region of Denmark. There will be two measurement occasions: one assessment 4 weeks after exposure and one follow-up 6 months after exposure. The trial will run over 18 months. During the first 12 months the intervention will run in 12 of the participating departments while the other 8 departments will function as a control group. During the last 6 months the intervention will run in all nineteen departments thus expanding the sample size of the intervention group. Since the intervention is carried out by health care professionals shortly after exposure, the children in the control group will not receive intervention at any point. Because the intervention investigated in this trial is conducted by HCPs as an integrated part of their general practice (as opposed to a specific treatment program), it was neither practically possible nor ethically sound to design the allocation of participants in the control and intervention group as a randomization of subjects. For this reason, the allocation was done, not on a subject level, but on the level of departments. This precluded an actual randomization of the trial. However, it should be noted that the allocation of the departments is done before the intervention and thus prior to the recruitment of participants. This means that the allocation of subjects was predetermined in advance of their admission to the hospital. This prevents some of the bias associated with the lack of randomization. In the allocation of departments the best possible match of patient groups was pursued. It is the hypotheses of this study that: The children and adolescents who receive the intervention, compared to participants in the control group, will demonstrate greater reduction in posttraumatic stress, anxiety and depressive symptoms and higher ratings on positive cognitive coping skills and social functioning. Sample Size and Power Calculation Given that this is the first trial to examine the effect of a trauma-informed intervention conducted by HCPs as part of their general health care practice, calculations of power were very limited. Furthermore, effect sizes from related trials such as early psychological interventions and interventions based on information provision have been small or non-significant. Given these limitations, however, calculations of power were conducted on the basis of the available evidence. Results from a study by Kenardy et al. (2008) on an information-provision intervention had an effect size of 0.63 (between groups) on child anxiety. In a similar study by Cox et al. (2010) effect size on child anxiety was 0.33. Based on the results from Cox et al. (2010) a power analysis revealed that 115 families per group (N = 230) will result in 80% power for detecting a significant difference between the intervention and the control group at the 5% significance level. However, with an estimated 25% attrition rate over the assessment time periods, a total of 290 families should be aimed at to be recruited for the study. Another power analysis was conducted to calculate the power given a specific sample size. This was done because the participating departments are required to include all eligible families presented to the departments during the intervention period. It was estimated that each department would receive two eligible participants per month. Because of the difference in size between the intervention group and the control this resulted in a sample size of 516 for the intervention group and 168 for the control group (N = 684). With an estimated 25 % attrition rate these numbers amount to 387 for the intervention group and 126 for the control group (N = 513). A post hoc power calculation resulted in 94 % power with an effect size of 0.33 and a 5 % significance level. Statistical Considerations Prior to statistical analysis of outcomes, the control group and the intervention group will be compared for pre-intervention equivalence on demographic information and type of PTE. If significant differences are identified, this will be taken into account by including these variables as covariates in outcome analyses. The outcome data will be analysed and reported in terms of statistical significance of difference between intervention and control group in change over time on outcome, clinical significance of the change, and effect sizes. The analysis will be performed by researchers who have not been involved in assessing eligibility of participants, collection of data or entering of data. Primary analysis will be undertaken on an intention-to-treat (ITT) basis, including all participants who have received intervention or have been registered to the control group, regardless of withdrawal from the study. The data will be analysed with a mixed effects model, which is able to include participants whose data may be partly missing. Analysis of covariance will be conducted with three pairs of outcome: 1) child and parent outcome, 2) symptom outcome and coping outcome, and 3) symptom outcome and social functioning outcome. Other outcomes such as demographics and types of PTE will be described.