- Email: [email protected]
Predicting Posttraumatic Stress Symptoms in Children After Road Traffic Accidents
Predicting Posttraumatic Stress Symptoms in Children After Road Traffic Accidents MARKUS A. LANDOLT, PH.D., MARGARETE VOLLRATH, PH.D., KARIN TIMM, M.D., HANSPETER E. GNEHM, M.D., AND FELIX H. SENNHAUSER, M.D.
ABSTRACT Objective: To prospectively assess the prevalence, course, and predictors of posttraumatic stress symptoms (PTSSs) in children after road traffic accidents (RTAs). Method: Sixty-eight children (6.5–14.5 years old) were interviewed 4–6 weeks and 12 months after an RTA with the Child PTSD Reaction Index (response rate 58.6%). Their mothers (n = 60) and fathers (n = 53) were assessed with the Posttraumatic Diagnostic Scale. Results: The prevalence of moderate to severe PTSSs in children was 16.2% at 4–6 weeks, and 17.6% at 12 months. Mean PTSS scores did not decrease between the two assessments. Five children showed a delayed onset of PTSSs. Twelve mothers (20%) and six fathers (11.3%) met criteria for posttraumatic stress disorder (PTSD) at 4 to 6 weeks. At 12 months, three mothers (5.7%) and no fathers met diagnostic criteria. Child PTSSs at 12 months was significantly predicted by PTSS at 4–6 weeks and by severity of father’s PTSD. Age, sex, injury severity, threat appraisal, and maternal PTSD did not significantly contribute to child PTSSs at follow-up. Conclusions: There is a need for careful psychological assessment of children and their parents after an RTA. The possibility of delayed onset of PTSSs implies a monitoring beyond the first weeks after the accident. The impact of fathers’ PTSD on child PTSSs suggests that fathers ought to be actively involved in family-based prevention and treatment interventions of child PTSSs after RTAs. J. Am. Acad. Child Adolesc. Psychiatry, 2005;44(12):1276–1283. Key Words: road traffic accident, injury, posttraumatic stress disorder.
In Western societies, road traffic accidents (RTAs) are among the most common traumatic events that children can face. There is now considerable evidence that children can suffer significant and long-lasting psychological distress following everyday RTAs. Notably, previous studies found that about 10% to 35% of trafficinjured children develop posttraumatic stress disorder (PTSD) or suffer from clinically relevant posttraumatic
Accepted July 26, 2005. Drs. Landolt and Sennhauser are with the University Children’s Hospital, Zurich, Switzerland; Dr. Vollrath is with the Division of Mental Health, Norwegian Institute of Public Health, Oslo, Norway; Dr. Timm is with the Children’s Hospital, St. Gallen, Switzerland; and Dr. Gnehm with the Children’s Hospital, Aarau, Switzerland. This study was funded by grants from the Gebert-Ruef-Foundation, the Hugo and Elsa-Isler-Foundation, and the Anna-Mueller-Grocholski-Foundation. Correspondence to Dr. Markus A. Landolt, University Children’s Hospital, Steinwiesstrasse 75, CH-8032 Zurich, Switzerland; e-mail: markus.landolt@ kispi.unizh.ch. 0890-8567/05/4412–12762005 by the American Academy of Child and Adolescent Psychiatry. DOI: 10.1097/01.chi.0000181045.13960.67
1276
stress symptoms (PTSSs; de Vries et al., 1999; Di Gallo et al., 1997; Ellis et al., 1998; Landolt et al., 2003; McDermott and Cvitanovich, 2000; Mirza et al., 1998; Stallard et al., 1998, 1999). Previous studies examining the predictors of PTSD in children after an RTA have identified many of the same risk factors that have been reported in the general childhood PTSD literature. Risk factors can be grouped into three categories: characteristics of the traumatic event and the injury, characteristics of the child, and characteristics of the child’s recovery environment. Regarding accident- and injury-related characteristics, injury severity and other medical characteristics have consistently been found to be unrelated to PTSSs and PTSD (Aaron et al., 1999; Bryant et al., 2004; Daviss et al., 2000; de Vries et al., 1999; Kassam-Adams and Winston, 2004; McDermott and Cvitanovich, 2000; Mirza et al., 1998; Stallard et al., 1998). Thus, even minor crashes can evoke marked distress. Recovery from physical injuries may be important. Two studies indicate that children who have not physically recovered 5–8 months postaccident were
J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 44:12, DECEMB ER 2005
PTSD IN CH ILD REN AFTER TRAFFIC A CCID EN TS
at greater risk of psychological problems (Ellis et al., 1998; Stallard et al., 2001). Findings also indicate that the type of accident can be important. Mirza et al. (1998) found that children who were passengers in a motor vehicle had a higher risk of developing PTSD as compared with pedestrians or bicycle riders; however, Stallard et al. (1998) could not support this finding. With regard to child characteristics, several authors have found an association between PTSSs and pretraumatic functioning (Aaron et al., 1999; Di Gallo et al., 1997; Mirza et al., 1998) and previous experience of trauma (Stallard et al., 1998). Also, female sex has been debated as a risk factor, and findings have been inconsistent (Aaron et al., 1999; Daviss et al., 2000; de Vries et al., 1999; Di Gallo et al., 1997; Ehlers et al., 2003; Mirza et al., 1998; Stallard et al., 1998). Child age seems to be unrelated to PTSSs (Aaron et al., 1999; Daviss et al., 2000; Kassam-Adams and Winston, 2004; Mirza et al., 1998). Finally, the important role of cognitions in predicting PTSD has consistently been demonstrated. Most studies confirmed that the child’s subjective perceptions of life threat or threat to the physical integrity were significant predictors of PTSSs (Aaron et al., 1999; Di Gallo et al., 1997; Ehlers et al., 2003; McDermott and Cvitanovich, 2000; Stallard et al., 1998). The role of the child’s recovery environment in the aftermath of an RTA has only been studied rarely. Daviss et al. (2000) found that children with accidental injuries were at higher risk of PTSD if the parents showed signs of acute distress after the incident. Stallard et al. (2001) noted that the relationship between child RTA victims and their caregivers can be complex, particularly if a parent was involved in the accident or even caused it. Notably, two studies have shown that parental involvement in the accident is an important predictor of poor psychological adjustment in the child (de Vries et al., 1999; Ellis et al., 1998); however, a third study did not support this finding (Mirza et al., 1998). Few studies among child RTA victims specifically examined the association of child and parental PTSD, and most studies only assessed the mothers. de Vries et al. (1999) and Winje and Ulvik (1998) both found mothers’ and children’s symptoms to be related. Conversely, Bryant et al. (2004) found no association between mother’s and child’s diagnosis of PTSD, irrespective of the mother’s involvement in the accident. To date, the role of the father in the child’s recovery has only been studied in a small sample by Winje and Ulvik (1998). These
authors found no associations between child and paternal PTSD. Taken together, studies on predictors of PTSSs in child RTA victims have focused on characteristics of the accident or injury and on characteristics of the child. The role of the family as the most important recovery environment has not been studied prospectively. This is astonishing because high levels of parental distress, along with poor family climate, are regarded as significant impediments to the child’s adjustment after a traumatic event (Green et al., 1991). There are some data on the importance of mothers. To date, the specific role of fathers in the aftermath of an RTA has not been studied prospectively. Aim of the Study
The purpose of this study was twofold: 1. We aimed at prospectively assessing the prevalence and course of PTSSs in children 4–6 weeks and 12 months after an RTA. Based on previous findings, we expected to find clinically significant PTSSs in about 20%–25% of the sample, with a decrease over time. 2. We aimed at investigating accident- and injuryrelated individual and parental predictors of PTSSs in children after an RTA. In accordance with earlier studies, we hypothesized that injury- and accidentrelated variables would not predict PTSSs. It was expected that greater perceived threat and the presence of PTSD in mothers and fathers would increase the rates of PTSSs in children. METHOD Participants Participants were assessed in four children’s hospitals in the Germanspeaking part of Switzerland. Children were consecutively recruited over a period of 24 months. Parents and children were asked to participate in the study within the first 2 weeks after occurrence of an RTA if the following criteria were met: (1) hospitalization of at least 24 hours; (2) no severe head trauma; (3) age between 6.5 and 14.5 years; (4) fluency in German; and (5) no evidence of mental retardation (physician’s rating). Of 116 patients who met these criteria, 38 (12 girls, 26 boys; mean age = 10.42 years) did not participate, mainly because the study seemed too time-consuming. Because of incomplete data at either of the two assessments, 10 children were excluded from further analyses. The final sample comprised 68 children from different families (58.6% response rate). Response rate in parents of the selected children was lower, with 60 mothers and 53 fathers participating in the study. Thus, 53 complete data sets from child, mother, and father were available. In seven cases, data from
J. AM. ACAD. CH ILD ADO LE SC. PSY CH IAT RY, 44:12, DE CEM BER 2 005
1277
LAND OL T ET A L.
child and mother and in eight cases, only data from children were accessible. None of the children received psychological treatment during the hospital stay. There were no significant differences between study completers and those who did not participate with regard to age (t = –1.61, p = .11) and sex (x2 = 1.98, p = .16). Similarly, those who dropped out after the first assessment did not differ from those who completed assessment in terms of age (t = 0.84, p = .40), injury severity (t = 0.12, p = .91), and PTSSs (t = –1.27, p = .25). Girls were overrepresented among completers (x2 = 4.56, p = .03).
Measures Child PTSSs. Child PTSSs were measured by the Child PTSD Reaction Index (RI) (Frederick et al., 1992), administered by a structured interview. The RI contains 20 items that address school-age children’s fears following a traumatic event, symptoms relating to memories of the event, avoidance, and general functioning. The frequency of each symptom is rated on a 5-point Likert scale ranging from none of the time (0) to most of the time (4). A total PTSS score is obtained by summing across all items. The RI does not provide a PTSD diagnosis according to the DSM-IV-TR (American Psychiatric Association, 2000), but there is a scoring system that establishes levels of PTSD (Pynoos et al., 1993). A score of 12–24 indicates a mild level, a score of 25–39 indicates a moderate level, a score of 40–59 indicates a severe level, and a score of >60 indicates a very severe level of PTSD. Scores >24 are interpreted as clinically relevant. The internal consistency of the RI is good (Nader et al., 1990), and there are strong associations of RI scores with DSM diagnoses of PTSD (McDermott and Cvitanovich, 2000; Pynoos et al., 1993). In this study, a German version of the RI was used. The translation procedure followed published guidelines, including the use of independent back-translation (Brislin et al., 1973). The RI achieved good internal consistency coefficients in this sample (a = .82 at 4–6 weeks, and a = .79 at 12 months). Parent PTSSs. PTSD of mothers and fathers was assessed by the Posttraumatic Diagnostic Scale (PDS; Foa et al., 1997). This selfreport measure yields a PTSD diagnosis according to DSM-IV-TR criteria and a rating of PTSD severity. Participants were asked to rate the presence of each of the 17 symptoms of PTSD on a 4-point Likert scale ranging from not at all (0) to very much (3). In its original English version, the PDS has demonstrated high internal consistency (a = .92) and good test-retest reliability (k = 0.74; Foa et al., 1997). Agreement between PTSD diagnoses obtained from the PDS and the Structured Clinical Interview for DSM-III-R SCID-PTSD module was 82%. The sensitivity of the PDS was 0.89 and its specificity was 0.75. The present study used the German version of the PDS, which was validated previously (Steil and Ehlers, 2000). Internal consistency in this sample was found to be a = .88 for mothers and a = .89 for fathers. In accordance with the DSM-IV-TR, PTSD was diagnosed if participants reported at least one reexperiencing symptom, three avoidance symptoms, two arousal symptoms, and impairment in one domain of life. A diagnosis of partial PTSD required at least one symptom in each of the three PTSD symptom clusters (Stein et al., 1997). Threat Appraisals. Child accident-related threat appraisals were assessed by a single item derived from an appraisal scale that was previously validated in pediatric patients (Vollrath et al., 2004). On a 3-point Likert scale ranging from not very threatening (0) to rather threatening (2), children were asked to tell how threatening their accident had been.
1278
Life Events. The occurrence of 12 major life events, such as change of domicile, unemployment, or parental separation was assessed by using the mothers as informants. Appropriate events for our sample were selected from previously published life event scales (Paykel, 1983). A sum score was computed by counting the number of life events. Life events were retrospectively assessed at initial assessment for the 12 months preceding the accident and at follow-up for the time period following the accident. Injury Severity. Severity of injuries was classified by a physician using the Modified Injury Severity Scale (MISS), a highly reliable and widely accepted scale (Mayer et al., 1980). The MISS allows the rating of the severity of injuries in different bodily systems. MISS values range from 1 to 75 (death), with scores >25 indicating a severe injury. Socioeconomic Status. Socioeconomic status was calculated by means of a score reflecting paternal occupation and maternal education (range 2–12 points). Three social classes were defined as follows: scores 2–5, lower class; scores 6–8, middle class; scores 9–12, upper class. This measure has been shown to be a reliable and valid indicator of socioeconomic status in our community (Landolt et al., 2002a). Procedure The study was approved by the ethics committees of all involved hospitals. Written informed consent was obtained from parents. Assessments were conducted 4–6 weeks after the child’s RTA (T1) and at 12 months (T2). None of the injured patients was in a lifethreatening condition after the accident. The children were assessed by means of a standardized face-to-face interview conducted by trained graduate students of psychology. The interviews lasted 30 to 60 minutes. Most interviews were conducted in the participants’ homes; some were conducted at the hospital. To ensure that children could openly express their own account of the accident, they were interviewed when possible without their parents present. Parents received their questionnaires at the same time and were asked to complete them separately and to return them by mail. Demographic and medical variables were retrieved from the patients’ hospital records. Statistical Analyses Data were analyzed using the statistical package SPSS, release 11 (SPSS Inc., Chicago, IL). Analyses were performed with twosided tests. x2 analyses were used to compare nominal variables. Kolmogorov-Smirnov tests revealed positively skewed distributions for the child RI scores at T1 and PDS scores of fathers at T1. Therefore, Spearman-Brown rank correlation tests were used to examine associations between child PTSSs and various predictors. Wilcoxon tests for pairwise comparisons were calculated to compare RI scores and PDS scores at T1 and T2. To examine prediction of children’s PTSSs after 1 year, linear multiple regression analysis was used. All complete 53 data sets were included in this analysis, with child RI at T2 as the dependent variable. The following independent variables were chosen on the basis of previous knowledge and significant correlations with the dependent variable: age, sex, threat appraisal at T1, number of preceding life events, MISS score, maternal PDS score at T1, paternal PDS score at T1, child RI at T1. PDS and RI scores were successfully normalized by means of square root transformation before using them in the regression analysis.
J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 44:12, DECEMB ER 2005
PTSD IN CH ILD REN AFTER TRAFFIC A CCID EN TS
RESULTS Characteristics of the Sample
Descriptive information about the sample is contained in Table 1. Of the sample, 85.3% was of Swiss origin, and 8.8% originated from a Mediterranean country (Italy, Spain, former Yugoslavia). Most families were from the upper or middle class, probably on account of the language requirement (many non-Swiss nationals are from the lower classes). Fifty-seven children (83.8%) lived with both biological parents. The children had received minor head injuries (58.8%), lower extremity fractures (16.2%), upper extremity fractures (14.7%), nonextremity fractures (36.8%), and internal injuries (16.2%); 48.5% of the children had combined injuries. The mean MISS score was 10.21 (SD = 7.58; median = TABLE 1 Demographic, Accident, and Medical Characteristics of the Sample (N = 68) Age at accident, yr Mean (SD) 9.82 (2.55) Median 9.38 Sex Female 31 (45.6%) Male 37 (54.4%) Socioeconomic status Lower 6 (8.8%) Middle 36 (52.9%) Upper 23 (33.8%) Unknown 3 (4.4%) Living with both 57 (83.8%) biological parents No. of preceding life events, mean (SD) 1.27 (1.26) No. of postaccident life events, mean (SD) 1.02 (1.71) Type of accident Car passenger 10 (14.7%) Bicycle/motorcycle 21 (30.9%) Pedestrian 30 (44.1%) Other 7 (10.3%) Parent involved 9 (13.2%) Modified Injury Severity Scale score Mean (SD) 10.21 (7.58) Median 9.00 Days in hospital, T1 Mean (SD) 8.09 (7.36) Median 5.50 Days in hospital, T2 Mean (SD) 10.05 (14.70) Median 5.50
9.0). Three children had incurred severe injuries (MISS >25). Two children (2.9%) were still hospitalized at initial assessment. Most children were injured either as pedestrians or as bicycle riders. In nine cases (13.2%), at least one parent was also involved in the accident. Prevalence and Course of PTSSs in Children
Descriptive statistics for RI scores at T1 and T2 are provided in Table 2. Of the 68 children, 11 (16.2%) had scores in the clinical range of the RI at T1, thus showing moderate to severe reactions. At 12-month follow-up, 12 children (17.6%) scored in the clinical range. Five of the latter showed a delayed onset of symptoms, thus scoring in the moderate to severe range only at 12 months, but not at 4–6 weeks. Four children with RI scores in the clinical range at T1 showed a decrease below the cutoff score at T2. On average, children reported symptoms in the mild range of PTSD at both T1 and T2. Analysis of RI items revealed that the most common symptoms reported were feelings of guilt, intrusive recollections of the accident, distress at exposure to cues that resemble aspects of the trauma, and acting or feeling as if the traumatic event was reoccurring. There was no significant decrease of mean RI scores between the two assessment points. This was also true if the five patients with delayed onset were excluded from analysis (Wilcoxon Z = –1.34, p = .18). At T2 parents of 12 children (17.6%) reported that their child had received some kind of outpatient psychological treatment following the accident. At T1 two of these children and at T2 five of these children showed PTSSs in the clinical range. Prevalence and Course of PTSD in Parents
Descriptive statistics for parental PTSD severity at T1 and T2 are provided in Table 2. Twelve mothers (20%) and six fathers (11.3%) met criteria for full PTSD at T1. An additional 37 mothers (61.7%) and 24 fathers (45.3%) suffered from partial PTSD. Prevalence of PTSD significantly decreased over time, with three mothers (5.7%) and no father meeting diagnostic criteria at 12 months. In seven mothers (13.2%) and six fathers (11.3%), partial PTSD was found at T2. Maternal PTSD at T1 and T2 was not associated with involvement in the accident (T1: x2 = 0.24, p = .62; T2: x2 = 1.87, p = .17). The same was true for PTSD of fathers at T1 (x2 = 0.12, p = .72). Severity of maternal PTSD at T2 was significantly associated with severity of PTSD
J. AM. ACAD. CH ILD ADO LE SC. PSY CH IAT RY, 44:12, DE CEM BER 2 005
1279
LAND OL T ET A L.
TABLE 2 Descriptive Statistics for Main Variables at 4–6 Weeks (T1) and at 12 Months (T2) T1 T2 Child RI Maternal PTSD severity Paternal PTSD severity
Mean
SD
Mean
SD
No.
Wilcoxon Z
p
14.44 11.55 7.46
10.98 8.13 6.73
14.02 3.34 2.00
10.01 4.31 2.67
68 60 53
–0.32 –5.93 –5.04
.75 <.0005 <.0005
Note: PTSD = posttraumatic stress disorder.
at T1 (r = 0.66, p < .001) and preceding life events (r = 0.32, p = .02), but not with length of hospitalization (r = 0.25, p = .09), injury severity (r = 0.11, p = .45), age (r = 0.05, p = .73), and sex of the child (r = 0.12, p = .38). Severity of fathers’ PTSD was significantly correlated with PTSD severity at T1 (r = 0.52, p < .001), length of hospitalization (r = 0.33, p = .03) and preceding life events (r = 0.33, p = .02), but not with injury severity (r = 0.19, p = .18), age (r = 0.03, p = .82), and sex of the child (r = –0.11, p = .46). Predictors of PTSSs in Children
Spearman correlations of RI scores at T1 and T2 with predictor variables are presented in Table 3. Child threat appraisals at T1 were significantly correlated with RI scores at T1 and T2. Notably, at both time points, TABLE 3 Spearman Correlation Between Predictor Variables and Child RI at 4–6 Weeks (T1) and 12 Months (T2) Reaction Reaction Index T1 Index T2 Age Sex Socioeconomic status Preceding life events Postaccident life events Modified Injury Severity Scale score Days in hospital, T1 Days in hospital, T2 Parent involved in accident Child appraisal of threat, T1 Maternal PTSD severity, T1 Paternal PTSD severity, T1 Child RI, T1
0.07 0.01 –0.02 0.14 –0.20
0.09 0.09 –0.03 0.25* –0.09
–0.05 –0.02
0.02 0.00 –0.05
0.03
0.21
0.37*** 0.06 0.01
0.29** 0.24 0.31* 0.64***
Note: 68 ‡ n ‡ 53; *p < .05; **p < .01; ***p < .001.
1280
sociodemographic variables such as age, sex, and socioeconomic status were not significantly associated with RI scores. The same was true for injury- and accident-related predictors, such as injury severity, length of hospital stay, and parental involvement in the accident. In addition, Kruskal-Wallis tests revealed that RI scores at T1 and T2 did not differ significantly with regard to the type of accident (T1: x2 = 0.54, p = .76; T2: x2 = 4.74, p = .09). Life events occurring after the accident were not related to RI scores at T1 and T2, but life events preceding the accident were related to RI scores at T2. Whereas maternal PTSD severity was not associated with RI scores, paternal PTSD severity was significantly correlated with child PTSSs at T2. Finally, RI scores at T1 were significantly correlated with RI scores at 12 months, indicating a substantial stability of PTSSs over time. Table 4 summarizes statistics for the regression model predicting child PTSSs at 12 months. Overall, the selected variables accounted for 52% of the variance in the RI scores, making the regression model highly significant. Child symptomatology at T1 and severity of TABLE 4 Regression Analysis Predicting Child PTSD Severity (RI Score) at 12 Months Variable Age, yr Sex Modified Injury Severity Scale score Preceding life events Threat appraisal, T1 Child RI score, T1 Maternal PDS score, T1 Paternal PDS score, T1
b
Significance b
Partial r
–.09 .08
.37 .44
–0.14 0.12
–.07 .11 .16 .57 .11 .31
.53 .36 .19 <.0005 .30 .01
–0.10 0.14 0.20 0.64 0.16 0.37
Note: n = 53; F = 8.065; p < .0005; R2 = 0.60, R2 adj. = 0.52. PTSD = posttraumatic stress disorder; PDS = Posttraumatic Diagnostic Scale.
J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 44:12, DECEMB ER 2005
PTSD IN CH ILD REN AFTER TRAFFIC A CCID EN TS
paternal PTSD at T1 both contributed with significant b weights. Notably, age, sex, number of preceding life events, injury severity, child threat appraisal, and maternal PDS score at T1 did not significantly contribute to the prediction of child PTSSs at 12 months. DISCUSSION
This prospective study found that clinically significant PTSSs occur in a small but significant portion of school-age children as a consequence of RTAs, thus confirming our initial hypothesis. The prevalence of moderate to severe PTSSs was 16.2% at 4–6 weeks, and 17.6% at 12-month follow-up. This is consistent with earlier findings that 10%–35% of traffic-injured children have PTSD or clinically relevant PTSSs (de Vries et al., 1999; Di Gallo et al., 1997; Landolt et al., 2003; McDermott and Cvitanovich, 2000; Mirza et al., 1998; Stallard et al., 1998, 1999). Also, PTSD was common in mothers and fathers, which supports the notion that witnesses of traumatic events and caregivers are at increased risk of posttraumatic maladjustment (Landolt et al., 2002b, 2005). Unexpectedly, based on mean RI scores, we did not find any decrease in severity of child PTSSs over time. This is probably the result of lower prevalence rates of moderate to severe PTSSs at T1 in our study as compared with other studies (de Vries et al., 1999; Di Gallo et al., 1997; Stallard et al., 1998). In this study, most children either had no PTSSs or recovered psychologically during the first weeks after the accident. Those who showed clinically significant PTSSs at 4–6 weeks had a high risk of maintaining their symptoms over a longer period of time; however, the course of symptoms was different in individual patients. Four of the 11 children (36%) with clinical levels of PTSSs at T1 showed a decrease below the cutoff score by T2. Conversely, five children had a delayed onset of symptoms, thus showing that this phenomenon is also present in children (Gillies et al., 2003). This study also examined the importance of several predictors of PTSSs in children after RTAs. Multiple regression analysis revealed two significant predictors of child PTSSs at 12 months: child PTSSs at 4–6 weeks and paternal PTSD severity at 4–6 weeks. The importance of early PTSSs for long-term outcome in children has been demonstrated in a previous study with RTA victims (Mirza et al., 1998). Also, findings by Daviss
et al. (2000) in injured children suggest an association between acute distress and subsequent PTSSs. Although the importance of early symptomatology has been demonstrated, this is the first study to show the importance of fathers in the psychological recovery of the child after an RTA. Severity of fathers’ PTSD at T1 contributed significantly to the prediction of child PTSSs at T2. The fact that fathers’ and children’s PTSSs were not associated at T1 and that paternal PTSSs had decreased significantly by T2 suggests that fathers’ distress symptoms in the early period after the accident affect the child in the long term. In contradiction to our hypothesis, mothers’ PTSD did not affect the children’s PTSSs at either time point. As noted earlier, previous studies tended to solely focus on the role of mothers, and fathers were neglected. The only study examining this issue found no evidence of the impact of fathers’ PTSD on child PTSSs (Winje and Ulvik, 1998); however, the sample in that study was small and included several sibling pairs, thus implying dependency of observation. Also, the assessment of PTSSs did not include hyperarousal symptoms. Interestingly, indirect support of our findings can be derived from a study in children with type 1 diabetes that found a positive relationship between children’s and fathers’ adjustment (Chaney et al., 1997). Why are fathers important for the posttraumatic adjustment of their children? One may speculate that children react more strongly to their fathers’ symptoms because this is less common for them. In Western society, fathers are expected to be strong and to handle difficult situations without displaying excessive emotions. Also, a father with PTSD may have difficulties supporting his wife, which in turn may negatively affect the child. Finally, in our sample, mainly fathers were responsible for the family income. A particularly distressed father can have difficulties at work and may therefore threaten the socioeconomic security of the family. This in turn may affect the child. Whereas in this study child threat appraisals were significantly associated with PTSSs at 12 months in the bivariate analyses, multiple regression analysis revealed no significant impact of threat appraisals on PTSSs at 12 months. This result is in contrast to our hypothesis and previous findings. Compared with other studies, however, the time lag of 11 months between the assessment of threat appraisals at initial assessment and PTSSs at follow-up is longer in our study. It is possible that in
J. AM. ACAD. CH ILD ADO LE SC. PSY CH IAT RY, 44:12, DE CEM BER 2 005
1281
LAND OL T ET A L.
children the impact of specific cognitions on PTSSs may not last as long as they do in adults. With regard to accident- and injury-related variables, our hypothesis was confirmed. We found that injury severity and other medical characteristics were unrelated to PTSSs, which is consistent with previous studies (Aaron et al., 1999; Bryant et al., 2004; Daviss et al., 2000; de Vries et al., 1999; Kassam-Adams and Winston, 2004; McDermott and Cvitanovich, 2000; Mirza et al., 1998; Stallard et al., 1998). In accordance with Stallard et al. (1998), our findings also showed that the type of accident was not related to the development of PTSSs. We could therefore not confirm a finding by Mirza et al. (1998) that children who were passengers in a motor vehicle sustaining an accident had a higher risk of PTSD. Moreover, in our study, parental involvement in the accident was not associated with child PTSSs. Although this result is in line with Mirza et al. (1998), it contradicts other studies (de Vries et al., 1999; Ellis et al., 1998). Presumably, the role of parental involvement with regard to child PTSSs is more complex and depends on the type of accident. One may also speculate that parental guilt may act as a moderating variable. This should be analyzed in further studies. Limitations
Some limitations of this study merit note. First, the measures used were aimed at PTSD and PTSSs. Inclusion of standardized instruments assessing symptoms of depression and anxiety in children and parents would have been useful. Second, in our study, assessment methods in children and parents were different: Children were interviewed, whereas parents were assessed by questionnaire. This may have biased our findings in an unknown way. Third, this study did not assess data on premorbid functioning of children and parents. We therefore may have missed an important additional predictor of PTSSs and PTSD. Also, because this study focused on child PTSSs, treatment interventions for parents’ PTSD during the study period were not assessed. A final issue potentially limiting generalizability of our findings is the participation rate of 59%. We have only basic sociodemographic information about nonparticipants. It is possible that some of the nonparticipants declined participation because of avoidance symptoms of PTSD or because they were well adjusted and the study was not relevant to them. Both cases would affect prevalence estimates of PTSD. Our participation rate is consistent
1282
with those of other studies that include comprehensive psychosocial evaluations, however. Also, comparisons of completers with nonparticipants and dropouts revealed differences only with regard to sex. Because we found no differences between boys and girls regarding PTSSs, it is unlikely that this influenced our findings. Clinical Implications
Despite these limitations, the present study has several strengths, not the least of which are its longitudinal design and its independent and standardized assessment of multiple informants. There are several clinical implications that can be drawn from our findings. Certainly, children after RTAs and their parents have to be assessed carefully with regard to PTSSs and PTSD. Delayed onset of symptoms in some children implies a monitoring beyond the first weeks after the accident. Emergency departments and hospitals could provide information on psychotraumatological issues to children and parents (e.g., by use of leaflets) and ensure that general practitioners are informed about particularly distressed children and families (Bryant et al., 2004). Psychological interventions need to be developed to prevent and treat psychological consequences of RTAs in children and if necessary in their parents. Such interventions should be offered to children irrespective of the severity of the injury. To date, such interventions are rarely made available (Gillies et al., 2003). In addition, our findings underscore that fathers need to be more actively involved in family-based prevention and treatment interventions of child PTSSs. Fathers play a specific, yet not well-understood role in child trauma recovery that needs to be studied in more detail. Disclosure: The authors have no financial relationships to disclose.
REFERENCES American Psychiatric Association (2000), Diagnostic and Statistical Manual of Mental Disorders, DSM-IV-TR, Text Revision. Washington, DC: American Psychiatric Association Aaron J, Zaglul H, Emery RE (1999), Posttraumatic stress in children following acute physical injury. J Pediatr Psychol 24:335–343 Brislin RW, Lonner WJ, Thorndike RM (1973), Cross-cultural Research Methods. New York: Wiley Bryant B, Mayou R, Wiggs L, Ehlers A, Stores G (2004), Psychological consequences of road traffic accidents for children and their mothers. Psychol Med 34:335–346 Chaney JM, Mullins LL, Frank RG, Peterson L, Mace LD, Kashani JH (1997), Transactional patterns of child, mother, and father adjustment in insulin-dependent diabetes mellitus: a prospective study. J Pediatr Psychol 22:229–244
J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 44:12, DECEMB ER 2005
PTSD IN CH ILD REN AFTER TRAFFIC A CCID EN TS
Daviss WB, Mooney D, Racusin R, Ford JD, Fleischer A, McHugo GJ (2000), Predicting posttraumatic stress after hospitalization for pediatric injury. J Am Acad Child Adolesc Psychiatry 39:576–583 de Vries APJ, Kassam-Adams N, Cnaan A, Sherman-Slate E, Gallagher PR, Winston FK (1999), Looking beyond the physical injury: posttraumatic stress disorder in children and parents after pediatric traffic injury. Pediatrics 104:1293–1299 Di Gallo A, Barton J, Parry-Jones W (1997), Road traffic accidents: early psychological consequences in children and adolescents. Br J Psychiatry 170:358–362 Ehlers A, Mayou R, Bryant B (2003), Cognitive predictors of posttraumatic stress disorder in children: results of a prospective longitudinal study. Behav Res Ther 41:1–10 Ellis A, Stores G, Mayou R (1998), Psychological consequences of road traffic accidents in children. Eur Child Adolesc Psychiatry 7:61–68 Foa EB, Cashman L, Jaycox L, Perry K (1997), The validation of a self-report measure of posttraumatic stress disorder: the Posttraumatic Diagnostic Scale. Psychol Assess 9:445–451 Frederick C, Pynoos R, Nader K (1992), Childhood Posttraumatic Stress Reaction Index—A Copyrighted Instrument. Los Angeles: University of California Gillies ML, Barton J, Di Gallo A (2003), Follow-up of young road accident victims. J Trauma Stress 16:523–526 Green BL, Korol M, Grace MC, Vary MG, Leonard AC, Gleser GC (1991), Children and disaster: age, gender, and parental effects on PTSD symptoms. J Am Acad Child Adolesc Psychiatry 30:945–951 Kassam-Adams N, Winston FK (2004), Predicting child PTSD: the relationship between acute stress disorder and PTSD in injured children. J Am Acad Child Adolesc Psychiatry 43:403–411 Landolt MA, Nuoffer JM, Steinmann B, Superti-Furga A (2002a), Quality of life and psychologic adjustment in children and adolescents with early treated phenylketonuria can be normal. J Pediatr 140:516–521 Landolt MA, Ribi K, Laimbacher J, Vollrath M, Gnehm HE, Sennhauser FH (2002b), Posttraumatic stress disorder in parents of children with newly diagnosed type 1 diabetes. J Pediatr Psychol 27:647–652 Landolt MA, Vollrath M, Laimbacher J, Gnehm HE, Sennhauser FH (2005), Prospective study of posttraumatic stress disorder in parents of children with newly diagnosed type 1 diabetes. J Am Acad Child Adolesc Psychiatry 44:682–689
Landolt MA, Vollrath M, Ribi K, Gnehm HE, Sennhauser FH (2003), Incidence and associations of parental and child posttraumatic stress symptoms in pediatric patients. J Child Psychol Psychiatry 44:1199–1207 Mayer T, Matlak ME, Johnson DG, Walker ML (1980), The Modified Injury Severity Scale in pediatric multiple trauma patients. J Pediatr Surg 15:719–726 McDermott BM, Cvitanovich A (2000), Posttraumatic stress disorder and emotional problems in children following motor vehicle accidents: an extended case series. Aust N Z J Psychiatry 34:446–452 Mirza KAH, Bdadrinath BR, Goodyer IM, Gilmour C (1998), Posttraumatic stress disorder in children and adolescents following road traffic accidents. Br J Psychiatry 172:443–447 Nader K, Pynoos R, Fairbanks L, Frederick C (1990), Children’s PTSD reactions one year after a sniper attack at their school. Am J Psychiatry 147:1526–1530 Paykel ES (1983), Methodological aspects of life event research. J Psychosom Res 27:341–352 Pynoos RS, Goenjian A, Tashjain M, Karakashian M, Manjikian R, Manoukian R (1993), Post-traumatic stress reactions in children after the 1988 Armenian earthquake. Br J Psychiatry 163:239–247 Stallard P, Velleman R, Baldwin S (1998), Prospective study of post-traumatic stress disorder in children involved in road traffic accidents. BMJ 317:1619–1623 Stallard P, Velleman R, Baldwin S (1999), Psychological screening of children for post-traumatic stress disorder. J Child Psychol Psychiatry 40:1075–1082 Stallard P, Velleman R, Baldwin S (2001), Children involved in road traffic accidents: psychological sequelae and issues for mental health services. Clin Child Psychol Psychiatry 6:23–39 Steil R, Ehlers A (2000), Posttraumatische Diagnoseskala (PDS). Jena, Germany: Universita¨t Jena Psychologisches Institut Stein MB, Walker JR, Hazen AL, Forde DR (1997), Full and partial posttraumatic stress disorder: findings from a community survey. Am J Psychiatry 154:1114–1149 Vollrath M, Landolt MA, Ribi K (2004), Illness appraisal in paediatric patients and their parents: a short scale. Swiss J Psychol 63:3–15 Winje D, Ulvik A (1998), Long-term outcome of trauma in children: the psychological consequences of a bus accident. J Child Psychol Psychiatry 39:635–642
J. AM. ACAD. CH ILD ADO LE SC. PSY CH IAT RY, 44:12, DE CEM BER 2 005
1283
ABSTRACT Objective: To prospectively assess the prevalence, course, and predictors of posttraumatic stress symptoms (PTSSs) in children after road traffic accidents (RTAs). Method: Sixty-eight children (6.5–14.5 years old) were interviewed 4–6 weeks and 12 months after an RTA with the Child PTSD Reaction Index (response rate 58.6%). Their mothers (n = 60) and fathers (n = 53) were assessed with the Posttraumatic Diagnostic Scale. Results: The prevalence of moderate to severe PTSSs in children was 16.2% at 4–6 weeks, and 17.6% at 12 months. Mean PTSS scores did not decrease between the two assessments. Five children showed a delayed onset of PTSSs. Twelve mothers (20%) and six fathers (11.3%) met criteria for posttraumatic stress disorder (PTSD) at 4 to 6 weeks. At 12 months, three mothers (5.7%) and no fathers met diagnostic criteria. Child PTSSs at 12 months was significantly predicted by PTSS at 4–6 weeks and by severity of father’s PTSD. Age, sex, injury severity, threat appraisal, and maternal PTSD did not significantly contribute to child PTSSs at follow-up. Conclusions: There is a need for careful psychological assessment of children and their parents after an RTA. The possibility of delayed onset of PTSSs implies a monitoring beyond the first weeks after the accident. The impact of fathers’ PTSD on child PTSSs suggests that fathers ought to be actively involved in family-based prevention and treatment interventions of child PTSSs after RTAs. J. Am. Acad. Child Adolesc. Psychiatry, 2005;44(12):1276–1283. Key Words: road traffic accident, injury, posttraumatic stress disorder.
In Western societies, road traffic accidents (RTAs) are among the most common traumatic events that children can face. There is now considerable evidence that children can suffer significant and long-lasting psychological distress following everyday RTAs. Notably, previous studies found that about 10% to 35% of trafficinjured children develop posttraumatic stress disorder (PTSD) or suffer from clinically relevant posttraumatic
Accepted July 26, 2005. Drs. Landolt and Sennhauser are with the University Children’s Hospital, Zurich, Switzerland; Dr. Vollrath is with the Division of Mental Health, Norwegian Institute of Public Health, Oslo, Norway; Dr. Timm is with the Children’s Hospital, St. Gallen, Switzerland; and Dr. Gnehm with the Children’s Hospital, Aarau, Switzerland. This study was funded by grants from the Gebert-Ruef-Foundation, the Hugo and Elsa-Isler-Foundation, and the Anna-Mueller-Grocholski-Foundation. Correspondence to Dr. Markus A. Landolt, University Children’s Hospital, Steinwiesstrasse 75, CH-8032 Zurich, Switzerland; e-mail: markus.landolt@ kispi.unizh.ch. 0890-8567/05/4412–12762005 by the American Academy of Child and Adolescent Psychiatry. DOI: 10.1097/01.chi.0000181045.13960.67
1276
stress symptoms (PTSSs; de Vries et al., 1999; Di Gallo et al., 1997; Ellis et al., 1998; Landolt et al., 2003; McDermott and Cvitanovich, 2000; Mirza et al., 1998; Stallard et al., 1998, 1999). Previous studies examining the predictors of PTSD in children after an RTA have identified many of the same risk factors that have been reported in the general childhood PTSD literature. Risk factors can be grouped into three categories: characteristics of the traumatic event and the injury, characteristics of the child, and characteristics of the child’s recovery environment. Regarding accident- and injury-related characteristics, injury severity and other medical characteristics have consistently been found to be unrelated to PTSSs and PTSD (Aaron et al., 1999; Bryant et al., 2004; Daviss et al., 2000; de Vries et al., 1999; Kassam-Adams and Winston, 2004; McDermott and Cvitanovich, 2000; Mirza et al., 1998; Stallard et al., 1998). Thus, even minor crashes can evoke marked distress. Recovery from physical injuries may be important. Two studies indicate that children who have not physically recovered 5–8 months postaccident were
J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 44:12, DECEMB ER 2005
PTSD IN CH ILD REN AFTER TRAFFIC A CCID EN TS
at greater risk of psychological problems (Ellis et al., 1998; Stallard et al., 2001). Findings also indicate that the type of accident can be important. Mirza et al. (1998) found that children who were passengers in a motor vehicle had a higher risk of developing PTSD as compared with pedestrians or bicycle riders; however, Stallard et al. (1998) could not support this finding. With regard to child characteristics, several authors have found an association between PTSSs and pretraumatic functioning (Aaron et al., 1999; Di Gallo et al., 1997; Mirza et al., 1998) and previous experience of trauma (Stallard et al., 1998). Also, female sex has been debated as a risk factor, and findings have been inconsistent (Aaron et al., 1999; Daviss et al., 2000; de Vries et al., 1999; Di Gallo et al., 1997; Ehlers et al., 2003; Mirza et al., 1998; Stallard et al., 1998). Child age seems to be unrelated to PTSSs (Aaron et al., 1999; Daviss et al., 2000; Kassam-Adams and Winston, 2004; Mirza et al., 1998). Finally, the important role of cognitions in predicting PTSD has consistently been demonstrated. Most studies confirmed that the child’s subjective perceptions of life threat or threat to the physical integrity were significant predictors of PTSSs (Aaron et al., 1999; Di Gallo et al., 1997; Ehlers et al., 2003; McDermott and Cvitanovich, 2000; Stallard et al., 1998). The role of the child’s recovery environment in the aftermath of an RTA has only been studied rarely. Daviss et al. (2000) found that children with accidental injuries were at higher risk of PTSD if the parents showed signs of acute distress after the incident. Stallard et al. (2001) noted that the relationship between child RTA victims and their caregivers can be complex, particularly if a parent was involved in the accident or even caused it. Notably, two studies have shown that parental involvement in the accident is an important predictor of poor psychological adjustment in the child (de Vries et al., 1999; Ellis et al., 1998); however, a third study did not support this finding (Mirza et al., 1998). Few studies among child RTA victims specifically examined the association of child and parental PTSD, and most studies only assessed the mothers. de Vries et al. (1999) and Winje and Ulvik (1998) both found mothers’ and children’s symptoms to be related. Conversely, Bryant et al. (2004) found no association between mother’s and child’s diagnosis of PTSD, irrespective of the mother’s involvement in the accident. To date, the role of the father in the child’s recovery has only been studied in a small sample by Winje and Ulvik (1998). These
authors found no associations between child and paternal PTSD. Taken together, studies on predictors of PTSSs in child RTA victims have focused on characteristics of the accident or injury and on characteristics of the child. The role of the family as the most important recovery environment has not been studied prospectively. This is astonishing because high levels of parental distress, along with poor family climate, are regarded as significant impediments to the child’s adjustment after a traumatic event (Green et al., 1991). There are some data on the importance of mothers. To date, the specific role of fathers in the aftermath of an RTA has not been studied prospectively. Aim of the Study
The purpose of this study was twofold: 1. We aimed at prospectively assessing the prevalence and course of PTSSs in children 4–6 weeks and 12 months after an RTA. Based on previous findings, we expected to find clinically significant PTSSs in about 20%–25% of the sample, with a decrease over time. 2. We aimed at investigating accident- and injuryrelated individual and parental predictors of PTSSs in children after an RTA. In accordance with earlier studies, we hypothesized that injury- and accidentrelated variables would not predict PTSSs. It was expected that greater perceived threat and the presence of PTSD in mothers and fathers would increase the rates of PTSSs in children. METHOD Participants Participants were assessed in four children’s hospitals in the Germanspeaking part of Switzerland. Children were consecutively recruited over a period of 24 months. Parents and children were asked to participate in the study within the first 2 weeks after occurrence of an RTA if the following criteria were met: (1) hospitalization of at least 24 hours; (2) no severe head trauma; (3) age between 6.5 and 14.5 years; (4) fluency in German; and (5) no evidence of mental retardation (physician’s rating). Of 116 patients who met these criteria, 38 (12 girls, 26 boys; mean age = 10.42 years) did not participate, mainly because the study seemed too time-consuming. Because of incomplete data at either of the two assessments, 10 children were excluded from further analyses. The final sample comprised 68 children from different families (58.6% response rate). Response rate in parents of the selected children was lower, with 60 mothers and 53 fathers participating in the study. Thus, 53 complete data sets from child, mother, and father were available. In seven cases, data from
J. AM. ACAD. CH ILD ADO LE SC. PSY CH IAT RY, 44:12, DE CEM BER 2 005
1277
LAND OL T ET A L.
child and mother and in eight cases, only data from children were accessible. None of the children received psychological treatment during the hospital stay. There were no significant differences between study completers and those who did not participate with regard to age (t = –1.61, p = .11) and sex (x2 = 1.98, p = .16). Similarly, those who dropped out after the first assessment did not differ from those who completed assessment in terms of age (t = 0.84, p = .40), injury severity (t = 0.12, p = .91), and PTSSs (t = –1.27, p = .25). Girls were overrepresented among completers (x2 = 4.56, p = .03).
Measures Child PTSSs. Child PTSSs were measured by the Child PTSD Reaction Index (RI) (Frederick et al., 1992), administered by a structured interview. The RI contains 20 items that address school-age children’s fears following a traumatic event, symptoms relating to memories of the event, avoidance, and general functioning. The frequency of each symptom is rated on a 5-point Likert scale ranging from none of the time (0) to most of the time (4). A total PTSS score is obtained by summing across all items. The RI does not provide a PTSD diagnosis according to the DSM-IV-TR (American Psychiatric Association, 2000), but there is a scoring system that establishes levels of PTSD (Pynoos et al., 1993). A score of 12–24 indicates a mild level, a score of 25–39 indicates a moderate level, a score of 40–59 indicates a severe level, and a score of >60 indicates a very severe level of PTSD. Scores >24 are interpreted as clinically relevant. The internal consistency of the RI is good (Nader et al., 1990), and there are strong associations of RI scores with DSM diagnoses of PTSD (McDermott and Cvitanovich, 2000; Pynoos et al., 1993). In this study, a German version of the RI was used. The translation procedure followed published guidelines, including the use of independent back-translation (Brislin et al., 1973). The RI achieved good internal consistency coefficients in this sample (a = .82 at 4–6 weeks, and a = .79 at 12 months). Parent PTSSs. PTSD of mothers and fathers was assessed by the Posttraumatic Diagnostic Scale (PDS; Foa et al., 1997). This selfreport measure yields a PTSD diagnosis according to DSM-IV-TR criteria and a rating of PTSD severity. Participants were asked to rate the presence of each of the 17 symptoms of PTSD on a 4-point Likert scale ranging from not at all (0) to very much (3). In its original English version, the PDS has demonstrated high internal consistency (a = .92) and good test-retest reliability (k = 0.74; Foa et al., 1997). Agreement between PTSD diagnoses obtained from the PDS and the Structured Clinical Interview for DSM-III-R SCID-PTSD module was 82%. The sensitivity of the PDS was 0.89 and its specificity was 0.75. The present study used the German version of the PDS, which was validated previously (Steil and Ehlers, 2000). Internal consistency in this sample was found to be a = .88 for mothers and a = .89 for fathers. In accordance with the DSM-IV-TR, PTSD was diagnosed if participants reported at least one reexperiencing symptom, three avoidance symptoms, two arousal symptoms, and impairment in one domain of life. A diagnosis of partial PTSD required at least one symptom in each of the three PTSD symptom clusters (Stein et al., 1997). Threat Appraisals. Child accident-related threat appraisals were assessed by a single item derived from an appraisal scale that was previously validated in pediatric patients (Vollrath et al., 2004). On a 3-point Likert scale ranging from not very threatening (0) to rather threatening (2), children were asked to tell how threatening their accident had been.
1278
Life Events. The occurrence of 12 major life events, such as change of domicile, unemployment, or parental separation was assessed by using the mothers as informants. Appropriate events for our sample were selected from previously published life event scales (Paykel, 1983). A sum score was computed by counting the number of life events. Life events were retrospectively assessed at initial assessment for the 12 months preceding the accident and at follow-up for the time period following the accident. Injury Severity. Severity of injuries was classified by a physician using the Modified Injury Severity Scale (MISS), a highly reliable and widely accepted scale (Mayer et al., 1980). The MISS allows the rating of the severity of injuries in different bodily systems. MISS values range from 1 to 75 (death), with scores >25 indicating a severe injury. Socioeconomic Status. Socioeconomic status was calculated by means of a score reflecting paternal occupation and maternal education (range 2–12 points). Three social classes were defined as follows: scores 2–5, lower class; scores 6–8, middle class; scores 9–12, upper class. This measure has been shown to be a reliable and valid indicator of socioeconomic status in our community (Landolt et al., 2002a). Procedure The study was approved by the ethics committees of all involved hospitals. Written informed consent was obtained from parents. Assessments were conducted 4–6 weeks after the child’s RTA (T1) and at 12 months (T2). None of the injured patients was in a lifethreatening condition after the accident. The children were assessed by means of a standardized face-to-face interview conducted by trained graduate students of psychology. The interviews lasted 30 to 60 minutes. Most interviews were conducted in the participants’ homes; some were conducted at the hospital. To ensure that children could openly express their own account of the accident, they were interviewed when possible without their parents present. Parents received their questionnaires at the same time and were asked to complete them separately and to return them by mail. Demographic and medical variables were retrieved from the patients’ hospital records. Statistical Analyses Data were analyzed using the statistical package SPSS, release 11 (SPSS Inc., Chicago, IL). Analyses were performed with twosided tests. x2 analyses were used to compare nominal variables. Kolmogorov-Smirnov tests revealed positively skewed distributions for the child RI scores at T1 and PDS scores of fathers at T1. Therefore, Spearman-Brown rank correlation tests were used to examine associations between child PTSSs and various predictors. Wilcoxon tests for pairwise comparisons were calculated to compare RI scores and PDS scores at T1 and T2. To examine prediction of children’s PTSSs after 1 year, linear multiple regression analysis was used. All complete 53 data sets were included in this analysis, with child RI at T2 as the dependent variable. The following independent variables were chosen on the basis of previous knowledge and significant correlations with the dependent variable: age, sex, threat appraisal at T1, number of preceding life events, MISS score, maternal PDS score at T1, paternal PDS score at T1, child RI at T1. PDS and RI scores were successfully normalized by means of square root transformation before using them in the regression analysis.
J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 44:12, DECEMB ER 2005
PTSD IN CH ILD REN AFTER TRAFFIC A CCID EN TS
RESULTS Characteristics of the Sample
Descriptive information about the sample is contained in Table 1. Of the sample, 85.3% was of Swiss origin, and 8.8% originated from a Mediterranean country (Italy, Spain, former Yugoslavia). Most families were from the upper or middle class, probably on account of the language requirement (many non-Swiss nationals are from the lower classes). Fifty-seven children (83.8%) lived with both biological parents. The children had received minor head injuries (58.8%), lower extremity fractures (16.2%), upper extremity fractures (14.7%), nonextremity fractures (36.8%), and internal injuries (16.2%); 48.5% of the children had combined injuries. The mean MISS score was 10.21 (SD = 7.58; median = TABLE 1 Demographic, Accident, and Medical Characteristics of the Sample (N = 68) Age at accident, yr Mean (SD) 9.82 (2.55) Median 9.38 Sex Female 31 (45.6%) Male 37 (54.4%) Socioeconomic status Lower 6 (8.8%) Middle 36 (52.9%) Upper 23 (33.8%) Unknown 3 (4.4%) Living with both 57 (83.8%) biological parents No. of preceding life events, mean (SD) 1.27 (1.26) No. of postaccident life events, mean (SD) 1.02 (1.71) Type of accident Car passenger 10 (14.7%) Bicycle/motorcycle 21 (30.9%) Pedestrian 30 (44.1%) Other 7 (10.3%) Parent involved 9 (13.2%) Modified Injury Severity Scale score Mean (SD) 10.21 (7.58) Median 9.00 Days in hospital, T1 Mean (SD) 8.09 (7.36) Median 5.50 Days in hospital, T2 Mean (SD) 10.05 (14.70) Median 5.50
9.0). Three children had incurred severe injuries (MISS >25). Two children (2.9%) were still hospitalized at initial assessment. Most children were injured either as pedestrians or as bicycle riders. In nine cases (13.2%), at least one parent was also involved in the accident. Prevalence and Course of PTSSs in Children
Descriptive statistics for RI scores at T1 and T2 are provided in Table 2. Of the 68 children, 11 (16.2%) had scores in the clinical range of the RI at T1, thus showing moderate to severe reactions. At 12-month follow-up, 12 children (17.6%) scored in the clinical range. Five of the latter showed a delayed onset of symptoms, thus scoring in the moderate to severe range only at 12 months, but not at 4–6 weeks. Four children with RI scores in the clinical range at T1 showed a decrease below the cutoff score at T2. On average, children reported symptoms in the mild range of PTSD at both T1 and T2. Analysis of RI items revealed that the most common symptoms reported were feelings of guilt, intrusive recollections of the accident, distress at exposure to cues that resemble aspects of the trauma, and acting or feeling as if the traumatic event was reoccurring. There was no significant decrease of mean RI scores between the two assessment points. This was also true if the five patients with delayed onset were excluded from analysis (Wilcoxon Z = –1.34, p = .18). At T2 parents of 12 children (17.6%) reported that their child had received some kind of outpatient psychological treatment following the accident. At T1 two of these children and at T2 five of these children showed PTSSs in the clinical range. Prevalence and Course of PTSD in Parents
Descriptive statistics for parental PTSD severity at T1 and T2 are provided in Table 2. Twelve mothers (20%) and six fathers (11.3%) met criteria for full PTSD at T1. An additional 37 mothers (61.7%) and 24 fathers (45.3%) suffered from partial PTSD. Prevalence of PTSD significantly decreased over time, with three mothers (5.7%) and no father meeting diagnostic criteria at 12 months. In seven mothers (13.2%) and six fathers (11.3%), partial PTSD was found at T2. Maternal PTSD at T1 and T2 was not associated with involvement in the accident (T1: x2 = 0.24, p = .62; T2: x2 = 1.87, p = .17). The same was true for PTSD of fathers at T1 (x2 = 0.12, p = .72). Severity of maternal PTSD at T2 was significantly associated with severity of PTSD
J. AM. ACAD. CH ILD ADO LE SC. PSY CH IAT RY, 44:12, DE CEM BER 2 005
1279
LAND OL T ET A L.
TABLE 2 Descriptive Statistics for Main Variables at 4–6 Weeks (T1) and at 12 Months (T2) T1 T2 Child RI Maternal PTSD severity Paternal PTSD severity
Mean
SD
Mean
SD
No.
Wilcoxon Z
p
14.44 11.55 7.46
10.98 8.13 6.73
14.02 3.34 2.00
10.01 4.31 2.67
68 60 53
–0.32 –5.93 –5.04
.75 <.0005 <.0005
Note: PTSD = posttraumatic stress disorder.
at T1 (r = 0.66, p < .001) and preceding life events (r = 0.32, p = .02), but not with length of hospitalization (r = 0.25, p = .09), injury severity (r = 0.11, p = .45), age (r = 0.05, p = .73), and sex of the child (r = 0.12, p = .38). Severity of fathers’ PTSD was significantly correlated with PTSD severity at T1 (r = 0.52, p < .001), length of hospitalization (r = 0.33, p = .03) and preceding life events (r = 0.33, p = .02), but not with injury severity (r = 0.19, p = .18), age (r = 0.03, p = .82), and sex of the child (r = –0.11, p = .46). Predictors of PTSSs in Children
Spearman correlations of RI scores at T1 and T2 with predictor variables are presented in Table 3. Child threat appraisals at T1 were significantly correlated with RI scores at T1 and T2. Notably, at both time points, TABLE 3 Spearman Correlation Between Predictor Variables and Child RI at 4–6 Weeks (T1) and 12 Months (T2) Reaction Reaction Index T1 Index T2 Age Sex Socioeconomic status Preceding life events Postaccident life events Modified Injury Severity Scale score Days in hospital, T1 Days in hospital, T2 Parent involved in accident Child appraisal of threat, T1 Maternal PTSD severity, T1 Paternal PTSD severity, T1 Child RI, T1
0.07 0.01 –0.02 0.14 –0.20
0.09 0.09 –0.03 0.25* –0.09
–0.05 –0.02
0.02 0.00 –0.05
0.03
0.21
0.37*** 0.06 0.01
0.29** 0.24 0.31* 0.64***
Note: 68 ‡ n ‡ 53; *p < .05; **p < .01; ***p < .001.
1280
sociodemographic variables such as age, sex, and socioeconomic status were not significantly associated with RI scores. The same was true for injury- and accident-related predictors, such as injury severity, length of hospital stay, and parental involvement in the accident. In addition, Kruskal-Wallis tests revealed that RI scores at T1 and T2 did not differ significantly with regard to the type of accident (T1: x2 = 0.54, p = .76; T2: x2 = 4.74, p = .09). Life events occurring after the accident were not related to RI scores at T1 and T2, but life events preceding the accident were related to RI scores at T2. Whereas maternal PTSD severity was not associated with RI scores, paternal PTSD severity was significantly correlated with child PTSSs at T2. Finally, RI scores at T1 were significantly correlated with RI scores at 12 months, indicating a substantial stability of PTSSs over time. Table 4 summarizes statistics for the regression model predicting child PTSSs at 12 months. Overall, the selected variables accounted for 52% of the variance in the RI scores, making the regression model highly significant. Child symptomatology at T1 and severity of TABLE 4 Regression Analysis Predicting Child PTSD Severity (RI Score) at 12 Months Variable Age, yr Sex Modified Injury Severity Scale score Preceding life events Threat appraisal, T1 Child RI score, T1 Maternal PDS score, T1 Paternal PDS score, T1
b
Significance b
Partial r
–.09 .08
.37 .44
–0.14 0.12
–.07 .11 .16 .57 .11 .31
.53 .36 .19 <.0005 .30 .01
–0.10 0.14 0.20 0.64 0.16 0.37
Note: n = 53; F = 8.065; p < .0005; R2 = 0.60, R2 adj. = 0.52. PTSD = posttraumatic stress disorder; PDS = Posttraumatic Diagnostic Scale.
J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 44:12, DECEMB ER 2005
PTSD IN CH ILD REN AFTER TRAFFIC A CCID EN TS
paternal PTSD at T1 both contributed with significant b weights. Notably, age, sex, number of preceding life events, injury severity, child threat appraisal, and maternal PDS score at T1 did not significantly contribute to the prediction of child PTSSs at 12 months. DISCUSSION
This prospective study found that clinically significant PTSSs occur in a small but significant portion of school-age children as a consequence of RTAs, thus confirming our initial hypothesis. The prevalence of moderate to severe PTSSs was 16.2% at 4–6 weeks, and 17.6% at 12-month follow-up. This is consistent with earlier findings that 10%–35% of traffic-injured children have PTSD or clinically relevant PTSSs (de Vries et al., 1999; Di Gallo et al., 1997; Landolt et al., 2003; McDermott and Cvitanovich, 2000; Mirza et al., 1998; Stallard et al., 1998, 1999). Also, PTSD was common in mothers and fathers, which supports the notion that witnesses of traumatic events and caregivers are at increased risk of posttraumatic maladjustment (Landolt et al., 2002b, 2005). Unexpectedly, based on mean RI scores, we did not find any decrease in severity of child PTSSs over time. This is probably the result of lower prevalence rates of moderate to severe PTSSs at T1 in our study as compared with other studies (de Vries et al., 1999; Di Gallo et al., 1997; Stallard et al., 1998). In this study, most children either had no PTSSs or recovered psychologically during the first weeks after the accident. Those who showed clinically significant PTSSs at 4–6 weeks had a high risk of maintaining their symptoms over a longer period of time; however, the course of symptoms was different in individual patients. Four of the 11 children (36%) with clinical levels of PTSSs at T1 showed a decrease below the cutoff score by T2. Conversely, five children had a delayed onset of symptoms, thus showing that this phenomenon is also present in children (Gillies et al., 2003). This study also examined the importance of several predictors of PTSSs in children after RTAs. Multiple regression analysis revealed two significant predictors of child PTSSs at 12 months: child PTSSs at 4–6 weeks and paternal PTSD severity at 4–6 weeks. The importance of early PTSSs for long-term outcome in children has been demonstrated in a previous study with RTA victims (Mirza et al., 1998). Also, findings by Daviss
et al. (2000) in injured children suggest an association between acute distress and subsequent PTSSs. Although the importance of early symptomatology has been demonstrated, this is the first study to show the importance of fathers in the psychological recovery of the child after an RTA. Severity of fathers’ PTSD at T1 contributed significantly to the prediction of child PTSSs at T2. The fact that fathers’ and children’s PTSSs were not associated at T1 and that paternal PTSSs had decreased significantly by T2 suggests that fathers’ distress symptoms in the early period after the accident affect the child in the long term. In contradiction to our hypothesis, mothers’ PTSD did not affect the children’s PTSSs at either time point. As noted earlier, previous studies tended to solely focus on the role of mothers, and fathers were neglected. The only study examining this issue found no evidence of the impact of fathers’ PTSD on child PTSSs (Winje and Ulvik, 1998); however, the sample in that study was small and included several sibling pairs, thus implying dependency of observation. Also, the assessment of PTSSs did not include hyperarousal symptoms. Interestingly, indirect support of our findings can be derived from a study in children with type 1 diabetes that found a positive relationship between children’s and fathers’ adjustment (Chaney et al., 1997). Why are fathers important for the posttraumatic adjustment of their children? One may speculate that children react more strongly to their fathers’ symptoms because this is less common for them. In Western society, fathers are expected to be strong and to handle difficult situations without displaying excessive emotions. Also, a father with PTSD may have difficulties supporting his wife, which in turn may negatively affect the child. Finally, in our sample, mainly fathers were responsible for the family income. A particularly distressed father can have difficulties at work and may therefore threaten the socioeconomic security of the family. This in turn may affect the child. Whereas in this study child threat appraisals were significantly associated with PTSSs at 12 months in the bivariate analyses, multiple regression analysis revealed no significant impact of threat appraisals on PTSSs at 12 months. This result is in contrast to our hypothesis and previous findings. Compared with other studies, however, the time lag of 11 months between the assessment of threat appraisals at initial assessment and PTSSs at follow-up is longer in our study. It is possible that in
J. AM. ACAD. CH ILD ADO LE SC. PSY CH IAT RY, 44:12, DE CEM BER 2 005
1281
LAND OL T ET A L.
children the impact of specific cognitions on PTSSs may not last as long as they do in adults. With regard to accident- and injury-related variables, our hypothesis was confirmed. We found that injury severity and other medical characteristics were unrelated to PTSSs, which is consistent with previous studies (Aaron et al., 1999; Bryant et al., 2004; Daviss et al., 2000; de Vries et al., 1999; Kassam-Adams and Winston, 2004; McDermott and Cvitanovich, 2000; Mirza et al., 1998; Stallard et al., 1998). In accordance with Stallard et al. (1998), our findings also showed that the type of accident was not related to the development of PTSSs. We could therefore not confirm a finding by Mirza et al. (1998) that children who were passengers in a motor vehicle sustaining an accident had a higher risk of PTSD. Moreover, in our study, parental involvement in the accident was not associated with child PTSSs. Although this result is in line with Mirza et al. (1998), it contradicts other studies (de Vries et al., 1999; Ellis et al., 1998). Presumably, the role of parental involvement with regard to child PTSSs is more complex and depends on the type of accident. One may also speculate that parental guilt may act as a moderating variable. This should be analyzed in further studies. Limitations
Some limitations of this study merit note. First, the measures used were aimed at PTSD and PTSSs. Inclusion of standardized instruments assessing symptoms of depression and anxiety in children and parents would have been useful. Second, in our study, assessment methods in children and parents were different: Children were interviewed, whereas parents were assessed by questionnaire. This may have biased our findings in an unknown way. Third, this study did not assess data on premorbid functioning of children and parents. We therefore may have missed an important additional predictor of PTSSs and PTSD. Also, because this study focused on child PTSSs, treatment interventions for parents’ PTSD during the study period were not assessed. A final issue potentially limiting generalizability of our findings is the participation rate of 59%. We have only basic sociodemographic information about nonparticipants. It is possible that some of the nonparticipants declined participation because of avoidance symptoms of PTSD or because they were well adjusted and the study was not relevant to them. Both cases would affect prevalence estimates of PTSD. Our participation rate is consistent
1282
with those of other studies that include comprehensive psychosocial evaluations, however. Also, comparisons of completers with nonparticipants and dropouts revealed differences only with regard to sex. Because we found no differences between boys and girls regarding PTSSs, it is unlikely that this influenced our findings. Clinical Implications
Despite these limitations, the present study has several strengths, not the least of which are its longitudinal design and its independent and standardized assessment of multiple informants. There are several clinical implications that can be drawn from our findings. Certainly, children after RTAs and their parents have to be assessed carefully with regard to PTSSs and PTSD. Delayed onset of symptoms in some children implies a monitoring beyond the first weeks after the accident. Emergency departments and hospitals could provide information on psychotraumatological issues to children and parents (e.g., by use of leaflets) and ensure that general practitioners are informed about particularly distressed children and families (Bryant et al., 2004). Psychological interventions need to be developed to prevent and treat psychological consequences of RTAs in children and if necessary in their parents. Such interventions should be offered to children irrespective of the severity of the injury. To date, such interventions are rarely made available (Gillies et al., 2003). In addition, our findings underscore that fathers need to be more actively involved in family-based prevention and treatment interventions of child PTSSs. Fathers play a specific, yet not well-understood role in child trauma recovery that needs to be studied in more detail. Disclosure: The authors have no financial relationships to disclose.
REFERENCES American Psychiatric Association (2000), Diagnostic and Statistical Manual of Mental Disorders, DSM-IV-TR, Text Revision. Washington, DC: American Psychiatric Association Aaron J, Zaglul H, Emery RE (1999), Posttraumatic stress in children following acute physical injury. J Pediatr Psychol 24:335–343 Brislin RW, Lonner WJ, Thorndike RM (1973), Cross-cultural Research Methods. New York: Wiley Bryant B, Mayou R, Wiggs L, Ehlers A, Stores G (2004), Psychological consequences of road traffic accidents for children and their mothers. Psychol Med 34:335–346 Chaney JM, Mullins LL, Frank RG, Peterson L, Mace LD, Kashani JH (1997), Transactional patterns of child, mother, and father adjustment in insulin-dependent diabetes mellitus: a prospective study. J Pediatr Psychol 22:229–244
J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 44:12, DECEMB ER 2005
PTSD IN CH ILD REN AFTER TRAFFIC A CCID EN TS
Daviss WB, Mooney D, Racusin R, Ford JD, Fleischer A, McHugo GJ (2000), Predicting posttraumatic stress after hospitalization for pediatric injury. J Am Acad Child Adolesc Psychiatry 39:576–583 de Vries APJ, Kassam-Adams N, Cnaan A, Sherman-Slate E, Gallagher PR, Winston FK (1999), Looking beyond the physical injury: posttraumatic stress disorder in children and parents after pediatric traffic injury. Pediatrics 104:1293–1299 Di Gallo A, Barton J, Parry-Jones W (1997), Road traffic accidents: early psychological consequences in children and adolescents. Br J Psychiatry 170:358–362 Ehlers A, Mayou R, Bryant B (2003), Cognitive predictors of posttraumatic stress disorder in children: results of a prospective longitudinal study. Behav Res Ther 41:1–10 Ellis A, Stores G, Mayou R (1998), Psychological consequences of road traffic accidents in children. Eur Child Adolesc Psychiatry 7:61–68 Foa EB, Cashman L, Jaycox L, Perry K (1997), The validation of a self-report measure of posttraumatic stress disorder: the Posttraumatic Diagnostic Scale. Psychol Assess 9:445–451 Frederick C, Pynoos R, Nader K (1992), Childhood Posttraumatic Stress Reaction Index—A Copyrighted Instrument. Los Angeles: University of California Gillies ML, Barton J, Di Gallo A (2003), Follow-up of young road accident victims. J Trauma Stress 16:523–526 Green BL, Korol M, Grace MC, Vary MG, Leonard AC, Gleser GC (1991), Children and disaster: age, gender, and parental effects on PTSD symptoms. J Am Acad Child Adolesc Psychiatry 30:945–951 Kassam-Adams N, Winston FK (2004), Predicting child PTSD: the relationship between acute stress disorder and PTSD in injured children. J Am Acad Child Adolesc Psychiatry 43:403–411 Landolt MA, Nuoffer JM, Steinmann B, Superti-Furga A (2002a), Quality of life and psychologic adjustment in children and adolescents with early treated phenylketonuria can be normal. J Pediatr 140:516–521 Landolt MA, Ribi K, Laimbacher J, Vollrath M, Gnehm HE, Sennhauser FH (2002b), Posttraumatic stress disorder in parents of children with newly diagnosed type 1 diabetes. J Pediatr Psychol 27:647–652 Landolt MA, Vollrath M, Laimbacher J, Gnehm HE, Sennhauser FH (2005), Prospective study of posttraumatic stress disorder in parents of children with newly diagnosed type 1 diabetes. J Am Acad Child Adolesc Psychiatry 44:682–689
Landolt MA, Vollrath M, Ribi K, Gnehm HE, Sennhauser FH (2003), Incidence and associations of parental and child posttraumatic stress symptoms in pediatric patients. J Child Psychol Psychiatry 44:1199–1207 Mayer T, Matlak ME, Johnson DG, Walker ML (1980), The Modified Injury Severity Scale in pediatric multiple trauma patients. J Pediatr Surg 15:719–726 McDermott BM, Cvitanovich A (2000), Posttraumatic stress disorder and emotional problems in children following motor vehicle accidents: an extended case series. Aust N Z J Psychiatry 34:446–452 Mirza KAH, Bdadrinath BR, Goodyer IM, Gilmour C (1998), Posttraumatic stress disorder in children and adolescents following road traffic accidents. Br J Psychiatry 172:443–447 Nader K, Pynoos R, Fairbanks L, Frederick C (1990), Children’s PTSD reactions one year after a sniper attack at their school. Am J Psychiatry 147:1526–1530 Paykel ES (1983), Methodological aspects of life event research. J Psychosom Res 27:341–352 Pynoos RS, Goenjian A, Tashjain M, Karakashian M, Manjikian R, Manoukian R (1993), Post-traumatic stress reactions in children after the 1988 Armenian earthquake. Br J Psychiatry 163:239–247 Stallard P, Velleman R, Baldwin S (1998), Prospective study of post-traumatic stress disorder in children involved in road traffic accidents. BMJ 317:1619–1623 Stallard P, Velleman R, Baldwin S (1999), Psychological screening of children for post-traumatic stress disorder. J Child Psychol Psychiatry 40:1075–1082 Stallard P, Velleman R, Baldwin S (2001), Children involved in road traffic accidents: psychological sequelae and issues for mental health services. Clin Child Psychol Psychiatry 6:23–39 Steil R, Ehlers A (2000), Posttraumatische Diagnoseskala (PDS). Jena, Germany: Universita¨t Jena Psychologisches Institut Stein MB, Walker JR, Hazen AL, Forde DR (1997), Full and partial posttraumatic stress disorder: findings from a community survey. Am J Psychiatry 154:1114–1149 Vollrath M, Landolt MA, Ribi K (2004), Illness appraisal in paediatric patients and their parents: a short scale. Swiss J Psychol 63:3–15 Winje D, Ulvik A (1998), Long-term outcome of trauma in children: the psychological consequences of a bus accident. J Child Psychol Psychiatry 39:635–642
J. AM. ACAD. CH ILD ADO LE SC. PSY CH IAT RY, 44:12, DE CEM BER 2 005
1283