This research was supported in part by a grant from Massey University (MURF: 1-0575- 67214A). Thanks to Ohakune and Waiouru Primary Schools and especially their children. The author wants to thank Tauri Morgan and Ellen Gould for their invaluable assistance.
Examined self-reported symptoms of post-traumatic stress in 118 children following a series of volcanic eruptions. A significant number of children reported symptoms that met criteria for the following symptom clusters: reexperiencing (65%), hyperarousal (24%), and psychic numbing/avoidance (14%). Additionally, 11% of the children reported symptoms that met criteria for all three of the clusters. In terms of the mediating effects of gender, age, and asthma, generally no differences were found in symptom endorsement with the exception of younger and asthmatic children reporting a greater frequency of hyperarousal symptoms. Anxiety and depression levels of children who met criteria for all three symptom clusters were found to be at clinically significant levels. Results are compared with findings from a study that looked at some of these issues following a more acute and more catastrophic disaster (Hurricane Andrew). The use of the current methodology in screening large samples of children and in school-based interventions following a natural disaster is discussed.
LaGreca et al. (in press) examined the frequency of self-reported PTSD symptom cluster
endorsement in a large sample of children following Hurricane Andrew in Florida, U.S. in 1992. A
large number of children met criteria for each symptom cluster three months following the hurricane:
89.8% of the children met criteria for the reexperiencing symptom cluster; 49.3% of the children met
criteria for the psychic numbing/avoidance cluster; 67% of the children met criteria for the
hyperarousal symptom cluster. Additionally, fully 39.1% of the children met criteria for all three
The current study addresses this issue of self-reported symptom cluster endorsement in a
sample of children following a series of volcanic eruptions (Mount Ruapehu) on North Island, New
Zealand in September, 1995. Unlike Hurricane Andrew, Mount Ruapehu did not result in loss of
human life or widespread damage. The primary problems following the eruptions were instead ashfall
in local communities and infrastructure disruptions resulting from increased ashfall (e.g., electrical
generator damage; Johnson, Ronan, & Houghton, in press). Also, unlike Hurricane Andrew, the
volcanic eruptions were not acute events with a defined endpoint. In fact, the first series of eruptions
continued for over a month followed by a second series of eruptions approximately 9 months later.
Thus, one aim of the current study was to examine the issue of frequency of symptom cluster
endorsement after a less catastrophic but more chronic natural hazard.
Another aim of the study was to examine symptom cluster endorsement as a function of age,
gender, and ashtmatic status. In terms of total scores on PTSD-related measures, younger children
and girls have been found to report higher total scores on these measures than boys in the more
immediate aftermath of a natural disaster (Belter & Shannon, 1993; Shannon et al., 1994; Vernberg et
al., in press). However, research has not looked into the effects of age and gender predicting the
increased likelihood of children meeting criteria for the PTSD symptom clusters (instead comparing
total scores on these measures). For example, it might be hypothesized that younger children would
have a tendency to experience greater frequencies of hyperarousal symptoms in line with their
developmental level. In terms of asthma, Ronan (1996a) hypothesized that asthmatic children may have viewed ashfall as
a potential precipitant for an asthma attack and thus exacerbate PTSD-related distress. This study
found that asthmatic children in general had higher scores on self-report indices of distress following
the volcanic eruptions. These findings were confirmed by parent and teacher reports. However, that
study did not examine the frequency of asthmatic versus nonasthmatic children who met criteria for
any of the PTSD symptom clusters. The question here was do asthmatic children report--in addition
to greater levels of distress--an increased frequency of actual symptom clusters. Given the earlier
findings (Ronan, 1996a) and what is known about about asthmatic children in terms of their potential
for hyperarousal when stressed (Butz & Alexander, 1993), it was hypothesized that asthmatic children
would be more likely to meet criteria for the hyperarousal cluster and for all three clusters. The final
major aim of the current study was to examine the level of self-reported anxiety and depression in
those children who did and did not meet criteria for each symptom cluster as well as those who did
and did not meet full symptom criteria. The question here relates to the general finding reported by
Keppel-Benson and Ollendick (1993) that traumatized samples may not differ from normative samples
in terms of scores on commonly used measures of anxiety and depression. Thus, the current study
looked at the mean scores on two commonly used measures of anxiety and depression for the whole
sample as well as those who did and did not meet symptom criteria. The hypothesis here was that
the total mean score of the current sample on these measures would not differ from the mean score
of the relevant normative sample for each measure. Additionally, the mean anxiety and depression
scores for those children who met symptom criteria were hypothesized to be significantly greater than
mean scores for those children who did not meet criteria as well as greater than the mean score of
each measure's normative sample.
Children were assessed in the month following the eruptions. Assessment was multitrait/multimethod: a battery of self-report measures, a teacher global rating, and a parent measure assessed the impact of the eruptions on a range of emotional, behavioural, cognitive, physiological, and coping factors in the children. The current study was primarily interested in answering questions related to selected instruments from this battery. These measures are now described.
Reaction Index. This 20 item measure is rated on a 5 point Likert scale (scored 0-4) and assesses features and symptoms of PTSD/Acute Stress Disorder. Originally, the RI was developed
based on DSM PTSD diagnostic criteria and intended to be used as a semistructured interview
procedure (Frederick, 1985). Widely used in research following disasters, it has recently been revised
to be used as either a structured interview or questionnaire (Frederick et al., 1992). An advantage of
the RI is its ability to assess PTSD symptoms in relation to the specific traumatic event (see below for
examples). Reliability and validity have been documented in previous studies (e.g., Frederick et al.,
1992; LaGreca et al., in press; Vernberg et al., in press). In the current study, alpha reliability was
found to be .88.
Vernberg et al. (in press) and LaGreca et al (in press) identifed subscales associated with
the three primary PTSD symptom clusters. In following these studies, the current study used these
factors to address questions related to PTSD symptom and diagnostic criteria. The first factor,
Reexperience phenomena included the following: Do you go over in your what happened--that is, do you see pictures in your mind or hear sounds in your mind about the volcano. Do thoughts about
the volcano come back to you even when you don't want them to? Do you have good or bad dreams
about the volcano or other bad dreams? Do things sometimes make you think it might happen again?
The second subscale, Psychic numbing/avoidance, included: Do you feel as good about things you liked to do before the volcano? Do you feel more alone inside, or more alone with your
feelings? Do you feel so scared, upset, or sad that you couldn't even talk or cry? Do you want to
stay away from things that make you remember what happened to you during the volcano?
The final factor, Hyperarousal, included: Do you startle more easily or feel more jumpy or nervous than before the volcano? Do you sleep well? Is it as easy to pay attention as before the
volcano? When something reminds you of the volcano, do you get tense or upset?
Using these symptom clusters, it was determined whether children met DSM-related criteria
for a particular symptom cluster. Thus, children "met criteria" for Reexperiencing if they endorsed at least one item (i.e, a rating of 2, 3, or 4; see LaGreca et al., in press); for Psychic numbing/avoidance, at least three items; for Hyperarousal, at least two items (APA, 1994). Additionally, the number of asthmatic versus non-asthmatic children who reported all three symptom
clusters was also determined to establish the number of these children who "met criteria" for a PTSD
diagnosis. While this allows for an approximation of criteria used to establish PTSD diagnosis, it
should not be construed as a substitute for a diagnosis as established through a structured diagnostic
interview (LaGreca et al., in press).
State-Trait Anxiety Inventory for Children (STAIC). This measure assesses both situational anxiety (STAIC-State) and prevailing tendencies to experience anxiety (STAIC-Trait)
(Spielberger, 1973). Each scale has 20 items rated on a 3-point Likert scale (scored 1, 2, or 3).
Reliability and validity data are extensive and have been documented (e.g., Kendall & Ronan, 1990).
Normative data are available (Spielberger, 1973).
Children's Depression Inventory (CDI). This 27-item inventory assesses affective, behavioural, and cognitive signs of depression (Kovacs, 1981). Each item has three choices from
which to choose that characterizes the child over the past two weeks (each item is then scored 0, 1,
2). Reliability and validity data are extensive (see review by Kendall, Cantwell, & Kazdin, 1989) and
normative data have been provided (Finch, Saylor, & Edwards, 1985).
The battery of measures was administered in small group settings and completed in approximately
30-45 minutes. Informed consent procedures were read aloud prior to assessment and children were
also told explicitly that they could choose not to participate and an alternative activity was available if
they so chose. Additionally, children were told that each item had no right or wrong answers, the
issue was "what you think and feel." After obtaining written consent from children, the instructions for
the first (and all those subsequent) measure was read aloud. Following instructions for each
measure, each item on the measure was then read aloud.
Age and Gender and Self-Reported Symptom Clusters
Frequencies of children who met criteria for each symptom cluster separately as well as those who met criteria for all three symptom clusters were calculated and compared. Using chi- square analysis, no differences were found in the frequency of children who met criteria for any or all symptom clusters as a function of gender or age, all p's > .20 with the exception of a significantly greater proportion of younger children who met criteria for the hyperarousal symptom cluster, X2 (8) = 16.31743, p < .05.
Asthma Status and Self-Reported Symptom Clusters
Table 1 presents data concerning the relationship between asthma and symptom presentation. Frequencies (and percentages) of asthmatic versus non-asthmatic children who met symptom criteria for each cluster and overall diagnosis are presented. Chi-square analysis (or Fisher's Exact Test for analyses having a cell that has fewer than 5 subjects) indicated no differences in proportions of asthmatic versus non-asthmatic children who met criteria for (a) reeexperiencing symptom cluster (b) psychic numbing/avoidance symptom cluster, and (c) all three symptom clusters, all p's > .10. However, regarding the hyperarousal symptom cluster, chi-square analysis was significant, X2 (1) = 5.22, p < .05, indicating that asthmatic children self-reported a significantly higher proportion of symptoms that met criteria for the hyperarousal symptom cluster. As may be seen in Table 1, 38% (13/34) of asthmatic children met criteria whereas only 18% (14/79) of non-asthmatic children met criteria for this symptom cluster.
|Asthma (n=34)||Non-Asthmatic (n=79)|
|Met||Did not meet||Met||Did not meet|
|Re-Experience||23 (65%)||12 (35%)||52 (66%)||27 (34%)|
|Numb/Av||5 (15%)||29 (85%)||11 (14%)||68 (86%)|
|Hyper||13 (38%)||21 (62%)||14 (18%)||65 (82%)|
|Met all three||4 (12%)||30 (88%)||7 (8%)||72 (91%)|
Diagnostic Status and Level of Self-Reported Anxiety and Depression
Scores on the STAIC State and Trait scale and the CDI were calculated for children who did and did not meet criteria for symptom clusters. These scores are presented in Table 2. In every instance, t-test comparison confirmed a difference between groups indicating that children who met criteria for symptom clusters also showed greater self-reported anxiety and depression.
|Met||Did not meet||Met||Did not meet||Met||Did not meet|
|Re-Experience||31.44 (8.4)||25.45 (4.5)||34.17 (8.8)||27.14 (6.0)||13.56 (10.1)||5.94 (5.9)|
|Numb/Av||41.44 (8.1)||27.32 (5.8)||43.18 (7.2)||29.67 (7.1)||22.93 (10.0)||9.12 (8.0)|
|Hyper||36.10 (8.9)||26.98 (5.9)||40.00 (7.5)||29.07 (7.2)||19.23 (8.6)||8.41 (8.4)|
|Met all three||41.83 (8.3)||27.86 (6.4)||44.67 (5.8)||30.19 (7.5)||25.40 (5.9)||9.59 (8.6)|
Clinical Significance: Normative Comparison
Finally, the scores on each measure (STAIC State and Trait and the CDI) were compared with scores from the normative samples. Looking first at the mean score for the entire sample (both those who met criteria and did not meet criteria), normative comparison indicated that the current sample's mean scores on the STAIC-S (M = 29.38, SD = 7.9), STAIC-T (M = 31.8, SD = 8.6), and CDI (M = 11.10, SD = 9.6) did not differ (by 1 SD or more) from those of relevant normative sample. The same pattern held for mean scores on these measures within the subsample of children who did not meet criteria-- that is, all means for the "did not meet criteria" subsample all were within 1 SD of the mean scores of the relevant normative samples. By contrast, clinically significant scores were indicated on the following measures: STAIC-S, children who met criteria for Psychic Numbing/Avoidance and for all three symptom clusters had a mean STAIC-State score greater than 1 SD above the normative mean; STAIC-T, those who met criteria for all three symptom clusters had a mean STAIC-Trait score more than 1 SD above the normative mean; CDI, children who met criteria for Psychic/Numbing Avoidance, for Hyperarousal, and for all three symptom clusters had mean CDI scores more than 1 SD above the normative mean. Thus, as evidenced on all measures, it appears that children who met all three symptom clusters also manifested clinically significant levels of anxiety and depression. Those who met criteria for Psychic Numbing/Avoidance showed clinically significant distress on two of the three measures (STAIC-S, CDI). Those who met criteria for Hyperarousal had clinically significant scores on one of the three measures (CDI). Those who met criteria for Reexperiencing did not show clinically significant levels of distress on any of the three measures.
These general findings were buttressed by an examination of self-reported anxiety and
depression levels in the children who met PTSD symptom criteria. Not only were the mean anxiety
and depression scores of children who met criteria significantly higher than the mean scores of those
children who did not meet criteria, but the symptomatic group's mean anxiety and depression levels
also exceeded the means of the relevant normative samples by more than one standard deviation.
Normative comparison such as that reported here has often been taken as an indicator of clinically
significant levels of psychopathology--in this case, clinically significant levels of anxiety and depression
appear to have been present in those children who met the full symptom picture. By contrast, the
anxiety and depression levels of the sample as a whole (symptomatic and non-symptomatic) did not
differ from the means of the relevant normative samples. This finding confirms earlier research that
has found that traumatized samples tend to show an increase in PTSD-related symptoms but, as a
whole, show nondeviant levels of anxiety and depression (Keppel-Benson & Ollendick, 1993). It
also suggests that clinically significant levels of distress may be experienced following a disaster by a
subsample of children who meet PTSD symptom-based criteria. This idea awaits further investigation.
With respect to the effects of age and gender, the only significant finding was that younger
children were more likely than older children to meet criteria for the hyperarousal symptom cluster.
This finding extends earlier findings that have found younger children tend to report higher scores on
PTSD-related measures following natural disasters (Belter & Shannon, 1993). However, other
findings regarding age and gender were not significant and this finding is particularly noteworthy as
regards the fact that equal numbers of younger versus older children and boys versus girls met
criteria for the full symptom picture. Thus, while younger children and girls have been found to report
a greater degree of distress in the aftermath of a disaster (Belter & Shannon, 1993; Vernberg et al., in
press), they did not here differ in terms of actually meeting the full symptom criteria for PTSD. These
findings await further exploration using structured diagnostic interviews.
Regarding the role of asthma, children with asthma were more likely to meet symptom criteria for the hyperarousal cluster than were children who did not have asthma. This finding met with expectations and confirms earlier research that has found that asthmatic children have a tendency to respond to stress-related situations with an increase in hyperarousal-like symptoms (Butz & Alexander, 1993; Miller & Wood, 1994). On the other, hand asthmatic and non-asthmatic children did not differ in the frequency of those who met criteria for the other two symptom clusters or, importantly, all three. Thus, while asthmatic children were found to have experienced a greater degree of eruption-related distress (Ronan, 1996a), they were not more likely than non-asthmatic children to meet the full symptom criteria for PTSD.
Limitations center largely around the reliance on the self-report methodology to estimate
PTSD-related symptom clusters (see also LaGreca et al., in press). Importantly, self-reports are not
substitutes for structured diagnostic interviews, and the findings here should not be interpreted as
confirming that children qualified for diagnoses. In order to make an actual diagnosis, a more
structured, comprehensive interview procedure is required. On the other hand, the purpose of the
study was not to establish the prevalence of diagnoses. The items on the Reaction Index (RI)
approximate relevant criteria used to determine a diagnosis of PTSD. Thus, this study used the RI in
a way that was deemed to be clinically useful and also allowed for direct comparison with a previous
study (LaGreca et al., in press). Given current concerns in the child-based literature on bridging the
gap between research purity and real-life pragmatics (Ronan, 1996b), findings here support the use
of the RI as an initial screening device for PTSD-related symptomatology. It appears to be an ideal
measure for screening large numbers of children that can then be followed up with a structured
diagnostic interview to establish actual diagnosis using a multiple gating methodology (Long, Ronan,
& Perreira-Laird, in press).
Future research is needed to look at the effects of school-based intervention approaches on PTSD-related symptoms. The question here is can interventions reduce the degree of distress (e.g., PTSD-related, anxiety, depression) while at the same time reducing the number of children who meet PTSD-related symptom criteria? Do different interventions have differing efficacy in this regard: are naturally occurring interventions (e.g., teacher-led counselling efforts) sufficient or is more extensive intervention necessary? Do different children respond differently to different interventions. The next generation of research looking at children and natural disasters needs to begin answering these questions.