The Association Between Parent PTSD/Depression Symptoms and Child PTSD Symptoms: A Meta-Analysis

Transcription

1 The Association Between Parent PTSD/Depression Symptoms and Child PTSD Symptoms: A Meta-Analysis Adam Morris, 1 MA, Crystal Gabert-Quillen, 1 MA, and Douglas Delahanty, 1,2 PHD 1 Department of Psychology, Kent State University and 2 Department of Psychology in Psychiatry, Northeast Ohio Medical University (NEOMED) All correspondence concerning this article should be addressed to Douglas Delahanty, Department of Psychology, Kent State University, 342 Kent Hall, Kent, OH ddelahan@kent.edu Received December 22, 2011; revisions received June 28, 2012; accepted June 30, 2012 Objective The present article presents a meta-analysis of studies examining the association between parent posttraumatic stress disorder (PTSD)/depression symptoms and child PTSD symptoms (PTSS) after a child s exposure to a traumatic event while considering multiple moderating factors to explain heterogeneity of effect sizes. Methods 35 studies were included: 32 involving the association between parent and child PTSS and 9 involving the association between parent depression and child PTSS. Results Across existing studies, both parent and child PTSS (r ¼ 0.31) and parent depression and child PTSS (r ¼ 0.32) yielded significant effect sizes. Parent gender, assessment type (interview vs. questionnaire), differences in assessment type for parents and children, and study design (cross-sectional vs. longitudinal) moderated the relationship between parent and child PTSS. Conclusions The current findings confirm the associations between parental posttraumatic responses and child PTSS and highlight important moderating factors to include in future studies of child PTSS. Key words depression; meta-analyses; parents; posttraumatic stress. Sixty-eight percent of children in the United States experience a potentially traumatic event (PTE) before the age of 16 years, and more than one-half of those children have experienced multiple PTEs (Copeland, Keeler, Angold, & Costello, 2007; Costello, Erkanli, Fairbank, & Angold, 2002). Despite high levels of exposure, only a minority of children meets full Posttraumatic Stress Disorder (PTSD) diagnostic criteria after a trauma, and estimates suggest that only 20% of child trauma victims develop clinically significant symptoms (Kahana, Feeny, Youngstrom, & Drotar, 2006; Kilpatrick et al., 2003). Thus, research has explored risk factors that would help identify children at risk for PTSD after a PTE (Copeland, Keeler, Angold, & Costello, 2007; Kassam-Adams & Winston, 2004). Identified risk factors include gender (e.g. Bokszczanin, 2007), age (e.g., Green et al., 1991), and peritraumatic reactions (e.g., dissociation: Bui et al., 2011). However, research on risk factors for PTSD in pediatric populations has yielded inconsistent findings, and has suggested that these factors account for only a small amount of variance in child PTSD symptoms (Alisic, Jongmans, van Wesel, & Kleber, 2011; Cox, Kenardy, & Hendrikz, 2008; Wolfe, Crooks, Lee, McIntyre-Smith, & Jaffe, 2003). Inconsistent findings with regards to individual-level risk factors may be because of a failure to consider the relationship between parental responses to the trauma and PTSD symptoms (PTSS) in children (de Vries et al., 1999; Pynoos et al., 2009). Recent research has attempted to address this shortcoming by examining the associations between parent posttraumatic sequelae (mainly PTSS and depression) and PTSS in children (e.g., Meiser-Stedman, Yule, Dagleish, Smith, & Glucksman, 2006; Nugent, Ostrowski, Christopher, & Delahanty, 2006; Ostrowski et al., 2011; Valentino, Berkowitz, & Stover, 2010). However, this research has also yielded equivocal findings. The meta-analytic framework provides a means by which to examine this inconsistent literature in an attempt to estimate the strength of the association between parent and Journal of Pediatric Psychology 37(10) pp , 2012 doi: /jpepsy/jss091 Advance Access publication September 27, 2012 Journal of Pediatric Psychology vol. 37 no. 10 ß The Author Published by Oxford University Press on behalf of the Society of Pediatric Psychology. All rights reserved. For permissions, please journals.permissions@oup.com

2 Parent PTSD/Depression and Child PTSS 1077 child posttraumatic distress (Lipsey & Wilson, 2000). In addition, meta-analytic reviews allow for the consideration of potential moderator variables (Lipsey & Wilson, 2000; Rosenthal & DiMatteo, 2001). Thus, the current article provides a meta-analytic review of studies examining the relationship between parent PTSS/depression and child PTSS while also examining the impact of a number of moderating variables on these relationships. The association between parent and child PTSS has been examined after a variety of child traumatic events including diagnosis with a chronic or terminal illness (e.g., Ben-Amitay et al., 2006; Bronner, Knoester, Bos, Last, & Grootenhuis, 2008; Brown et al., 2007; Kazak et al., 2004; Magal-Vardi et al., 2004), injuries (e.g., Coakley et al., 2010; Hall et al., 2009; Kassam-Adams et al., 2004; Nugent et al., 2006; Ostrowski, Christopher, & Delahanty, 2007), disasters (e.g., Birmes et al., 2009; Kilic, Özguven, & Sayil, 2003; Koplewicz et al., 2002; Jones, Ribbe, Cunningham, Weddle, & Langley, 2002), and war experiences (e.g., Quota, Punamäki, & El Sarraj, 2003; Smith, Perrin, Yule, & Rabe-Hesketh, 2001; Thabet, Tawahina, Sarraj, & Vostanis, 2008). In general, this literature has produced mixed findings regarding the concordance of parent and child PTSS, with some studies reporting a significant association (e.g., Hall et al., 2006; Nugent et al., 2006; Ostrowski et al., 2011) and others reporting no such association (e.g., Bryant et al., 2004; Kassam-Adams, Fleisher, & Winston, 2009; Landolt, Vollrath, Ribi, Gnehm, & Sennhauser, 2003). Equivocal findings may be due in part to methodological differences between studies. For example, when parents report on their own their child s symptoms, there is a higher correlation between parent and child symptoms than in studies in which children self-reported their own symptoms (Shemesh et al., 2005; Smith et al., 2001). Parent depression has also been associated with child PTSS after pediatric medical events (Ben-Amitay et al., 2006; Meiser-Stedman, Yule, Dagleish, Smith, & Glucksman, 2006; Valentino et al., 2010; Zatzick et al., 2006), disasters (Birmes et al., 2009; Wickrama & Kaspar, 2007; Kilic et al., 2003), and war experiences (Smith et al., 2001; Qouta et al., 2003). Although the association between parent depression and a variety of adverse child outcomes are well established, several metaanalyses have revealed substantial variability across studies (Goodman et al., 2011; Connell & Goodman, 2002). Further, to our knowledge, an empirical review of studies examining the relationship between parent depression and child PTSS after a traumatic event has not been conducted. For the current meta-analysis, we calculated the weighted mean effect sizes of the relationships between parent and child PTSS and between parent depressive symptoms and PTSS in children. A comprehensive empirical review of the strength of these associations will aid in answering questions about how much variance in child PTSS is, and is not, accounted for by parent PTSS/depression, while also taking into account differences in sample size. Several potential moderating factors have also been proposed in the literature; the present meta-analysis considers the role of these moderators to inform future study design. Moderators A number of variables are hypothesized to moderate the relationship between parent PTSS/depression and child PTSS. Females report greater levels of PTSD than males (Tolin & Foa, 2006), and maternal PTSD symptoms may differentially impact boy versus girl trauma victims (Ostrowski et al., 2007). Further, intergenerational studies of PTSD suggest that maternal, and not paternal, PTSD is associated with an increased prevalence of PTSD in offspring (Yehuda, Bell, Bierer, & Schmeidler, 2008). Additionally, it has been suggested that mixing crosssectional and longitudinal studies in reviews may lead to misleading findings, as longitudinal studies are more likely to capture several distinct symptom trajectories in parents and children, and the strength of the association between parent and child PTSS increases over time (Alisic et al., 2011; Koplewicz et al., 2002; Le Brocque, Hendrikz, & Kenardy, 2010; O Donnell, Elliot, Lau, & Creamer, 2007). In addition to proposed differences between cross-section and longitudinal studies, it is likely that there are differences based on assessment type. Effect sizes for randomized controlled trials that use self-report to measure improvement in internalizing symptoms are typically smaller than effect sizes of studies that use clinician-rated scales (Cuijpers, Hofmann, & Andersson, 2010). Although differences based on assessment type have not been explicitly investigated with parent child dyads, these findings suggest a possible lack of equivalency between assessment types. Therefore, the current metaanalysis investigated whether gender, study design (longitudinal or cross-sectional), or the type of assessment measure used (self-report vs. interview) influenced the association between parent PTSS/depression symptoms and child PTSS. Incidence rates for PTSD also differ depending on the index trauma experienced (Breslau, Chilcoat, Kessler, & Davis, 1999). The fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV: American Psychiatric Association, 2000) recognized that diagnosis of a life-threatening or terminal disease and

3 1078 Morris, Gabert-Quillen, and Delahanty learning that one s child has a life-threatening illness is a PTE that can lead to the development of PTSD. This led to a large number of studies examining rates of PTSD after a child s diagnosis with a life-threatening illness, most commonly a diagnosis of cancer (Bruce, 2006). These studies have found that parents of children with cancer have elevated levels of PTSS compared with healthy control subjects (Brown, Madan-Swain, & Lambert, 2003; Barakat et al., 1997; Pelcovitz et al., 1998), and that children with cancer typically have lower levels of PTSS compared with their parents and siblings (Kazak et al., 2004; Manne Du Hamel, Gallelli, Sorgen, & Redd, 1998; Stuber, Christakis, Houskamp, & Kazak, 1996). However, when comparisons are made between rates of PTSD in children diagnosed with a medical illness compared with other child trauma samples, the results have been inconsistent (Schwartz & Drotar, 2006; Stoppelbein, Greening, & Elkin, 2006). To our knowledge, no study has compared whether the association between parent and child posttraumatic distress differs depending on whether the index trauma involves a medical diagnosis. Therefore, trauma type (medical diagnosis vs. other traumas) was also considered as a potential moderator. Finally, one of the more controversial components of the PTSD diagnosis is the requirement that a trauma victim experiences or witnesses an event that involves actual or perceived life threat (criterion A1), and that they respond with intense fear, helplessness, or horror (criterion A2; APA, 2000). Given the debate concerning the diagnostic necessity of including criterion A in the upcoming DSM-5 (O Donnell, Creamer, McFarlane, Silove, & Bryant, 2010), whether criterion A was used as an inclusionary criterion was also examined as a potential moderator. Hypotheses Past research has demonstrated that child posttraumatic responses are likely influenced by their parents responses (e.g., Meiser-Stedman, Yule, Dagleish, Smith, & Glucksman, 2006; Nugent et al., 2006; Ostrowski et al., 2011; Valentino et al., 2010); accordingly, it was hypothesized that both parent PTSS and depression would be positively associated with child PTSS. Given the smaller sample sizes for the moderation analyses, these analyses were largely considered exploratory; however, directional hypotheses were made based on existent literature. For instance, because females are more likely than males to display PTSS (Brewin, Andrews, & Valentine, 2000), and research has suggested that maternal symptoms afford a specific risk to child PTSS (Ostrowski et al., 2007; Yehuda et al., 2008), it was hypothesized that associations between female caregivers and children would yield stronger effect sizes compared with associations between paternal symptoms and child PTSS. Further, with theoretical arguments that longitudinal studies are more likely to capture the natural course of symptoms after an event (Alisic et al., 2011), we hypothesized that longitudinal studies would have larger effect sizes. Given the mixed findings with regards to the rates of PTSD because of a medical diagnosis (Schwartz & Drotar, 2006; Stoppelbein, Greening, & Elkin, 2006), we hypothesized that studies assessing traumas other than medical diagnoses would have larger relative effect sizes. With respect to assessment type, because research on internalizing disorders has demonstrated that interviews typically yield stronger effect sizes (Cuijpers et al., 2010), it was hypothesized that studies using clinical interviews to assess symptoms would have larger effect sizes. To account for differences in measurement type, we conducted exploratory analyses on whether the mode of assessment was the same or different for parents and children. Finally, studies requiring that children endorse Criterion A as an inclusionary criterion for participation were hypothesized to have larger effect sizes, as they were more likely to include children at a higher risk for PTSS. Methods Literature Search for Relevant Studies Literature searches were conducted by the lead author (A.M.) using PsycINFO,MEDLINE, Google Scholar, and PILOTS databases for publications from 1980 to 2011 to align with the years that PTSD has been included in the DSM. The following key word(s) were used: trauma, posttraumatic stress disorder, posttraumatic stress symptoms, child(ren), PTSD, PTSS, stress, parent(s), caregiver, predictor(s), MVA, accident(s), assault, fall, burn, natural disaster(s), medical events, cancer, diabetes, depression, and depressive symptoms. According to the ancestry method (Johnson & Eagly, 2000), the reference sections of the 206 empirical articles that assessed parents and children simultaneously were reviewed to locate additional studies. Inclusion/Exclusion Criteria Several criteria were used to determine whether a study was included in the meta-analysis. First, studies had to assess parents and children at the same time point. They had to include either an assessment of PTSS or depression for parents while also assessing PTSS in children. Further, children had to report on their own symptoms; therefore, studies using parent report of child symptoms were excluded. All articles had to be written in English, and

4 Parent PTSD/Depression and Child PTSS 1079 child participants had to be <18 years of age to maintain a focus on child and adolescent PTSS. Criteria B, C, and D needed to be assessed in both parents and children. Included studies had to present data on the association between parent PTSS and child PTSS or parent depression and child PTSS. If data on parent child associations were not available, enough statistical information had to be included to compute meaningful analyses. Accordingly, studies were required to provide, in some combination, Pearson product-moment correlation coefficients (r), sample sizes, means and standard deviations, or chisquare values (w 2 ). Because the included studies encompassed a wide range of research designs and effect size estimates, methodological accuracy was ensured through multiple strategies. Effect size estimates were aggregated for studies that ran analyses separated by gender or time point to avoid a distortion of the standard error estimates (Gleser & Olkin, 1994). Additionally, if a study reported a non-significant association, but did not include any information about the association, a p value of. 50 was assumed (Rosenthal, 1991). Following this procedure, 582 articles were obtained. Of these, 206 referenced both parents and children. Within this subsample, 65 were reviews, and 98 did not meet the inclusion criteria. Therefore, 35 studies were retained in the current metaanalysis. Included studies are summarized in Table I. Coding of Studies Each study was coded on a number of sample and methodological characteristics by the primary and secondary authors (A.M. and C.Q.). High interrater reliability was found between coders, with Cohen s kappa ranging from 0.70 to 1.00 for all main outcomes and moderator variables. Differences were resolved through discussion, and the authors of the included publications were contacted if clarification was necessary. Two variables were coded to reflect (1) parent depression and (2) parent PTSS. The coding scheme and analysis plan included both categorical and continuous outcomes, with all variables converted into r values to reflect continuous levels of symptoms of depression and PTSS. Only one study provided associations separated by child gender (Ostrowski et al., 2007); thus, child gender could not be examined as a moderator. For parent gender, we classified studies that provided separate associations for (1) mothers, (2) fathers, and studies that were (3) mixed. A categorical variable (1 ¼ cross-sectional and 2 ¼ longitudinal) was created for study type. Trauma type was coded as either (1) medical diagnosis or (2) non-medical diagnosis. Owing to a lack of diversity of studies on the association between parent and child symptoms in medical populations, the medical diagnosis category consisted entirely of children diagnosed with cancer. The non-medical diagnosis category contained 12 studies of traumatic injury, six of disaster trauma, three of war trauma, one of unexpected death of a parent, and four that included mixed trauma samples. Coding for type of measurement was (1) interview or (2) self-report. For differences in measurement type, we coded whether the mode of assessment was (1) the same or (2) different for parents and children. Finally, each study was coded for whether DSM-IV-TR criterion A was used as a screener before enrollment in the study, using (1) assessed and (2) not assessed. Data Analysis Recommendations for all statistical analyses were obtained from Lipsey and Wilson (2000), and the Comprehensive Meta-Analysis, Version 2 statistical software was used for all analyses (Borenstein, Hedges, Higgins, & Rothstein, 2005). Effect sizes for the associations between parent symptoms and child PTSS were calculated using the Pearson product-moment correlation (r), because r has several favorable interpretive qualities compared with Cohen s d (Rosenthal, 1991; Rosenthal & DiMatteo, 2001). Higher r values indicated stronger relationships between parent symptoms and child PTSS, with 0.10 classified as small, 0.25 as medium, and 0.40 as large effects. If a study reported the number of members in each group (PTSD vs. no PTSD), as well as means and standard deviations with t, F, w 2, or odds ratio, their results were converted to r (Lipsey & Wilson, 2000; Rosenthal, 1991). All r values were then converted to Fisher s Z (Z r ), as r is heavily influenced by sample size (Lipsey & Wilson, 2000). A relative weight for Z r was calculated for each included study to take into account the variability in sample sizes, and the (SE Zr ) of each effect size was calculated. After this, the weighted effect size (wes) for each individual study was calculated by multiplying each effect size by its weight. To yield an accurate and interpretable overall effect size, the weighted mean effect size was then converted back to an r value. The Q statistic was calculated to examine the homogeneity of effect, and the corresponding degrees of freedom (k 1) were examined on a w 2 distribution table. The presence of a significant Q statistic suggests that the variance in the effect sizes may be because of systematic differences in the studies themselves (Lipsey & Wilson, 2000). All moderator variables were categorical; therefore, an analog to the analysis of variance was used (Lipsey & Wilson, 2000).

5 1080 Morris, Gabert-Quillen, and Delahanty Table I. Studies Included in the Current Meta-Analysis Study N Parent gender Child measure Parent measure Trauma type Study Type CritA Parent outcome Appelbaum & Burns, Mixed I I Other CS No FTSS Ben-Amitay et al., Mothers I SR MD L No DEP Ben-Amitay et al., Fathers I SR MD L No DEP Birmes et al., Mothers SR SR Other CS No PTSS, DEP Birmes et al., Fathers SR SR Other CS No PTSS, DEP Bryant et al., Mothers I SR Other CS No FTSS Bronner et al., Mothers SR SR MD L No PTSS Bronner et al., Fathers SR SR MD L No FTSS Brown et al., Mothers SR SR MD CS No FTSS Coakley et al., Mixed SR SR Other CS No FTSS Hall et al., Mixed I SR Other CS No FTSS Jones et al., No information I I Other CS No FTSS Kazak et al., Fathers SR SR MD CS No FTSS Kazak et al., Mothers SR SR MD CS No FTSS Kassam-Adams et al., Mixed I SR Other CS No FTSS Kilic et al., Mothers SR SR Other CS No DEP Kilic et al., Fathers SR SR Other CS No DEP Koplewicz et al., Mixed SR SR Other L No FTSS Landolt et al, 2003 ISO Mothers I SR Other CS No FTSS Landolt et al, Fathers 1 SR Other CS No FTSS Magal-Vardi et al, Fathers 1 SR MD L No FTSS Magal-Vardi et al, Mothers I SR MD L No FTSS Meiser-Steadman et al., Mixed SR SR Other L No DEP Morris, Mothers SR SR Other CS Yes FTSS Nugent et al, Mixed I SR Other L Yes FTSS Ostrowski et al., Mothers I I Other L Yes FTSS Ostrowski et al., Mixed 1 1 Other L Yes FTSS Ozono etal., Fathers SR SR MD CS No FTSS Ozono etal., Mothers SR SR MD CS No FTSS Pelcovitz etal., Mothers I I MD CS No FTSS Phipps et al., Mixed SR SR MD CS No FTSS Qouta et al., Mothers I SR Other CS No DEP Qouta et al., Mothers I I Other CS No FTSS Self-Brown et al., Mixed SR SR Other CS No FTSS Smith et al., Mothers SR SR Other CS No FTSS; DEP Stoppelbein et al., Mixed SR SR Other CS No FTSS Stoppelbein et al., Mixed SR SR MD CS No PTSS, DEP Stuber et al., Fathers SR SR MD CS No FTSS Stuber et al., Mothers SR SR MD CS No FTSS Stuber et al., Fathers SR SR MD CS No FTSS Stuber et al., Mothers SR SR MD CS No PTSS Sturms et al., Mixed SR SR Other LS No PTSS Thabet et al., Mixed SR SR Other CS No PTSS Valentino, Berkowitz, & Stover, Mixed I SR Other CS Yes FTSS Wickrama & Kaspar, Mothers SR SR Other CS No PTSS, DEP Zatzick et al., Mixed I SR Other CS No FTSS, DEP Note. MD¼ Medical Diagnosis; SR ¼ self report; I ¼ interview; CS ¼ cross-sectional; L ¼ longitudinal. Some citations appear twice to allow a description of the separate effect sizes for mothers and fathers.

6 Parent PTSD/Depression and Child PTSS 1081 Results Study Sample In total, 35 independent studies published or submitted for publication met the inclusion criteria. In total, these studies included 3,925 parent child dyads. Sample sizes for individual studies ranged from 16 to 336 families (M ¼ 79.48, SD ¼ 78.14). Of the 35 studies included, 32 assessed parent PTSS and nine assessed parent depression (Table I). The mean ages of the children in these studies ranged from years, with an overall mean of years (SD ¼ 2.10). Parent PTSS/Parent Depression In the determination of an overall weighted effect size for parent PTSS (Table II), the w 2 test for homogeneity indicated that the effect sizes were heterogeneous (Q [ 31 ] ¼ , p <.001); therefore, the random-effects model was applied. Overall, a significant weighted mean effect size was found for the association between child and parent PTSS (weighted r ¼.31, p <.001), with a 95% confidence interval indicating medium-to-large effect sizes ranging from 0.25 to In the analysis of the overall weighted effect size for parent depression (Table II), the w 2 test for homogeneity was again significant (Q [ 8 ] ¼ 20.59, p <.001); therefore, a random-effects model was applied. A significant overall weighted mean effect size was also found for the association between parent depression and child PTSS (weighted r ¼ 0.32, p <.001), with a 95% confidence interval indicating medium-to-large effect sizes ranging from 0.21 to Next, the effect sizes were compared. This analysis revealed no significant differences in effect sizes based on parent symptom type (Q [ 39 ] ¼.46, ns). Moderation Effects The minimum requirement for inclusion in each moderation analysis was having two or more studies on all levels of the moderator. Several studies were excluded from specific depression analyses because they were the only study on a given level of the moderator. Thus, several moderators could not be explored for the relationship Table II. Weighted Mean Effect Sizes for the Relationship Between Parent Symptoms and Child PTSS Parent symptom type k r Q z 95% CI Parent PTSS a *** b *** 9.15*** Parent Depression a *** b ** 7.52*** Note. a Based on random effects model. b Based on fixed effects model. **p <.01, ***p <.001. between parent depression and child PTSS. Spearman s rho correlations were performed to examine correlations among potential moderator variables. The results of this analysis indicated that parent measure type and child measure type (r ¼ 0.50), as well as child assessment type and inclusion of criterion A (r ¼ 0.33) were significantly intercorrelated. Hunter & Schmidt (2004) suggest that confounding moderator variables should be combined to create subgroups. Although we were able to do this for parent and child measure type, for study type, and criterion A, this would result in further dividing a relatively small number of studies into smaller subgroups, prohibiting the examination of moderation. Therefore, we examined the effects of each moderator variable separately and emphasized that the findings regarding study type and criterion A should be interpreted with caution. Results for the moderation analyses can be found in Table III. A moderation effect of parent gender was found for the association between parent and child PTSS (Qb [ 1 ] ¼ 8.45, p <.001) such that maternal PTSS (r ¼ 0.29) had a larger effect size compared with paternal PTSS (r ¼ 0.13). No significant moderation effect for parent gender was found for the association between parent depression and child PTSS (Qb [ 1 ] ¼ 0.01, ns). Studies that used a parent interview (r ¼ 0.38) had a larger effect size than studies that used self-report (r ¼ 0.25; Qb [ 1 ] ¼ 10.37, p <.001). Parent assessment type could not be investigated as a moderator in studies investigating the relationship between parent depression and child PTSS, as all studies used parent self-report of depressive symptoms. Child assessment type did not significantly moderate the strength of the relationships between child PTSS and parent PTSS (Qb [ 1 ] ¼ 1.21, ns) or child PTSS and parent depression (Qb [ 1 ] ¼ 0.26, ns). However, studies that used the same mode of assessment for parents and children yielded a stronger effect size (r ¼ 0.32) than studies that used different modes of assessment for the parent and child (r ¼ 0.16; Qb [ 1 ] ¼ 17.29, p <.001). Differences in mode of assessment did not moderate the association between parent depression and child PTSS (Qb [ 1 ] ¼ 0.26, ns). The type of study design moderated the strength of the relationship between parent and child PTSS (Qb [ 1 ] ¼ 6.66, p <.01) such that longitudinal studies yielded a stronger effect size (r ¼ 0.40) compared with crosssectional studies (r ¼ 0.26). Type of study design did not moderate the relationship between parent depression and child PTSS (Qb [ 1 ] ¼ 0.01, ns). Similarly, type of trauma did not moderate the associative strength between child and parent PTSS (Qb [ 1 ] ¼ 0.07, ns), and the moderating influence of trauma type could not be explored for the relationship between child PTSS and parent depression.

7 1082 Morris, Gabert-Quillen, and Delahanty Table III. Moderation Analyses Moderator Parent IV Category k ESþ z 95% CI Parent gender PTSS Fathers ** Q B ¼ 8.45*** Mothers *** Depression Fathers Q B ¼ 0.01 Mothers Trauma type PTSS Medical diagnosis *** Q B ¼ 0.07 Other *** Study type PTSS Cross-sectional *** Q B ¼ 6.66* Longitudinal *** Depression Cross-sectional *** Q B ¼ 0.01 Longitudinal *** Criterion A PTSS Assessed *** Q B ¼ 1.36 Not assessed *** Child assessment type PTSS Interview *** Q B ¼ 1.21 Self report *** Depression Interview *** Q B ¼ 0.26 Self report *** Parent assessment type PTSS Interview *** Q B ¼ 10.37** Self report *** Assessment type comparison PTSS Different *** Q B ¼ 17.29*** Same *** Depression Different *** Q B ¼ 0.26 Same *** Note. ESþ¼mean weighted effect size. *p <.05, **p <.01, ***p <.001. Finally, the use of criterion A as a condition of participation did not moderate the associative strength between parent and child PTSS (Qb [ 1 ] ¼ 1.36, ns) and could not be tested for the relationship between parent depression and child PTSS. Publication Bias Given that published studies were the predominant source of data in the current meta-analysis, there is a risk for publication bias. Consequently multiple methods were used to test for such bias. First, using the Comprehensive Meta-Analysis, version 2, we calculated Orwin s (1983) fail-safe N for each mean effect size to determine the number of unpublished studies with null findings that would be necessary to reduce the mean effect size to a trivial criterion effect size of r ¼ 0.10 (Borenstein, Hedges, Higgins, & Rothstein, 2005; Naragon-Gainey, 2010; Orwin, 1983). Orwin s fail-safe N was 59 for the association between parent and child PTSS, and 19 for the association between parent depression and child PTSS. Next, the estimated effect sizes from each individual study were compared against their respective sample sizes using funnel plots (Sterne & Egger, 2001). All funnel plots were generally symmetrical and formed a funnel shape. A linear regression method was next used to measure funnel plot asymmetry (Egger, Smith, Schneider, & Minder, 1997). The intercept was not significantly different than zero for either the relationship between parent and child PTSS (t ¼ 1.98, ns, df ¼ 31) or the relationship between parent depression and child PTSS (t ¼ 0.55, ns, df ¼ 7), suggesting that publication bias is not likely. Discussion The literature regarding the association between parent and child posttraumatic symptoms after a child s traumatic event has increased substantially over the past few years. Although this literature has described an association between parent and child internalizing symptoms, in the child trauma literature, the extent to which parent symptoms are related to child PTSS is more mixed. The current findings support a significant positive association between parent posttraumatic distress and child PTSS. Cohen s (1988) widely adopted criteria indicate similar medium effect sizes for both parent PTSS and depression. Several mechanisms may explain the association between

8 Parent PTSD/Depression and Child PTSS 1083 parent posttraumatic distress and child PTSS including the following: a shared genetic diathesis (Koenen, Amstadter, & Nugent, 2009; Koenen et al., 2008), modeling and reinforcement of anxious or avoidant behaviors by parents (Fisak & Grills-Taqueche, 2007), and decreased positive parent child interactions (Ewell-Foster, Garber, & Durlak, 2007; Ruscio et al.; 2002). However, future longitudinal research is needed to confirm the role of these mechanisms. In addition to the main effects, several significant moderators were discovered. Consistent with our hypotheses, parent gender was a significant moderator of the relationship between child and parent PTSS. The direction of the findings suggested that maternal PTSS had a stronger relationship to child PTSS than did paternal PTSS. These findings suggest that maternal PTSS may be a salient risk factor for child PTSS. Studies on the intergenerational transmission of PTSD have found that maternal PTSS may confer risk through epigenetic mechanisms related to biological factors conferred in utero or through variations in maternal care that influence how children respond to events (see Yehuda & Bierer, 2009, for review). Parent gender did not serve as a moderator of the association between parent depression and child PTSS; however, it is likely that the relatively small number of studies of parent depression that provided associations separated by gender impacted our ability to find moderation based on parent gender. More studies on parent depression and child PTSS are needed to examine possible parent gender differences. Further, the current study was unable to explore differences between studies of boy versus girl trauma victims or potential interactive effects between parent and child gender. Prior research has suggested that the gender characteristics of the parent child dyad (i.e., mother daughter, mother son, father daughter, and father son) may influence the relationship between parent symptoms and child outcomes (Reeb, Conger, & Wu, 2010; Russell & Saebel, 1997). Future research should examine the extent to which male and female children are differentially influenced by symptomatic distress in fathers and mothers. In addition to parent gender, type of study (longitudinal vs. cross-sectional) and assessment type (interview vs. self-report) moderated the relationship between parent and child PTSS. Consistent with our hypothesis, the pattern of effect sizes associated with study type suggested that longitudinal studies yielded a stronger association between parent and child PTSS. This finding is in line with the theory that longitudinal studies are more likely to capture several distinct trajectories of symptoms (Alisic et al., 2011). However, given the problems with subject retention in longitudinal studies, this finding may result from inherent differences in participants who typically dropout versus those who continue to participate in studies with multiple follow-up assessments. The significant moderator of parent assessment type suggests that the strength of the relationship of parent and child PTSS was greater in studies that used an interview format to assess parent symptoms. Further, studies that used the same type of assessment for parents and children yielded stronger relationships between parent and child PTSS. It is unclear how differences in assessment type contribute to differences in the effect sizes of the relationship between parent and child PTSS; however, the present findings suggest that there are differences between assessment measures of PTSD, with a higher consistency found for studies using parent interviews. Further, the moderating effect of same versus different assessment type may also partially explain the inconsistencies in research on the association between parent and child PTSS. Findings for several moderators ran counter to hypotheses. Notably, the use of criterion A as a study inclusion parameter did not moderate the strength of the association between parent and child PTSS. The usefulness of Criterion A as a screener has been a persistent problem in the PTSD literature, due in part to researchers ignoring the subjective component of trauma (Weathers & Keane, 2007). A majority of the studies included in the current meta-analysis inferred that events were traumatic, without asking participants for their subjective appraisal of the event. Our results suggest that this did not appear to influence the strength of the relationship between parent symptomatic distress and child PTSS. However, there may be an overlap between the use of the A-criteria and interview as mode of administration, as most child interviews also used criterion A as an inclusion criteria. Additional self-report studies using criterion A are necessary to tease apart this overlap. Finally, the type of index trauma (medical diagnosis vs. other) also did not moderate the strength of the associations between parent and child PTSS, suggesting that the relationship between parent and child PTSS is similar across medical and nonmedical diagnosis traumas, regardless of the objective nature of the event. Limitations and Future Directions The present investigation had several limitations. First, the number of studies included in the meta-analysis was constrained by the number of individual studies that investigated the associations between parent posttraumatic symptoms and child PTSS. Similarly, the types of moderating variables were limited to the methodological variables included in the individual studies. Furthermore,

9 1084 Morris, Gabert-Quillen, and Delahanty the current study was limited to calculating the association between two variables and did not include multiple interaction effects. To overcome this limitation, future meta-analyses should consider meta-regression techniques. Although requiring a much larger number of studies, metaregression allows for the investigation of continuous, as well as categorical variables, while also allowing for the effects of multiple factors to be investigated simultaneously (Thompson & Higgins, 2002). Despite these limitations, the current findings highlight a need to expand current assessment models of child PTSS to include the influence of parent posttraumatic distress and depression. Clinically, these results suggest a family-focused approach to the treatment of child PTSS after a trauma. Failure to address parental, primarily maternal, posttraumatic stress may hinder PTSD treatment in child trauma victims. Further, ongoing assessment of child and parent symptoms should be implemented in therapeutic approaches, given that the strength of the relationship between parent and child symptoms is strongest in longitudinally designed studies. References *indicates which studies were included in the current meta-analysis. Alisic, E., Jongmans, M. J., van Wesel, F., & Kleber, R. J. (2011). Building child trauma theory from longitudinal studies: A meta-analysis. Clinical Psychology Review, 31, American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed. revised). Washington, DC. *Applebaum, D., & Burns, G. (1991). Unexpected childhood death: posttraumatic stress disorder in surviving siblings and parents. Journal of Clinical Child Psychology, 20, Barakat, L. P., Kazak, A. E., Meadows, A. T., Casey, R., Meeske, K., & Stuber, M. L. (1997). Families surviving childhood cancer: A comparison of posttraumatic stress symptoms with families of healthy children. Journal of Pediatric Psychology, 22, *Ben-Amitay, G., Kosov, I., Reiss, A., Toren, P., Yoran- Hegesh, R., Kotler, M., & Mozes, T. (2006). Is elective surgery traumatic for children and their parents? Journal of Pediatrics and Child Health, 42, *Birmes, P., Raynaud, J. P., Daubisse, L., Brunet, A., Arbus, C., Klein, R.,... Schmitt, L. (2009). Children s enduring PTSD symptoms are related to their family s adaptability and cohesion. Community Mental Health Journal, 45, Bokszczanin, A. (2007). PTSD symptoms in children and adolescents 28 months after a flood: Age and gender differences. Journal of Traumatic Stress, 20, Borenstein, M., Hedges, L., Higgins, J., & Rothstein, H. (2005). Comprehensive Meta-analysis Version 2. Englewood, NJ: Biostat. Breslau, N., Chilcoat, H. D., Kessler, R. C., & Davis, G. C. (1999). Previous exposure to trauma and PTSD effects of subsequent trauma: Results from the Detroit area survey of trauma. American Journal of Psychiatry, 156, Brewin, C. R., Andrews, B., & Valentine, J. D. (2000). Meta-analysis of risk factors for posttraumatic stress disorder in trauma-exposed adults. Journal of Consulting and Clinical Psychology, 68, *Bronner, M. B., Knoester, H., Bos, A. P., Last, B. F., & Grootenhuis, M. A. (2008). Posttraumatic stress disorder (PTSD) in children after paediatric intensive care treatment compared to children who survived a major fire disaster. Child and Adolescent Psychiatry and Mental Health, 9, 1 9. *Brown, R., Fuemmeler, B., Anderson, D., Jamieson, S., Simonian, S., Hall, R., & Brescia, F. (2007). Adjustment of children and their mothers with breast cancer. Journal of Pediatric Psychology, 2, Brown, R. T., Madan-Swain, A., & Lambert, R. (2003). Posttraumatic stress symptoms in adolescent survivors of childhood cancer and their mothers. Journal of Traumatic Stress, 16, Bruce, M. (2006). A systematic and conceptual review of posttraumatic stress in childhood cancer survivors and their parents. Clinical Psychology Review, 26, *Bryant, B., Mayou, R., Wiggs, L., Ehlers, A., & Stores, G. (2004). Psychological consequences of road traffic accidents for children and their mothers. Psychological Medicine, 34, Bui, E., Brunet, A., Olliac, B., Very, E., Allenou, C., Raynaud, J.,... Birmes, P. (2011). Validation of the peritraumatic dissociative experiences questionnaire and peritraumatic distress inventory in school-aged victims of road traffic accidents. European Psychiatry, 26, *Coakley, R. M., Forbes, P. W., Kelley, S. D., LeBovidge, J., Beasley, P., DeMaso, D. R., & Waber, D. P. (2010). Family functioning and posttraumatic stress symptoms in youth and their

10 Parent PTSD/Depression and Child PTSS 1085 parents after unintentional pediatric injury. Journal of Traumatic Stress, 23, Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2 ed.). Hillsdale, NJ: Lawrence Earlbaum Associates. Connell, A. M., & Goodman, S. H. (2002). The association between psychopathology in fathers versus mothers and children s internalizing and externalizing behavior problems: A meta-analysis. Psychological Bulletin, 128, Copeland, W. E., Keeler, G., Angold, A., & Costello, E. J. (2007). Traumatic events and posttraumatic stress in childhood. Archives of General Psychiatry, 64, Costello, E. J., Erkanli, A., Fairbank, J. A., & Angold, A. (2002). The prevalence of potentially traumatic events in childhood and adolescence. Journal of Traumatic Stress, 15, Cox, C. M., Kenardy, J. A., & Hendrikz, J. K. (2008). A meta-analysis of risk factors that predict psychopathology following accidental trauma. Journal for Specialists in Pediatric Nursing, 13, Cuijpers, P., Li, J., Hofmann, S. G., & Andersson, G. (2010). Self-reported versus clinician-rated symptoms of depression as outcome measures in psychotherapy research on depression: A meta-analysis. Clinical Psychology Review, 30, de Vries, A. P. J., Kassam-Adams, N., Cnaan, A., Sherman-Slate, E., Gallagher, P. R., & Winston, F. K. (1999). Looking beyond the physical injury: Posttraumatic stress disorder in children and parents after pediatric traffic injury. Pediatrics, 104, Egger, M., Smith, G., Schneider, M., & Minder, C. (1997). Bias in meta-analysis detected by a simple, graphical test. BMJ, 315, Ewell Foster, C. J., Garber, J., & Durlak, J. A. (2007). Current and past maternal depression, maternal interaction behaviors, and children s externalizing and internalizing symptoms. Journal of Abnormal Child Psychology, 35, Fisak, B., & Grills-Taquechel, A. E. (2007). Parental modeling, reinforcement, and information transfer: Risk factors in the development of child anxiety? Clinical Child and Family Psychology Review, 10, Gleser, L. J., & Olkin, I. (1994). Stochastically dependent effect sizes. In H. M. Cooper, & L. V. Hedges (Eds.), The handbook of research synthesis (pp ). New York, NY: Sage. Goodman, S., Rouse, M., Connell, A., Broth, M., Hall, C., & Heyward, D. (2011). Maternal depression and child psychopathology: a meta-analytic review. Clin Child Fam Psychol Rev, 14, Green, B., Korol, M., Grace, M., Vary, M., Leonard, A., Gleser, G., & Smitson-Cohen, S. (1991). Children and disaster: Age, gender, and parental effects on PTSD symptoms. Journal of the American Academy of Child and Adolescent Psychiatry, 30, *Hall, E., Saxe, G., Stoddard, F., Kaplow, J., Koenen, K., Chawla, N.,... King, D. (2006). Posttraumatic stress symptoms in parents of children with acute burns. Journal of Pediatric Psychology, 31, Hunter, J. E., & Schmidt, F. L. (2004). Methods of meta-analysis. Thousand Oaks, CA: Sage Publications. Johnson, B. T., & Eagly, A. H. (2000). Quantitative synthesis of social psychological research. In H. T. Reis & C. M. Judd (Eds.), Handbook of research methods in social psychology (pp ). Cambridge, England: Cambridge University Press. *Jones, R. T., Ribbe, D. P., Cunningham, P. B., Weddle, J. D., & Langley, A. K. (2002). Psychological impact of fire disaster on children and their parents. Behavior Modification, 26, Kahana, S. Y., Feeny, N. C., Youngstrom, E. A., & Drotar, D. (2006). Posttraumatic stress in youth experiencing illnesses and injuries: An exploratory meta-analysis. Traumatology, 12, *Kassam-Adams, N., Fleisher, C. L., & Winston, F. K. (2009). Acute stress disorder and posttraumatic stress disorder in parents of injured children. Journal of Traumatic Stress, 22, Kassam-Adams, N., & Winston, F. (2004). Predicting child PTSD: The relationship between acute stress disorder and PTSD in injured children. Journal of American Academy of Child and Adoescent Psychiatry, 43, *Kazak, A., Alderfer, M., Rourke, M., Simms, S., Streisand, R., & Grossman, J. (2004). Posttraumatic stress symptoms (PTSS) and posttraumatic stress disorder (PTSD) in families of adolescent childhood cancer survivors. Journal of Pediatric Psychology, 29, *Kilic, E. Z., Özguven, H. D., & Sayil, I. (2003). The psychological effects of parental mental health on children experiencing disaster: The experience of bolu earthquake in Turkey. Family Process, 42, Kilpatrick, D. G., Ruggiero, K. J., Acierno, R., Saunders, B. E., Resnick, H. S., & Best, C. (2003).

11 1086 Morris, Gabert-Quillen, and Delahanty Violence and risk of PTSD, major depression, substance abuse/dependence, and comorbidity: Results from the national survey of adolescents. Journal of Consulting and Clinical Psychiatry, 71, Koenen, K. C., Amstadter, A. B., & Nugent, N. R. (2009). Gene-environment interaction in posttraumatic stress disorder: An update. Journal of Traumatic Stress, 22, Koenen, K. C., Fu, Q. J., Ertel, K., Lyons, M. J., Eisen, S. A., True, W. R.,... Tsuang, M. T. (2008). Common genetic liability to major depression and posttraumatic stress disorder in men. Journal of Affective Disorders, 105, *Koplewicz, H. S., Vogel, J. M., Solanto, M. V., Morrissey, R. F., Alonso, C. M., Abikoff, H.,... Novick, R. M. (2002). Child and parent response to the 1993 world trade center bombing. Journal of Traumatic Stress, 15, *Landolt, M. A., Vollrath, M., Ribi, K., Gnehm, H. E., & Sennhauser, F. H. (2003). Incidence and associations of parental and child posttraumatic stress symptoms in pediatric patients. Journal of Child Psychology and Psychiatry, 44, Le Brocque, R., Hendrikz, J., & Kenardy, J. (2010). The course of post traumatic stress in children: Examination of recovery trajectories following traumatic injury. Journal of Pediatric Psycholog, 35, Lipsey, M. W., & Wilson, D. B. (2000). Practical meta-analysis. Thousand Oaks, CA: Sage Publications. *Magal-Vardi, O., Laor, N., Toren, A., Strauss, L., Wolmer, L., Bielorai, B.,... Toren, P. (2004). Psychiatric morbidity and quality of life in children with malignancies and their parents. The Journal of Nervous and Mental Disease, 192, Manne, S. L., Du Hamel, K., Gallelli, K., Sorgen, K., & Redd, W. H. (1998). Posttraumatic stress disorder among mothers of pediatric cancer survivors: Diagnosis, comorbidity, and utility of the PTSD checklist as a screening instrument. Journal of Pediatric Psychology, 23, *Meiser-Stedman, R. A., Yule, W., Dalgleish, T., Smith, P., & Glucksman, E. (2006). The role of the family in child and adolescent posttraumatic stress following attendance at an emergency department. Journal of Pediatric Psychology, 31, *Morris, A. (2010). Caregiver characteristics and the development of PTSD following traumatic injury in children: The impact of affectionless control and subjective appraisals on post-injury adjustment. Retrieved from OhioLINK ETD ( ). Naragon-Gainey, K. (2010). Meta-analysis of the relations of anxiety sensitivity to the depressive and anxiety disorders. Psychological Bulletin, 136, *Nugent, N., Ostrowski, S., Christopher, N., & Delahanty, D. (2006). Parental posttraumatic stress symptoms as a moderator of child s acute biological response and subsequent posttraumatic stress symptoms in pediatric injury patients. Journal of Pediatric Psychology, 32, O Donnell, M., Creamer, M., McFarlane, A., Silove, D., & Bryant, R. (2010). Should A2 be a diagnostic requirement for posttraumatic stress disorder in DSM-V? Psychiatry Research, 176, O Donnell, M. L., Elliott, P., Lau, W., & Creamer, M. (2007). PTSD symptom trajectories: From early to chronic response. Behaviour Research and Therapy, 45, Orwin, R. G. (1983). A fail-safe N for effect size in meta-analysis. Journal of Educational Statistics, 8, *Ostrowski, S. A., Ciesla, J. A., Lee, T. J., Irish, L., Christopher, N. C., & Delahanty, D. L. (2011). The impact of caregiver distress on the longitudinal development of child acute post-traumatic stress disorder symptoms in pediatric injury victims. Journal of Pediatric Psychology, 36, Ostrowski, S. A., Christopher, N., & Delahanty, D. L. (2007). Brief report: The impact of maternal posttraumatic stress disorder symptoms and child gender on risk for persistent posttraumatic stress disorder symptoms in child trauma victims. Journal of Pediatric Psychology, 32, *Ozono, S., Saeki, T., Mantani, T., Ogata, A., Okamura, H., & Yamawaki, S. (2007). Factors related to posttraumatic stress in adolescent survivors of childhood cancer and their parents. Supportive Care in Cancer, 15, *Pelcovitz, D., Libov, B. G., Mandel, F., Kaplan, S., Weinblatt, M., & Septimus, A. (1998). Posttraumatic stress disorder and family functioning in adolescent cancer. Journal of Traumatic Stress, 11, *Phipps, S., Larson, S., Long, A., & Rai, S. N. (2006). Adaptive style and symptoms of posttraumatic stress in children with cancer and their parents. Journal of Pediatric Psychology, 31, Pynoos, R. S., Steinberg, A. M., Layne, C. M., Briggs, E. C., Ostrowski, S. A., & Fairbank, J. A. (2009). DSM-V PTSD diagnostic criteria for children and adolescents: A developmental perspective and