sources. Sixty unduplicated references were generated by the
search. Abstracts were reviewed to identify publications
describing scientific studies of child interventions for
disasters and terrorist incidents, but interventions addressing
threat of terrorism, political conflict, war, and various singleincident
traumas such as accidents were also included.
Throughout the search process, the reference sections of
publications, review articles, and studies known to the
authors that were not generated in the search were used to
locate additional materials. Studies using experimental
control conditions were selected for analysis. In all, 28
research papers, which assessed PTSD or PTSD symptoms
or reactions as outcomes, were judged appropriate for
inclusion in this review. One article described two separate
investigations using the same intervention delivered in
individual and group format [14]; these studies were
analyzed separately. Thus, the final database of independent
samples numbered 29.
2.1. Coding procedures
Information and categories to be coded and analyzed were
determined after a preliminary review of the available
studies. Every study was coded to denote the treatment
approach and the quality of the description of the
intervention; characteristics of the sample, methodology,
and instruments; assessor training; treatment adherence; and
efficacy of the intervention. After piloting the procedures on
publications from the adult trauma and child maltreatment
literatures, two investigators coded all studies with respect to
various categories. Toward ensuring adequate inter-rater
agreement, piloting continued until coders reached an
agreement of 80% on the coding variables for five practice
articles. The same two coders then coded each of the studies.
Cohen's kappa coefficients for the final sample ranged from
.29 to .91 (median = .71) and percent agreement ranged
from 52% to 94% (median = 85%). Of 19 coding variables,
7 yielded kappa at or below .70 and 12 yielded kappa above
.70. There were no statistically significant differences in
kappa for the two raters on the first 15 articles and the final
sample. The research team met to discuss discrepancies in
coding and to reach a reasoned consensus.
The research team sorted interventions into treatment
types based on the techniques used in each study. The
treatment types used were as follows: strict Cognitive
Behavioral Therapy (CBT), eclectic with no cognitive
behavioral components, eclectic with some cognitive
behavioral components, Eye Movement Desensitization
and Reprocessing (EMDR), exposure, massage, relaxation,
and psychological debriefing/crisis intervention. Other novel
approaches were coded as described in the intervention
studies as spiritual hypnosis [15] and as Huggy Puppy [14].
The analysis of each intervention study identified the
treatment approach, the use of the diagnosis of PTSD and/or
PTSD symptoms or reactions as target symptoms or
outcome, and the use of inclusion and exclusion criteria;