Predictors and correlates
Abramson et al. (2010)
Four annual waves from Jan–Aug 2006 (5–12 months post Katrina) to Nov. 2009–2010 (50–54 months)
Community/displaced; age: 4–17 years
Gender and ethnicity not given
Serious emotional disturbance (SED; psychological, emotional, and behavioral dysfunction)
Multiple individual, social, and contextual factors (e.g., Parental mental health, living situation); parental and child’s prior mental health; hurricane exposure
Banks and Weems (2014)
24 and 30 months post Katrina
School age: 8–15 years; 55 % male; 99 % minority
PTSD, anxiety, and depression
Parent and peer social support
Hensley-Maloney and Varela (2009)
6.5 and 17.5 months post Katrina
School age: 10–15; 61 % female; 60 % minority
Panic symptoms at T1 and T2
Child’s age, gender, parent level of education, family income, ethnicity, and panic at T1; anxiety sensitivity at T1 and T2; Hurricane trauma exposure
Kelley et al. (2010)
5 and 15.5 months post Katrina
Community; age: 8–16 years;76 % minority; gender not given
PTSD symptoms at T2
PTSD symptoms at T1; hurricane exposure and violence exposure
Kilmer et al. (2009)
12 and 24 months post Katrina
Community age: 7–10 years
54 % female; 85 % minority
Posttraumatic growth at T2
Hurricane exposure; posttraumatic growth at T1; PTSD symptoms at T1and T2
Kilmer and Gil-Rivas (2010)
12 and 24 months post Katrina
Community age: 7–10 years; 56 % female; 85 % minority
Child’s posttraumatic growth at T2
Child report of multiple cognitive factors; PTSD symptoms and posttraumatic growth at T1; parent report of parenting behaviors, parental posttraumatic growth, and parental mental health at T1, PTSD symptoms at T1 and T2
Kronenberg et al. (2010)
24 and 36 months post Katrina
School/displaced age: 9–18 years; 54.5 % female; 27 % minority
PTSD and depression symptoms at T2; Individual symptom recovery patterns from T1 to T2
Hurricane exposure; multiple cognitive, social, and contextual factors; previous trauma, multiple life stressors (e.g., school problems), PTSD and depression symptoms at T1
McLaughlin et al. (2009)
Longitudinal/cross-sectional analyses only; three annual waves from Mar to Nov 2007 and Aug to Nov 2008 (18–27 months post Katrina)
Community age: 4–17 years; gender and ethnicity not given
Parent report of child’s hurricane- and non-hurricane-related serious emotional disturbance (psychological, emotional, and behavioral dysfunction)
Multiple demographic (age, gender, ethnicity, etc.) and contextual factors; Parent report of child’s hurricane exposure (e.g., death of a loved one), child’s hurricane-related stress, parent’s mental health
Ontogenic, microsystem, exosystem
McLaughlin et al. (2010)
Three annual waves from Mar–Nov 2007 and Aug–Nov 2008 (18–39 months post Katrina)
Community age: 4–17 years; gender and ethnicity not given
Parent report of child’s hurricane- and non-Hurricane-related serious emotional disturbance (SED; psychological, emotional, and behavioral dysfunction) at T1 and T2
Multiple demographic (e.g., age, gender, ethnicity, etc.) and contextual factors, parent report of child’s pre-hurricane functioning (parent retrospective report); hurricane-related stress at T1 and T2
Exosystem, ontogenic, microsystem
Rohrbach et al. (2009)
13 months pre-Rita and follow-ups at 7 and 19 months post Rita
School age: ninth graders (M = 14.4 years); 68 % female; 32 % minority
Substance abuse at T2 and T3
Age, gender, and ethnicity; substance abuse at T1; hurricane exposure, life disruption, negative life events and PTSD symptoms at T2
Microsystem, ontogenic, exosystem
Ward et al. (2008)
Academic years: 2004–2005 (pre-Katrina), 2005–2006 (post Katrina), and 2006–2007 (post Katrina)
School/displaced age: Pre-K to grade 12
Ethnicity, SES, school attendance, disciplinary problems (e.g., suspension, expulsion), achievement, grade retention, and dropout rates at T2 and T3
School attendance, disciplinary problems (e.g., suspension, expulsion), achievement, grade retention, and dropout rates at T1
24 and 30 months post Katrina
School age: 8–15 years; 55 % male; 99 % minority
PTSD symptoms and PTSD severity Group membership at T2; test anxiety; academic achievement
Age, gender, hurricane exposure, life events, PTSD symptoms, PTSD severity group Membership at T1
Ontogenic, microsystem, macrosystem
24 and 30 months post Katrina and 1 month post Gustav
Grades fourth through eighth
Identified trajectories of PTSD symptoms; PTSD symptoms
Subsequent exposure, non-avoidant coping; TV viewing of Gustav
Weems et al. (2014)
94 and 141
Sample 1: 13, 20, and 26 months post Katrina and 5 months post Hurricane Gustav
Sample 2: Time 1; 24 months post Katrina) and 1 month and 8 months post Gustav
Sample 1; (initial grade 9th followed to 11th Sample 2; grades 4th through 8th
Memories of Katrina
The ontogenic level is the ecology of the individual and represents factors within the individual that influence developmental adaptation. First at this level is what the literature indicates about the course of mental health following disasters. As noted earlier, while there appear to be general declines over time, researchers have begun to define and identify multiple trajectories of symptoms following disasters (Bonanno et al., 2010, 2011; see Chap. 1, this volume) and the importance of identifying resilient subsamples (Masten & Obradovic, 2008). In other words, there may not be just one single course but a set of potential long-term outcomes. Conceptually, resilience emphasizes “that some children facing adversity nonetheless do well (or return to positive functioning following a period of maladaptation)” (Sroufe, 1997, p. 256) . Resilience then is defined by (1) exposure to some risk (e.g., facing disaster-related adversity) as well as (2) the relatively positive functioning of some compared to others. As a subcomponent of (1), the “resilient” should have the same level of risk exposure as the non-resilient (for related discussion, see Luthar, Cicchetti, & Becker, 2000) . Clearly, exposure to a disaster represents risk, but samples of disaster-exposed youth can be very heterogeneous with regard to the level of exposure. Thus, a group with a stable low trajectory may seem “resilient” but in fact may simply be relatively less exposed to traumatic experiences.
Weems and Graham (2014) followed up at a third assessment point (right after Hurricane Gustav) in a study of the stability of PTSD symptoms among ethnic minority youth in grades 4 through 8 who were assessed at 24 and 30 months post Hurricane Katrina. The investigators initially found a generally high level of stability in PTSD symptoms and that younger age, being female, and continued disrepair to the child’s home were predictive of stable, elevated PTSD symptoms (Weems et al., 2010). Weems and Graham found that there were general declines in symptoms. However, subsamples of youth with increasing, decreasing, chronic, and stable low levels of PTSD symptoms were identified, consistent with previous theorizing. High subsequent Gustav exposure was associated with more chronic symptoms, while low Gustav exposure was associated with declines in symptoms. A relatively large group (43 % of the sample) with a stable low trajectory was identified. The stable low trajectory group had significantly fewer exposure experiences, although the level of exposure was not homogeneous. A subsample of the stable low group was defined as resilient. The identification of this resilient group was made according to commensurability in level of risk exposure (total sample average or higher levels of disaster experiences) with the non-resilient group (i.e., chronic/stable high). In this case, a much smaller portion (n = 16 %) was considered resilient. Moreover, comparison of all stable low with stable high would have led to an erroneous conclusion about the use of avoidant coping strategies. It was the truly resilient (low symptoms and high exposure) who reported the lowest levels of avoidant coping strategies.
Kronenberg et al. (2010) assessed youth aged 9–18 years for symptoms of PTSD and depression at 2 and 3 years following Hurricane Katrina and classified youth into the outcome trajectories of stress resistant, normal response and recovery, delayed breakdown, and breakdown without recovery. One encouraging finding was that overall symptomatology decreased between the 2nd and 3rd years, and the majority (72.3 %) of participants were classified as either stress resistant or having a normal response and recovery. However, the remaining participants, who reported higher hurricane exposure, continued to report significant symptoms of PTSD and depression, and were classified as breakdown without recovery or delayed breakdown. Younger age, being female, and the presence of family concerns (e.g., parent still unemployed following the storm) were all associated negative responses to the disaster and poor recovery patterns.
Disaster experiences that involve life threat can challenge one’s sense of control and self-efficacy in containing the threat, leading to emotional reactions that may be difficult to regulate (Norris et al., 2002) and ultimately increasing the risk for the development of psychopathology. Theoretically, at the individual level, direct exposure to life threat in a disaster increases activity of the hypothalamic–pituitary–adrenal (HPA) axis as part of a normative fight–flight reaction (see Chap. 9, this volume). Fear reactions are associated with elevations in the secretion of cortisol, a corticosteroid hormone produced by the adrenal cortex that can be assayed from blood, urine, or saliva samples (see Nadar & Weems, 2011, for a review). The intense taxing of this system in disasters may lead to dysregulation of the system. Dysregulation in this system has been associated with a number of emotional and behavioral disorders (Weems & Carrión, 2009; Yehuda, 2006) . Research suggests that after a period of relative cortisol hypersecretion, elevated levels may reverse in trauma-exposed individuals (Weems & Carrión, 2007) to relatively low levels of cortisol (Yehuda, 2006). For example, Pfeffer, Altemus, Heo, and Jiang (2007) examined cortisol levels among bereaved youth following the September 11 attacks, finding that while bereaved youth had higher cortisol levels, those with PTSD had relatively lower evening cortisol levels than those without PTSD. This low cortisol may result from an enhanced negative feedback loop at the pituitary–adrenal level of the HPA axis (Yehuda et al., 1995) . One mechanism for this sensitization may be an increased number of glucocorticoid receptors in the HPA axis that facilitates the negative feedback loop (Yehuda, 2006).
The experience of traumatic stress has been associated with atypical brain development in youth, including attenuated volume in brain regions that are involved in cognitive emotional processing such as the hippocampus (Carrión, Weems, & Reiss, 2007). Changes in neural structure and function following stress (Carrión et al., 2007; Carrión, Garrett, Menon, Weems, & Reiss, 2008; Carrión, Weems, Richert, Hoffman, & Reiss, 2010) , and susceptibility to dysregulation in the normative stress response may characterize individual risk for mental health problems among disaster victims. Recently, La Greca, Lai, Joormann, Auslander, and Short (2013) found molecular genetic evidence for susceptibility, in that they showed that while greater exposure to hurricane stressors was related to more symptoms of PTSD and depression in a sample of youth, this effect was stronger for children with the met allele of the brain-derived neurotrophic factor (BDNF). There is also some evidence to suggest that differential brain activation may occur between trauma and non-trauma-exposed youth. Yang et al. (2011) examined a sample of Taiwanese adolescents using functional magnetic resonance imaging (fMRI) 14 months after exposure to a severe earthquake. Following the presentation of traumatic reminders (words) of the earthquake, the PTSD participants demonstrated activation in the bilateral visual cortex, bilateral cerebellum, and left parahippocampal gyrus; such activation was not present among the control participants.
Research has suggested that preexisting characteristics of the child can influence the impact of disaster exposure on mental health outcomes. In particular, previous research has documented that pre-hurricane trait anxiety and negative affect levels predict PTSD symptoms above and beyond exposure to the trauma (La Greca, Silverman, & Wasserstein, 1998; Weems et al., 2007). Anxiety sensitivity is another trait that has been studied in relation to hurricane exposure and panic (Hensley-Maloney & Varela, 2009) . The authors assessed youth at 5–8 months and 17–18 months post Katrina and found that while hurricane exposure and anxiety sensitivity did predict panic symptoms at the initial assessment, the association between exposure and panic was not significant when youth were assessed at 17–18 months post disaster. The authors theorized that the temporal association between the trauma and panic symptoms could be due to an elevation in anxiety sensitivity following trauma which then generalizes to non-trauma-related cues and ultimately results in anxiety sensitivity having an association with panic over time.
One important consideration with regard to post-disaster outcomes over time is disaster-related secondary stressors (Silverman & La Greca, 2002). These stressors may function as cascading or cumulative events which then have the potential to promote the persistence of long-term psychopathology. For example, McLaughlin et al. (2010) found that several Hurricane Katrina-exposed children and adolescents developed new serious emotional disturbance (SED) during the period between baseline (18–27 months post Katrina) and follow-up 12–18 months later, and the onsets of SED during this period were more frequent among participants reporting high stress following from disaster-related secondary adversities. Additionally, post-hurricane negative life events such as family illness, financial strain, and arguments with parents have been found to be significant predictors of adolescents’ increased substance use (i.e., alcohol, marijuana, and cigarettes) from 13 months pre Hurricane Rita to 19 months post disaster (Rohrbach, Grana, Vernberg, Sussman, & Sun, 2009) . This finding is in line with the extant literature indicating that adverse life events are associated with psychological distress, which may lead to substance use in an effort to alleviate the distress (Newcomb, Huba, & Bentler, 1986; Wills, 1986; Wills, Vaccaro, & McNamara, 1992) .
The microsystem is the next level up from the individual child and represents the proximal ecologies within which the child develops, including the family/home and school environments and peer relationships (see Chap. 12, this volume) . Disasters can affect youth adaptation by causing stress and disruptions in the family, school, and neighborhood environments. Disaster research on the microsystems impacted by disaster has tended to focus on the family environment, with studies showing that marital stress, domestic violence, and parental psychopathology increase after disasters (Larrance, Anastario, & Lawry, 2007; Norris et al., 2002). Research by Scheeringa and Zeanah (2008) and Spell et al. (2008) shows the importance of parental mental health for child functioning at two distinct developmental points. For preschool children, Scheeringa and Zeanah (2008) found that the onset of new mental health problems in preschool children was significantly correlated with the onset of new mental health problems in their caregivers. For school-age children (8–16 years), Spell et al. (2008) obtained a similar finding among a sample of 260 displaced mother–child dyads impacted by Katrina with maternal psychological distress a significant predictor of child mental health .
Focusing on long-term outcomes, Abramson et al. (2010) reported that data collected in a longitudinal study across 4 years post Katrina showed that youth whose parents reported mental health illness were more than five times as likely to have serious emotional disturbance than youth whose parents did not suffer from such distress. As noted by Scaramella, Sohr-Preston, Callahan, and Mirabile (2008) , parental emotional distress and mental illness undermine parenting efficacy, resulting in increases in parental irritability and decreases in consistent discipline, both of which increase the risk for child mental health problems. In a longitudinal study, Kelley et al. (2010) found that youth who reported greater hurricane exposure had parents who reported more maladaptive coping; these parents were then more likely to employ the use of corporal punishment, which was associated with youths’ PTSD symptoms at both 3–7 and 14–17 months post disaster .
Disasters may also result in the disruption of the school microsystem. There is evidence to suggest an indirect effect of disaster exposure (i.e., indirectly through PTSD symptoms fostering increased test anxiety) on youths’ academic achievement (standardized test scores), a full 30 months following Katrina (Weems et al., 2013). Disasters may also necessitate relocation to another school. Ward, Shelley, Kaase, and Pane (2008) conducted a longitudinal study to examine the differences in achievement and school-related behavior between displaced and non-displaced students in Mississippi 2 years after Hurricane Katrina. The results indicated that displaced students were more likely to have disciplinary problems (i.e., suspension and expulsion) and had lower academic achievement than their non-displaced peers . Based on data collected 1 year before Hurricane Katrina, the authors determined that there were preexisting differences in behavior and achievement between the two groups before the storm, but there was an increase in suspension and expulsion post disaster for both displaced and non-displaced students. Additionally, displaced students had greater odds of being suspended or expelled, suggesting that not only were the students previously demonstrating disciplinary problems but that the forced relocation may also have amplified the effects of the trauma and been associated with the increase in disciplinary problems .
A critical factor of youths’ microsystems is the social support received. Banks and Weems (2014) found that higher levels of social support from family and peers were associated with lower levels of psychological distress (i.e., symptoms of PTSD , anxiety, and depression) in a large sample of minority youth when assessed 36–65 months after Hurricane Katrina. The results from the study also showed that there was an association between lower social support and higher psychological distress longitudinally, similar to other studies that have examined longitudinal associations between social support and post-disaster distress in youth (e.g., Jaycox et al., 2010; La Greca, Silverman, Vernberg, & Prinstein, 1996; La Greca, Silverman, Lai, & Jaccard, 2010) . Peer social support, in particular, showed multiple associations with lower distress concurrently and longitudinally in Banks and Weems’ study, which makes theoretical sense as peer relationships become more salient across the developmental range of youth included in the study and so youth may have turned to peers more so than family for assistance in coping and decreasing isolation following the disaster (e.g., Pynoos & Nader, 1988; Vernberg & Vogel, 1993). Importantly, however, the findings from the study suggested the possibility that high hurricane exposure may potentially overwhelm the benefit of peer social support against the development of PTSD symptoms for some youth .
Increased risk created by disasters within microsystem environments may be offset by the presence of protective factors within the other microsystems surrounding the child. For example, school-based mental health services represent a protective factor within the school microsystem that can offset the negative developmental outcomes associated with disaster exposure (Abramson & Garfield, 2006; Pynoos, Goenjian, & Steinberg, 1998). Salloum and Overstreet (2008) and Weems et al. (2009, 2015) demonstrate the effectiveness of school-based interventions for children in post-disaster environments. Weems et al. (2009, 2015) tested the effects of a school-based test anxiety intervention on reducing PTS symptoms in which Hurricane Katrina-exposed youth with elevated test anxiety completed a primarily behavioral (e.g., relaxation training combined with gradual exposure to anxiety-provoking test-related stimuli) group-administered, test-anxiety-reduction intervention. Their findings yielded a statistically significant effect of the intervention on test anxiety levels and academic performance with evidence of positive secondary effects on PTS symptoms. Moreover, change in test anxiety predicted change in PTS, and there appeared to be no negative effects on natural PTS symptom decline. Test anxiety was highly associated with anxiety and depressive symptoms and was also predictive of PTSD symptoms .