Prevalence of Posttraumatic Stress Disorder among Children and Adolescents following Road Traffic Accidents: A Meta-Analysis
Abstract
Abstract
Objective:
Children and adolescents are among the most vulnerable road users, and road traffic accidents (RTAs) can lead to not only physical injuries but also adverse psychological outcomes, particularly posttraumatic stress disorder (PTSD). However, estimates of the prevalence of PTSD among children and adolescents following RTAs varied considerably across studies. Therefore, this study aimed to estimate the pooled prevalence of PTSD among this population.
Methods:
A systematic search for literature was performed in the electronic databases of PubMed, Web of Science, PsycINFO, and Embase. Heterogeneity was assessed using the Cochran’s chi-square test and quantified by the I2 value. Meta-regression analyses were carried out to identify the effects of some potential moderators on the overall heterogeneity. Subgroup analyses were performed to estimate the pooled prevalence of PTSD according to some sample characteristics.
Results:
Eleven eligible studies with a total of 1532 children and adolescents who were involved in RTAs were included. The overall heterogeneity (I2 = 89.7, P < 0.001) was high across the eligible studies, and the pooled prevalence of PTSD was 19.95% (95% confidence interval, 13.63% to 27.09%) by a random-effects model. No significant moderators of the overall heterogeneity were identified using meta-regression analyses. Subgroup analyses showed that the pooled prevalence of PTSD differed significantly according to the study location and gender (P < 0.05).
Conclusions:
One-fifth of children and adolescents developed PTSD in the aftermath of RTAs, indicating the need for regular assessment of PTSD and timely and effective psychological interventions among this population. Furthermore, more population-based studies with a large sample size are warranted. The protocol was registered in the PROSPERO database (No. CRD42018087941).
Résumé
Objectif:
Les enfants et les adolescents sont parmi les utilisateurs de la route les plus vulnérables, et les accidents de la circulation routière (ACR) peuvent entraîner non seulement des blessures physiques mais aussi des résultats psychologiques indésirables, particulièrement le trouble de stress post-traumatique (TSPT). Toutefois, les estimations de la prévalence du TSPT chez les enfants et les adolescents suivant des ACR variaient considérablement entre les études. Cette étude visait donc à estimer la prévalence regroupée du TSPT dans cette population.
Méthodes:
Une recherche systématique de la littérature a été menée dans les bases de données électroniques PubMed, Web of Science, PsycINFO et Embase. L’hétérogénéité a été évaluée à l’aide du test chi-carré de Cochran et quantifiée par la valeur I2. Des analyses de méta-régression ont été effectuées pour discerner les effets de certains modérateurs potentiels sur l’hétérogénéité globale. Des analyses de sous-groupe ont été exécutées pour estimer la prévalence regroupée du TSPT selon certaines caractéristiques de l’échantillon.
Résultats:
Onze études admissibles comptant 1 532 enfants et adolescents qui avaient été impliqués dans un ACR ont été incluses. L’hétérogénéité globale (I2 = 89,7; P < 0,001) était élevée dans toutes les études admissibles, et la prévalence regroupée du TSPT était de 19,95% (IC à 95% 13,63% à 27,09%) par un modèle d’effets aléatoires. Aucun modérateur significatif de l’hétérogénéité globale n’a été identifié à l’aide des analyses de méta-régression. Des analyses de sous-groupe ont montré que la prévalence regroupée du TSPT différait significativement selon l’endroit de l’étude et le sexe (P < 0,05).
Conclusions:
Un cinquième des enfants et des adolescents ont développé un TSPT suite à un ACR, indiquant le besoin d’une évaluation régulière du TSPT ainsi que d’interventions psychologiques ponctuelles et efficaces dans cette population. En outre, plus d’études dans la population avec de larges tailles d’échantillon sont justifiées. Le protocole a été enregistré dans la base de données PROSPERO (No CRD42018087941).
The size of the population involved in road traffic accidents (RTAs) has grown worldwide over the past few decades.1,2 According to the statistics by the World Health Organization (WHO), globally, 1.2 million people are killed in RTAs each year, and as many as 50 million are injured or disabled.3 Among all types of road users, children and adolescents are considered the most vulnerable.1,4 In 2004, road traffic injuries accounted for almost 262,000 deaths among children and adolescents aged 0 to 19 years, and among those aged 15 to 19 years and 5 to 14 years, road traffic injuries are the first and second leading cause of death, respectively.5 In addition to causing deaths and injuries, RTAs can lead to adverse psychological outcomes, particularly posttraumatic stress disorder (PTSD).6,7
PTSD could develop after exposure to an unintentional traumatic event, and RTA is the most common unintentional traumatic event that may involve children and adolescents.8,9 According to the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5), PTSD is characterized by 4 symptom clusters, including marked intrusion, avoidance, negative alterations in cognitions and mood, and marked alterations in arousal and reactivity symptoms, and these full symptom clusters must last at least 1 month.10 Studies on effects of various kinds of traumatic events have consistently shown that childhood PTSD could have long-term adverse effects on the quality of life. For example, Zatzick et al.11 interviewed 108 adolescents who experienced physical injuries and found that early PTSD symptoms were associated with a wide range of functional impairment at 1-year follow-up after the physical injury, and in a population-based survey of 596 children by Jia et al.,12 it was found that PTSD was strongly negatively associated with health-related quality of life 15 months after the 2008 Sichuan earthquake. Furthermore, Landolt et al.13 conducted a prospective study among 68 children following RTAs and found that early PTSD symptoms could have long-term adverse effects on health-related quality of life.
There has been a growing research interest in PTSD among children and adolescents following RTAs in recent years, with the estimates of the prevalence of PTSD varying considerably from 4.9% to 34.5% across the individual studies.4,14–16 This variation could be accounted for by some factors such as the differences in the interval between the occurrence of a trauma and assessment of PTSD (trauma assessment interval), tool used to assess PTSD, and sample characteristics, including gender, type of RTA, and injury severity.15–18 In addition, differences in social support, family cohesion, and parenting behaviors may contribute to the variation in the prevalence of PTSD among children and adolescents.19–21 A valid estimate of the pooled prevalence of PTSD among children and adolescents following RTAs is needed, since it could not only facilitate the efforts of the mental health providers to identify the accurate number of children and adolescents who may develop PTSD but also help balance the cost of prevention and treatment of PTSD by allocating psychological intervention resources proportionally. However, systematic research on PTSD among children and adolescents following RTAs is limited, and to date, no study has reported the pooled prevalence of PTSD among this population. Therefore, this study aimed to identify the pooled prevalence of PTSD among children and adolescents following RTAs by synthesizing relevant evidence through meta-analysis.
Methods
Search Strategy
This meta-analysis was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (see Suppl. File S1). The protocol was registered in the PROSPERO database (No. CRD42018087941).
A systematic search for literature was performed in the electronic databases of PubMed, Web of Science, PsycINFO, and Embase up to January 2018. A combination of the subject headings related to PTSD, RTA, and children or adolescents was used as the search terms. For example, the search terms for the database of PubMed were as follows: ((“Stress Disorders, Post-Traumatic”[Mesh]) AND (“Child”[Mesh] OR “Adolescent”[Mesh])) AND “Accidents, Traffic”[Mesh]). A full list of the search terms is in Supplemental File S2. Furthermore, all reference lists of the included studies were manually searched for further relevant publications.
Eligibility Criteria
Two reviewers (W.D. and A.C.K.) independently identified the eligibility of studies. The inclusion criteria were 1) observational studies investigating PTSD among children and adolescents who were involved in an RTA, regardless of type of RTA; 2) PTSD was assessed at least 1 month after the RTA using the self-report questionnaires or structured interviews according to the DSM criteria; 3) data on the sample size and the prevalence of PTSD among children and adolescents following an RTA were reported; and (4) full article was written in English. In addition, for longitudinal studies, only the prevalence of PTSD at baseline was included, and if duplicated data were observed, only the first publication was included. Studies were excluded if they were abstracts, case reports, comments, reviews, dissertations, or book chapters. Any discrepancies between the 2 reviewers were resolved via consensus.
Data Extraction
Two reviewers (W.D. and A.C.K.) independently extracted the following data and assessed the risk of bias from each eligible study: first author, year of publication, study location, study design, sample source, age, male proportion, trauma assessment interval, tool used to assess PTSD, number of subjects with PTSD, sample size, and the prevalence of PTSD. Furthermore, if available, data regarding the type of RTA and parental reporting of children’s PTSD were extracted to perform subgroup analyses. Any disagreements between the foregoing reviewers were resolved via discussion with a third reviewer (J.D.), and an unweighted κ value was calculated to evaluate the agreement between reviewers.
Risk-of-Bias Assessment
The Loney criteria were used to assess the risk of bias for each eligible study. This scale is widely used to assess the quality of observational studies on the prevalence of health-related outcomes.22–24 It consists of 8 items to comprehensively evaluate the risk of bias such as unbiased sampling frame, adequate sample size, and standard tools. A study is given 1 point for each item it meets, making the maximal total score to be 8 points, with higher scores suggesting lower risk of bias.
Statistical Analysis
All statistical analyses were conducted in R statistical software version 3.4.1 (https://www.r-project.org) using the “meta” and “metafor” packages.
Estimates of the prevalence of PTSD among eligible studies were transformed using the Freeman-Tukey double arcsine method. Heterogeneity was assessed using Cochran’s χ2 test and quantified by the I2 value, which indicates the percentage of the total variation due to interstudy heterogeneity rather than chance. An I2 value of <25%, 25% to 75%, and >75% suggests low, moderate, and high degree of heterogeneity, respectively.25 The pooled prevalence of PTSD was synthesized by a random-effects model if significant heterogeneity (P value for Cochran’s χ2 test <0.05) was observed across studies. In contrast, a fixed-effects model was applied.26,27 For each pooled estimate, its corresponding 95% confidence interval (CI) was calculated.
Meta-regression analyses with a mixed-effect model were carried out to identify the effects of potential moderators on the overall heterogeneity using the restricted maximum-likelihood estimator method if significant heterogeneity was observed across all included studies. Potential moderators included the study design, mean age, male proportion, tool used to assess PTSD, and score of risk-of-bias assessment.
Subgroup analyses were performed to estimate the pooled prevalence of PTSD with regard to study location, trauma assessment interval, tool used to assess PTSD, gender, type of RTA, and parental reporting of children’s PTSD. The χ2 test was used to assess the differences across each subgroup, and the significance level was set at P < 0.05.
Sensitivity analysis was performed by serially removing studies one by one to show the effect of that on the effect sizes.28 Publication bias was evaluated not only visually by presenting a funnel plot for asymmetry but also statistically by performing Begg’s rank test and Egger’s linear test. Furthermore, the trim-and-fill method was used to obtain an adjusted pooled estimate in the presence of publication bias.29
Results
Search Results
Figure 1 shows the PRISMA flowchart of the selection process of the eligible studies. A total of 622 records were initially retrieved from the databases by the search strategy. After removal of duplicated records, 388 records were screened. After abstract screening, 41 full-text articles were shortlisted for potential eligibility. Among these 41 articles, 4 were excluded for not identifying PTSD at least 1 month after the RTA, 2 were excluded for not identifying PTSD according to the DSM criteria, 16 were excluded for not reporting the prevalence of PTSD among children and adolescents following the RTA, and 8 were excluded for duplicated data (Suppl. File S3). Therefore, 11 eligible articles were ultimately included in this study.
Study Characteristics
Table 1 shows the characteristics of the eligible studies. The 11 eligible studies were carried out in 5 countries, including the United States, United Kingdom, Greece, France, and China. In addition, 7 were longitudinal and 4 were cross-sectional; 10 recruited participants from hospital and 1 recruited participants via the official police accident reports; 2 made the diagnosis of PTSD using exclusively self-report questionnaire, 8 using exclusively structured interview, and 1 using both the structured interview and the self-report questionnaire. The trauma assessment interval of these studies ranged from 1 month to 18 months following the RTA, and the mean age of participants ranged from 6.5 to 13.97 years.
| First Author | Year of Publication | Study Location | Study Design | Sample Source | Age, y | Male Proportion, % | Trauma Assessment Interval | Tool Used to Assess PTSD | No. of Subjects with PTSD | Sample Size | PTSD Prevalence, % | ||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Range | Mean (SD) | Self-report questionnaire | Structured interview | ||||||||||
| Mirza30 | 1998 | United Kingdom | Longitudinal | Hospital based | 8-16 | 13.61 (2.44) | 68.1 | 9-13 weeks | FRI | K-SADS-L (DSM-IV) | 33 | 119 | 27.7 |
| Stallard16 | 1998 | United Kingdom | Cross-sectional | Hospital based | 5-18 | 13.97 (3.59) | 57.1 | 40.3 days | — | CAPS-C (DSM-IV) | 41 | 119 | 34.5 |
| de Vries17 | 1999 | United States | Cross-sectional | Hospital based | 3-17 | 9.4 (3.5) | 66.7 | 7-12 months | PCL-C (DSM-IV) | — | 26 | 102 | 25.5 |
| Keppel-Benson31 | 2002 | United States | Cross-sectional | Official police accident reports | 7-16 | 11.6 (3.2) | 58.0 | 2-18 months | — | DICA-R (DSM-III-R) | 7 | 50 | 14.0 |
| Zink32 | 2003 | United States | Longitudinal | Hospital based | 7-15 | 10.8 (2.6) | 59.4 | 2 months | — | DICA-R (DSM-III) | 26 | 143 | 18.2 |
| Kassam-Adams33 | 2004 | United States | Longitudinal | Hospital based | 8-17 | 11.3 (2.5) | 74.9 | 3-12 months | — | CAPS-CA (DSM-IV) | 10 | 177 | 5.6 |
| Bryant34 | 2004 | United Kingdom | Longitudinal | Hospital based | 5-16 | 12.27 (2.86) | 55.0 | 3 months | Questions based on DSM-IV criteria | — | 19 | 81 | 23.5 |
| Pervanidou35 | 2007 | Greece | Longitudinal | Hospital based | 7-18 | 10.96 (2.51) | 71.4 | 1 month | — | K-SADS-L (DSM-IV) | 23 | 56 | 41.1 |
| Bui15 | 2010 | France | Cross-sectional | Hospital based | 8-15 | 11.7 (2.2) | 53.4 | 5 weeks | — | CAPS-CA (DSM-IV) | 5 | 103 | 4.9 |
| Wu18 | 2016 | China | Longitudinal | Hospital based | 1-13 | 6.8 (0.9) | 53.8 | 3 months | — | CAPS-CA (DSM-IV) | 133 | 537 | 24.8 |
| Meiser-Stedman6 | 2017 | United Kingdom | Longitudinal | Hospital based | 2-10 | 6.5 (2.8) | 52.1 | 6 months | — | CAPS-CA (DSM-IV) | 6 | 45 | 13.3 |
CAPS-C, Clinician-Administered PTSD Scale for Children; CAPS-CA, Clinician-Administered PTSD Scale for Children and Adolescents; DSM, Diagnostic and Statistical Manual of Mental Disorders; DICA-R, Diagnostic Interview for Children and Adolescents–Revised Version; FRI, Frederick’s Reaction Index; K-SADS-L, Kiddie-Schedule for Affective Disorders and Schizophrenia-Lifetime Version; PCL-C, PTSD checklist for children; PTSD, post-traumatic stress disorder; SD, standard deviation.
Risk-of-Bias Assessment
The results of risk-of-bias assessment are shown in Supplemental File S4. Based on the Loney criteria, 7 of the 11 eligible studies were scored 5 points, 3 scored 4 points, and 1 scored 6 points. None of them used a random sampling method or recruited participants from the whole population. Furthermore, none applied unbiased sampling frame, and only 1 had an adequate sample size of >300. Agreement between the 2 reviewers assessing the risk of bias using the Loney criteria was good (κ = 0.69).
Pooled Prevalence of PTSD among Children and Adolescents following RTA
Figure 2 shows the forest plot of all included studies. A total of 1532 children and adolescents were screened in the aftermath of RTAs, of which 329 were identified with PTSD. Estimates of the prevalence of PTSD across the 11 eligible studies ranged from 4.9% to 41.1%. The lowest prevalence of PTSD (4.9%) was reported in France in a hospital-based study conducted 5 weeks after the RTA,15 and the highest prevalence of PTSD (41.1%) was reported in Greece in a hospital-based study conducted 1 month after the RTA.35 Significantly high heterogeneity (I2 = 89.7, P < 0.001) was observed across all included studies, and the pooled prevalence of PTSD among children and adolescents following the RTA was 19.95% (95% CI, 13.63% to 27.09%) using a random-effects model.
Meta-Regression Analyses
The results of meta-regression analyses indicated that the study design (P = 0.761), male proportion (P = 0.724), mean age of participants (P = 0.684), tool used to assess PTSD (P = 0.581), and score of risk-of-bias assessment (P = 0.244) were not significant moderators of the overall heterogeneity. Detailed data are shown in Supplemental File S5.
Subgroup Analyses
Table 2 shows the results of subgroup analyses. For studies conducted <3 months and ≥3 months following the RTA, the pooled prevalence of PTSD was 23.49% (95% CI, 11.96% to 37.41%) and 17.76% (95% CI, 9.23% to 28.25%), respectively. For studies using exclusively self-report questionnaire and exclusively structured interview to assess PTSD, the pooled prevalence of PTSD was 24.58% (95% CI, 18.55% to 31.14%) and 17.96% (95% CI, 10.19% to 27.28%), respectively. For the studies conducted in the United States and United Kingdom, the pooled prevalence of PTSD was 14.93% (95% CI, 6.57% to 25.80%) and 25.46% (95% CI, 18.04% to 33.64%), respectively. In addition, for male and female subjects, the pooled prevalence of PTSD was 22.21% (95% CI, 18.66% to 25.96%) and 34.45% (95% CI, 21.94% to 48.12%), respectively. Furthermore, the pooled prevalence of PTSD among subjects who reported PTSD by themselves was 20.32% (95% CI, 12.07% to 30.01%), and for those whose symptoms were reported by their parents, the pooled prevalence was 19.76% (95% CI, 15.30% to 24.61%).
| Subgroup/Category | No. of Studies | No. of Subjects with PTSD | Sample Size | Pooled Prevalence (95% CI), % | χ2 Value | P Value |
|---|---|---|---|---|---|---|
| Trauma assessment interval | 1.951 | 0.162 | ||||
| <3 months | 5 | 128 | 540 | 23.49 (11.96 to 37.41) | ||
| ≥3 months | 5 | 194 | 942 | 17.76 (9.23 to 28.25) | ||
| Tool used to assess PTSD | 1.684 | 0.194 | ||||
| Self-report questionnaire | 2 | 45 | 183 | 24.58 (18.55 to 31.14)a | ||
| Structured interview | 8 | 251 | 1230 | 17.96 (10.19 to 27.28) | ||
| Study location | 20.252 | <0.001b | ||||
| United States | 4 | 69 | 472 | 14.93 (6.57 to 25.80) | ||
| United Kingdom | 4 | 99 | 364 | 25.46 (18.04 to 33.64) | ||
| Gender | 7.090 | 0.008b | ||||
| Male | 4 | 114 | 510 | 22.21 (18.66 to 25.96)a | ||
| Female | 4 | 118 | 391 | 34.45 (21.94 to 48.12) | ||
| Type of RTA | 4.673 | 0.097 | ||||
| Auto occupant | 3 | 40 | 130 | 30.37 (22.53 to 38.79)a | ||
| Pedestrian | 3 | 36 | 118 | 30.43 (22.29 to 39.21)a | ||
| Bicyclist | 3 | 23 | 116 | 19.65 (12.70 to 27.60) | ||
| Parental reporting of children’s PTSD | 0.462 | 0.497 | ||||
| Yes | 3 | 58 | 290 | 19.76 (15.30 to 24.61)a | ||
| No | 8 | 271 | 1242 | 20.32 (12.07 to 30.01) |
CI, confidence interval; PTSD, posttraumatic stress disorder; RTA, road traffic accident.
a Data were combined using a fixed-effects model.
bP value for χ2 tests <0.05.
The results of χ2 tests indicated that the pooled prevalence of PTSD among children and adolescents following RTAs differed significantly according to the study location and gender (P < 0.05). Specifically, the pooled prevalence of PTSD among subjects in the United States was significantly lower than that among subjects in the United Kingdom (P < 0.05), and the pooled prevalence of PTSD among female subjects was significantly higher than that among male subjects (P < 0.05). In addition, the pooled prevalence of PTSD among subjects did not differ significantly according to the trauma assessment interval, tool used to assess PTSD, type of RTA, and parental reporting of children’s PTSD (P > 0.05).
Sensitivity Analysis and Publication Bias
After serially removing studies one by one, the I2 values ranged from 83.9% to 90.7%, and the pooled prevalence of PTSD ranged from 18.30% (95% CI, 12.18% to 25.31%) to 21.95% (95% CI, 16.18% to 28.30%), indicating low sensitivity.
Figure 3 shows the funnel plot of all included studies. The funnel plot was almost symmetrical, and both the results of Begg’s rank test (z = –0.468, P = 0.639) and Egger’s linear test (t = –0.274, P = 0.790) indicated no significant publication bias.
Discussion
To our knowledge, this is the first meta-analysis to estimate the pooled prevalence of PTSD among children and adolescents following an RTA. This study included 11 eligible studies with a total of 1532 children and adolescents who were involved in RTAs. The results indicated that one-fifth of these children and adolescents developed PTSD in the aftermath of RTAs, and in comparison with results of similar meta-analyses exploring the pooled prevalence of PTSD following specific traumatic events, the pooled prevalence of PTSD found in this study (19.95%) was lower than that among earthquake survivors (23.66%),23 adult RTA survivors (22.25%),36 adult patients after acute orthopaedic trauma (26.6%),37 and children exposed to a war (47%).38 However, it was much higher than that among flood survivors (15.74%),39 patients of stroke and transient ischemic attack (13%),40 patients with chronic pain (9.7%),41 and patients with breast cancer (9.6%).42 The high pooled prevalence of PTSD among children and adolescents following RTAs suggests that, in addition to the physical treatment, the service providers should pay special attention to the mental health of this population, and routine assessment of PTSD is imperative. Furthermore, given the fact that childhood PTSD may lead to long-term adverse effects on the quality of life, timely and effective psychological interventions such as cognitive behavior therapy are needed for children and adolescents following RTAs.43
Numerous studies have consistently shown that the prevalence of PTSD was associated with the timing of PTSD assessment following specific traumatic events. For example, Edmondson et al.40 conducted a meta-analysis exploring the pooled prevalence of PTSD among survivors of stroke and transient ischemic attack and found that the pooled prevalence of PTSD assessed within 1 year following stroke and transient ischemic attack (23%) was significantly higher than that assessed at more than 1 year (11%). In addition, Hiller et al.44 conducted a meta-analysis of longitudinal studies on PTSD among those aged 5 to 18 years and found moderate declines in the prevalence of PTSD as well as the symptom severity of PTSD over the first 3 to 6 months posttrauma. This variation could be explained by the possibility of recovery from PTSD over time.45,46 For example, Truss et al.46 found that the prevalence of clinically posttraumatic symptoms among 122 children following pediatric concussion was 16%, 10%, and 6% at 2 weeks, 1 month, and 3 months postconcussion, respectively, and 25% of the children recovered from PTSD during the 3 months. However, in this study, although the results of subgroup analyses indicated that the pooled prevalence of PTSD assessed at less than 3 months after RTAs (23.49%) was higher than that assessed at least 3 months after RTAs (17.76%), this variation was not significant. Possible reasons could be that the number of relevant studies was limited when synthesizing the data and the heterogeneity across relevant studies was high.
It has been widely understood that the self-report questionnaires are more likely to overestimate the prevalence of PTSD than the structured interviews. For example, Siqveland et al.41 synthesized the results of 21 studies exploring the prevalence of PTSD among patients with chronic pain and found that the prevalence of PTSD identified by self-reported questionnaire (20.4%) was significantly higher than that identified by structured interview (4.5%). Similar result was observed in a study exploring the pooled prevalence of PTSD among cancer patients.47 However, no significant difference in the pooled prevalence of PTSD according to the tool used to assess PTSD was observed in this study. It is worth noting here that only 2 eligible studies used self-report questionnaires to assess PTSD when synthesizing relevant data, which may explain the inconsistency of the preceding results with those of previous similar studies.
Furthermore, the pooled prevalence of PTSD in this study differed significantly according to the study location and gender. Country difference in PTSD prevalence could be attributed to the differences in sample characteristics, including socioeconomic, sociocultural, and biological characteristics.48,49 In addition, gender difference in the pooled prevalence of PTSD could be accounted for by the gender differences in the interpretation of stressful events and the strategies used to cope with stressful events.50,51 These findings indicate that the service providers should allocate the psychological intervention resources to those who were involved in RTAs proportionally according to the degree of the prevalence of PTSD in a country, and special attention should be paid to girls.
Few studies have compared the parent-child discrepancy in reporting children’s PTSD symptoms. In a study by Dyb et al.,52 it was found that parents reported significantly less children’s PTSD symptoms than children themselves, and this parent-child discrepancy was more prominent among much younger children. The results of our study showed that the pooled prevalence of children’s PTSD reported by children and adolescents did not differ significantly from that reported by their parents. Given the limited number of studies that provided relevant data, future studies are still needed to clarify this discrepancy.
In terms of risk-of-bias assessment, it is worth noting here that none of the 11 eligible studies were population based, and almost all eligible studies recruited participants from hospital, which may not only induce selection bias but also preclude from generalizing our findings to population-based studies. Also, only 1 of the eligible studies had a sample size of >300. Studies with a small sample size may tend to overestimate the effect size.53 In this regard, more population-based studies with a large sample size are warranted. In addition, many previous studies have consistently shown that risk of bias was an important source of the heterogeneity.54,55 However, the results of meta-regression analyses in this study showed that risk-of-bias assessment score was not a significant moderator of the overall heterogeneity. This may be due to the fact that, although the risk-of-bias assessment score differed from 4 to 6 points across the eligible studies, these studies were mostly at moderate risk of bias.
Some limitations should be acknowledged. First, subgroup analysis according to the injury severity was unable to be performed since the instruments used to assess injury severity were quite inconsistent across the eligible studies. Among the eligible studies, de Vries et al.17 found that injury severity was not related to PTSD, while Keppel-Benson et al.31 found that severe physical injury predicted more PTSD symptoms. Therefore, more future studies using consistently standard scales such as an injury severity scale to assess injury severity are needed. Second, none of the 11 eligible studies assessed PTSD based on the DSM-5 criteria, which may preclude from generalizing our findings to the studies using the tools according to the DSM-5 criteria to identify PTSD. Third, the differences in the pooled prevalence of PTSD within each subgroup were compared in the absence of adjustment for potential confounders. In addition, the overall heterogeneity was high across the eligible studies, and no significant moderators of the overall heterogeneity were identified in this study, indicating that future studies should explore more factors that may contribute to the differences in the prevalence of PTSD among children and adolescents following RTAs, such as history of psychiatric disorders, history of previous traumatic events, social support, and parenting behaviors.
Conclusions
The pooled prevalence of PTSD among children and adolescents following RTAs was 19.95% (95% CI, 13.63% to 27.09%), which underscores the importance of regular assessment of PTSD among this population, as well as the need for the implementation of timely and effective psychological interventions. No significant moderators of the overall heterogeneity were identified in this study, indicating that future studies should explore more factors associated with PTSD among children and adolescents who have been involved in RTAs. Furthermore, more population-based studies with a large sample size as well as an unbiased sampling frame are needed.
Acknowledgments
We are grateful to all authors of the full-text articles.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported in part by the Canadian Institutes of Health Research Foundation (FDN-148438); the Specialized Research Fund for the Doctoral Program of Higher Education (20130162110054); the Natural Science Foundation of Hunan Province, China (2016JJ2153); and the Fundamental Research Funds for the postgraduates of Central South University (2015zzts282).
ORCID iD
Wenjie Dai, MD https://orcid.org/0000-0002-0922-7223
References
1. Meiser-Stedman R, Smith P, Glucksman E, et al. The posttraumatic stress disorder diagnosis in preschool- and elementary school-age children exposed to motor vehicle accidents. Am J Psychiatry. 2008;165(10):1326–1337.
2. Stallard P, Velleman R, Baldwin S. Recovery from post-traumatic stress disorder in children following road traffic accidents: the role of talking and feeling understood. J Community Appl Soc Psychol. 2001;11(1):37–41.
3. World Health Organization. World report on road traffic injury prevention. http://www.who.int/violence_injury_prevention/publications/road_traffic/world_report/en/. Accessed June 6, 2018.
4. Schafer I, Barkmann C, Riedesser P, et al. Posttraumatic syndromes in children and adolescents after road traffic accidents—a prospective cohort study. Psychopathology. 2006;39(4):159–164.
5. World Health Organization. World report on child injury prevention. http://www.who.int/violence_injury_prevention/child/injury/world_report/en/. Accessed June 6, 2018.
6. Meiser-Stedman R, Smith P, Yule W, et al. Posttraumatic stress disorder in young children 3 years posttrauma: prevalence and longitudinal predictors. J Clin Psychiatry. 2017;78(3):334–339.
7. Williams JL, Rheingold AA, Knowlton AW, et al. Associations between motor vehicle crashes and mental health problems: data from the National Survey of Adolescents–Replication. J Trauma Stress. 2015;28(1):41–48.
8. Blanchard EB, Hickling EJ, Taylor AE, et al. Who develops PTSD from motor vehicle accidents? Behav Res Ther. 1996;34(1):1–10.
9. Ellis A, Stores G, Mayou R. Psychological consequences of road traffic accidents in children. Eur Child Adolesc Psychiatry. 1998;7(2):61–68.
10. American Psychiatric Association (APA). Diagnostic and statistical manual of mental disorders. 5th ed. Washington, DC: American Psychiatric Association; 2013.
11. Zatzick DF, Jurkovich GJ, Fan MY, et al. Association between posttraumatic stress and depressive symptoms and functional outcomes in adolescents followed up longitudinally after injury hospitalization. Arch Pediatr Adolesc Med. 2008;162(7):642–648.
12. Jia Z, Tian W, He X, et al. Mental health and quality of life survey among child survivors of the 2008 Sichuan earthquake. Qual Life Res. 2010;19(9):1381–1391.
13. Landolt MA, Vollrath ME, Gnehm HE, et al. Post-traumatic stress impacts on quality of life in children after road traffic accidents: prospective study. Aust N Z J Psychiatry. 2009;43(8):746–753.
14. Tierens M, Bal S, Crombez G, et al. The traumatic impact of motor vehicle accidents in high school students. J Pediatr Psychol. 2012;37(1):1–10.
15. Bui E, Brunet A, Allenou C, et al. Peritraumatic reactions and posttraumatic stress symptoms in school-aged children victims of road traffic accident. Gen Hosp Psychiatry. 2010;32(3):330–333.
16. Stallard P, Velleman R, Baldwin S. Prospective study of post-traumatic stress disorder in children involved in road traffic accidents. BMJ. 1998;317(7173):1619–1623.
17. de Vries AP, Kassam-Adams N, Cnaan A, et al. Looking beyond the physical injury: posttraumatic stress disorder in children and parents after pediatric traffic injury. Pediatrics. 1999;104(6):1293–1299.
18. Wu F, Meng WY, Hao CZ, et al. Analysis of posttraumatic stress disorder in children with road traffic injury in Wenzhou, China. Traffic Inj Prev. 2016;17(2):159–163.
19. Lee SH, Kim EJ, Noh JW, et al. Factors associated with post-traumatic stress symptoms in students who survived 20 months after the Sewol ferry disaster in Korea. J Korean Med Sci. 2018;33(11):e90.
20. Williamson V, Creswell C, Fearon P, et al. The role of parenting behaviors in childhood post-traumatic stress disorder: a meta-analytic review. Clin Psychol Rev. 2017;53:1–13.
21. Dai W, Chen L, Tan H, et al. Association between social support and recovery from post-traumatic stress disorder after flood: a 13-14 year follow-up study in Hunan, China. BMC Public Health. 2016;16:194.
22. Loney PL, Chambers LW, Bennett KJ, et al. Critical appraisal of the health research literature: prevalence or incidence of a health problem. Chronic Dis Can. 1998;19(4):170–176.
23. Bassuk EL, Richard MK, Tsertsvadze A. The prevalence of mental illness in homeless children: a systematic review and meta-analysis. J Am Acad Child Adolesc Psychiatry. 2015;54(2):86–96.e82.
24. Dai W, Chen L, Lai Z, et al. The incidence of post-traumatic stress disorder among survivors after earthquakes: a systematic review and meta-analysis. BMC Psychiatry. 2016;16:188.
25. Ades AE, Lu G, Higgins JP. The interpretation of random-effects meta-analysis in decision models. Med Decis Making. 2005;25(6):646–654.
26. Li X, Wang J, Zhou J, et al. The association between post-traumatic stress disorder and shorter telomere length: a systematic review and meta-analysis. J Affect Disord. 2017;218:322–326.
27. Huo Q, Zhang N, Yang Q. Epstein-Barr virus infection and sporadic breast cancer risk: a meta-analysis. PLoS One. 2012;7(2):e31656.
28. Syrjanen K. Geographic origin is a significant determinant of human papillomavirus prevalence in oesophageal squamous cell carcinoma: systematic review and meta-analysis. Scand J Infect Dis. 2013;45(1):1–18.
29. Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics. 2000;56(2):455–463.
30. Mirza KA, Bhadrinath BR, Goodyer IM, et al. Post-traumatic stress disorder in children and adolescents following road traffic accidents. Br J Psychiatry. 1998;172:443–447.
31. Keppel-Benson JM, Ollendick TH, Benson MJ. Post-traumatic stress in children following motor vehicle accidents. J Child Psychol Psychiatry. 2002;43(2):203–212.
32. Zink KA, McCain GC. Post-traumatic stress disorder in children and adolescents with motor vehicle–related injuries. J Spec Pediatr Nurs. 2003;8(3):99–106.
33. Kassam-Adams N, Winston FK. Predicting child PTSD: the relationship between acute stress disorder and PTSD in injured children. J Am Acad Child Adolesc Psychiatry. 2004;43(4):403–411.
34. Bryant B, Mayou R, Wiggs L, et al. Psychological consequences of road traffic accidents for children and their mothers. Psychol Med. 2004;34(2):335–346.
35. Pervanidou P, Kolaitis G, Charitaki S, et al. Elevated morning serum interleukin (IL)–6 or evening salivary cortisol concentrations predict posttraumatic stress disorder in children and adolescents six months after a motor vehicle accident. Psychoneuroendocrinology. 2007;32(8-10):991–999.
36. Lin W, Gong L, Xia M, et al. Prevalence of posttraumatic stress disorder among road traffic accident survivors: a PRISMA-compliant meta-analysis. Medicine. 2018;97(3):e9693.
37. Muscatelli S, Spurr H, O’Hara NN, et al. Prevalence of depression and posttraumatic stress disorder after acute orthopaedic trauma: a systematic review and meta-analysis. J Orthop Trauma. 2017;31(1):47–55.
38. Attanayake V, McKay R, Joffres M, et al. Prevalence of mental disorders among children exposed to war: a systematic review of 7,920 children. Med Confl Surviv. 2009;25(1):4–19.
39. Chen L, Liu A. The incidence of posttraumatic stress disorder after floods: a meta-analysis. Disaster Med Public Health Prep. 2015;9(3):329–333.
40. Edmondson D, Richardson S, Fausett JK, et al. Prevalence of PTSD in survivors of stroke and transient ischemic attack: a meta-analytic review. PLoS One. 2013;8(6):e66435.
41. Siqveland J, Hussain A, Lindstrom JC, et al. Prevalence of posttraumatic stress disorder in persons with chronic pain: a meta-analysis. Front Psychiatry. 2017;8:164.
42. Wu X, Wang J, Cofie R, et al. Prevalence of posttraumatic stress disorder among breast cancer patients: a meta-analysis. Iran J Public Health. 2016;45(12):1533–1544.
43. Morina N, Koerssen R, Pollet TV. Interventions for children and adolescents with posttraumatic stress disorder: a meta-analysis of comparative outcome studies. Clin Psychol Rev. 2016;47:41–54.
44. Hiller RM, Meiser-Stedman R, Fearon P, et al. Research review: changes in the prevalence and symptom severity of child post-traumatic stress disorder in the year following trauma—a meta-analytic study. J Child Psychol Psychiatry. 2016;57(8):884–898.
45. Dikmen-Yildiz P, Ayers S, Phillips L. Longitudinal trajectories of post-traumatic stress disorder (PTSD) after birth and associated risk factors. J Affect Disord. 2018;229:377–385.
46. Truss K, Godfrey C, Takagi M, et al. Trajectories and risk factors for post-traumatic stress symptoms following pediatric concussion. J Neurotrauma. 2017;34(14):2272–2279.
47. Swartzman S, Booth JN, Munro A, et al. Posttraumatic stress disorder after cancer diagnosis in adults: a meta-analysis. Depress Anxiety. 2017;34(4):327–339.
48. Markkula N, Lehti V, Gissler M, et al. Incidence and prevalence of mental disorders among immigrants and native Finns: a register-based study. Soc Psychiatry Psychiatr Epidemiol. 2017;52(12):1523–1540.
49. Kronick R. Mental health of refugees and asylum seekers: assessment and intervention. Can J Psychiatry. 2018;63(5):290–296.
50. de Barros ACS, Furlan AER, Marques LHN, et al. Gender differences in prevalence of psychiatric disorders, levels of alexithymia, and coping strategies in patients with refractory mesial temporal epilepsy and comorbid psychogenic nonepileptic seizures. Epilepsy Behav. 2018;82:1–5.
51. Wilhsson M, Svedberg P, Hogdin S, et al. Strategies of adolescent girls and boys for coping with school-related stress. J Sch Nurs. 2017;33(5):374–382.
52. Dyb G, Holen A, Braenne K, et al. Parent-child discrepancy in reporting children’s post-traumatic stress reactions after a traffic accident. Nord J Psychiatry. 2003;57(5):339–344.
53. Kjaergard LL, Villumsen J, Gluud C. Reported methodologic quality and discrepancies between large and small randomized trials in meta-analyses. Ann Intern Med. 2001;135(11):982–989.
54. Dai W, Liu A, Kaminga AC, et al. Prevalence of acute stress disorder among road traffic accident survivors: a meta-analysis. BMC Psychiatry. 2018;18(1):188.
55. Savovic J, Turner RM, Mawdsley D, et al. Association between risk-of-bias assessments and results of randomized trials in Cochrane reviews: the ROBES meta-epidemiologic study. Am J Epidemiol. 2018;187(5):1113–1122.
Supplementary Material
Supplemental material files:
