Service Utilization and Event Reaction Patterns Among Children Who Received Project Liberty Counseling Services

By its very nature, terrorism is designed to create social chaos and a psychological climate of fear and intimidation ( 1 ). Even when terrorist activities are aimed at adults, the physical and mental well-being of children invariably is negatively affected ( 2 ). Children affected by terrorism generally have been found to manifest the same broad symptoms of traumatic stress as adults ( 2 ). Children also exhibit these symptoms of traumatic stress for six or more months after a traumatic event ( 3 , 4 ), and these symptoms vary according to the children's age and cognitive abilities ( 5 , 6 ). Accurate assessment of reactions to trauma among children as a prelude to rapid provision of therapeutic intervention ( 7 ) is particularly important.

The terrorist attacks on the World Trade Center on September 11, 2001, created widespread emotional distress among both adults and children. In a survey of 434 parents conducted two months after the attacks, parents reported that 18 percent of their children had severe or very severe stress reactions and 27 percent of their children exhibited problems in their daily functioning ( 8 ). A substantial number of New York City children continued to have elevated rates of psychiatric disorders for up to six months after the attacks. A needs assessment survey of 8,236 public schoolchildren estimated that 28.6 percent of fourth through 12th graders experienced at least one mental health problem serious enough to require intervention six months after September 11, 2001 ( 3 ). Children's symptom patterns also were found to vary as a function of age, gender, and race or ethnicity. Younger children, girls, and non-Caucasian children were more likely to report symptoms of anxiety-related disorders. Family exposure to the attacks was associated with a probable mental disorder to a larger degree than personal exposure ( 3 ).

Surveys found low rates of service utilization for children who experienced symptoms after the World Trade Center attacks. Two phone surveys of 434 and 112 parents, respectively, reported that 10 percent ( 8 ) and 22 percent ( 9 ) of symptomatic children received treatment; the school-based needs assessment ( 3 ) reported that only 33 percent of children who met criteria for posttraumatic stress disorder (PTSD) had sought counseling services. According to survey results, such services were more likely to have been accessed by children whose parents reported current PTSD ( 9 ), those who had lost a relative or friend in the attacks, those who had received previous counseling, and those who manifested severe PTSD ( 8 ). Survey information is very helpful in estimating population needs and service utilization after a disaster. However, it is also important to examine trends in service use by employing actual service utilization data to provide disaster mental health community planners, providers, researchers, and program managers with information to enhance opportunities for disaster preparedness and response.

Project Liberty, a federally funded program designed to respond to mental health needs after the terrorist attacks of September 11, 2001, provided large-scale, free, disaster-related outreach, individual and group counseling, psychoeducation, and referral services to both adults and children residing in the greater New York City metropolitan area ( 10 ). In order to be reimbursed for services, Project Liberty providers were required to complete service encounter logs, thus providing a source of data for examining demographic characteristics and event reaction profiles of clients served.

This study used data from Project Liberty to examine the characteristics of service utilization among children served, compare service utilization by children and adults who received individual or family counseling services, examine service utilization of children in relation to census data, and examine children's reactions to the World Trade Center disaster.

Methods

Under the Federal Emergency Management Agency (FEMA) funding conditions, all services provided under Project Liberty were anonymous. All of the 177 community-based Project Liberty service providers served both adults and children. Provider agencies included large and small mental health agencies, consumer-run organizations, faith-based social service agencies, and agencies with experience serving particular ethnic, cultural, or racial groups. Providers completed 753,015 service logs indicating date and location of service (for example, school or place of employment), borough or county where service was received, and risk category (for example, family of a missing or deceased person, persons who were injured in attack, or fire, police, or rescue personnel).

These service logs included both group education and individual (including family) counseling services provided from Project Liberty's inception through December 31, 2003. Individual counseling service logs included information about whether the contact was a first or follow-up visit, demographic information (client age, gender, preferred language, and race or ethnicity), and ratings of the presence of up to 31 behavioral, emotional, physical, and cognitive reactions to the events that the recipient identified as a reaction that he or she was experiencing or that the crisis counselor observed during that day's session. The 31 event reactions were adapted by Project Liberty from those included in the data management toolkit developed for the FEMA-funded disaster mental health crisis counseling program by the federal Center for Mental Health Services ( 11 ). Data from the separate Project Liberty program administered by the New York City Department of Education from January 2003 to December 2004 were not included in the analyses reported here; rather, this article focuses on the services delivered by community-based agencies, including services that community-based agencies provided in school settings.

The analyses focused on the logs of 681,318 service encounters (99 percent of the 687,848 individual crisis counseling sessions) for individual counseling sessions where counselors coded age. Visit logs were not grouped within each provider for these analyses. The Mount Sinai School of Medicine Institutional Review Board found this secondary analysis of anonymous data to be exempt from review.

The study first examined characteristics of service utilization by children, and chi square analyses were applied to examine whether event reactions for adults differed from those for children and whether younger and older children who received services differed in terms of demographic characteristics, event reactions, and possible major depressive disorder or PTSD ( 12 ). Logs were linked to census statistics by county (date of census information April 1, 2000), and analyses used chi square goodness-of-fit tests to examine whether the observed rates of children's service use and of children's race or ethnicity differed from what would be expected on the basis of census data. This analysis permitted determination of whether Project Liberty services were being accessed by children in representative proportion to their presence in the population.

Throughout, because of the very large number of service logs, group differences were nearly always significant, even when they differed by just a few percentage points. Hence, this report notes differences that were both statistically significant (p<.05) and programmatically meaningful (defined as having a difference between groups of 10 percent or greater, a difference large enough to suggest changes in service delivery to target underserved populations).

To examine whether providers influenced the likelihood of receiving follow-up visits, for each provider, the proportion of first visits versus follow-up visits and the proportion of first visits provided to children compared with adults were calculated. These proportions were then correlated. Chi square statistics were then applied to determine whether follow-up visits were more likely in school settings. After finding that significantly more follow-up visits occurred in school settings, we used hierarchical logistic regressions to determine whether children were more likely to receive a follow-up visit after the analyses controlled for location of service (school versus other). The first logistic regression model included one independent measure, location of service. The second logistic regression model included two independent measures, location of service and whether the participant was a child or an adult.

Results

Of 681,318 individual service logs for first and follow-up visits, 105,244 (15 percent) were for services provided to children either individually or in family counseling. Because services were provided anonymously, it was not possible to determine how many different children had follow-up visits and, among those with follow-up visits, how many each child had. When only first visits were examined, 39,962 of 463,277 service logs (9 percent) were for services provided to children. As shown in Table 1 , children received first-time visits most frequently from community providers in schools (41 percent), and most first visits with children (69 percent) were provided to those aged 12 to 17 years.

Enlarge table

Chi square goodness-of-fit tests indicated that, across all areas served by Project Liberty, children constituted a smaller proportion of individual counseling service logs than their representation in the population. Compared with adults, significantly fewer first-visit service logs were for children (regardless of whether children six years of age or younger were included or excluded), given the census data (most p values <.001 for chi square goodness-of-fit comparisons for children versus adults). However, among those served, the proportions of gender and race or ethnicity were generally consistent with what would have been expected on the basis of the census data.

Comparisons of preschool, grade school, and older children who received individual counseling services showed no between-group differences for gender, for whether services were received in English or another language, or for whether the child received a referral for professional mental health services. However, preschool children served were more likely to be Caucasian (31 percent compared with 20 percent of grade school children and 17 percent of older children) and less likely to be African American (22 percent compared with 33 percent of grade school children and 35 percent of older children) ( χ² =244, df=6, p<.001).

As shown in Table 2 , on the basis of first-visit service logs, children and adults who received individual counseling services did not differ meaningfully from each other in terms of gender, preferred language, or likelihood of referral for professional mental health services. However, compared with the adults, significantly more of the children served were Hispanic and significantly fewer were Caucasian.

Enlarge table

Compared with the proportion of adults (90 percent) among the service logs for first visits, a significantly smaller proportion of adults (70 percent) was seen among the service logs for follow-up visits ( χ² =51,570, df=1, p<.001). Providers who devoted a higher proportion of first visits to children were also more likely to offer proportionately more follow-up visits overall (r=-.21, p<.005). Significantly more follow-up visits with both children and adults (60,341 visits, or 28 percent) than first visits with both children and adults (36,862 visits, 8 percent) occurred in school settings ( χ² =47,131, df=1, p<.001). Proportionally more of these follow-up visits in school settings were provided to children, than to adults (49,710 follow-up visits for children, or 76 percent, compared with 10,529 follow-up visits for adults, or 7 percent; χ² =109,494, df=1, p<.001).

Results of logistic regression suggest that adults were .36 times less likely than children to receive a follow-up visit ( χ² =14,511, df=1, p<.001), even after the analyses controlled for whether the follow-up visit was provided in a school setting (those who received services in school settings were 2.5 times more likely to receive a follow-up visit; χ² =10,277, df=1, p<.001).

Table 3 summarizes the percentage of adults and children within each age group whose counselors reported specific event reactions, categorized as behavioral, emotional, physical, and cognitive. Children were less likely than adults to report at least one physical event reaction (48 percent compared with 61 percent; χ² =2,700, df=1, p<.001). No significant differences were found between children and adults for the emotional, behavioral, or cognitive reaction categories. As illustrated in Table 3 , Project Liberty providers reported that the most common reactions among children (20 percent or greater) were similar to the most common reactions among adults. Additionally, the percentage of individuals with possible major depressive disorder or possible PTSD increased with age ( χ² =346, df=5, p<.001 for major depressive disorder and χ² =868, df=5, p<.001 for PTSD).

Enlarge table

Discussion

Children represented 9 percent of those who received individual counseling services from community-based agencies participating in Project Liberty between program inception and December 2003, when service delivery by these agencies ceased. Project Liberty also funded additional counseling and other services to New York City schoolchildren between January 2003 and December 2004 that were organized through the New York City Department of Education. Although data processing delays precluded the inclusion of detailed data from the Department of Education's service logs in these analyses, we know that an additional 36,337 children received individual crisis counseling during this later phase of Project Liberty. With these additional services included, 20 percent of Project Liberty's total service logs for individual (including family) counseling were provided to children, which is comparable to the census information identifying 25 percent of New York City's population as children. Community-based Project Liberty providers reached children of both genders and varying races or ethnicities in expected proportions, given the census statistics.

The relative frequencies of traumatic stress reactions recorded for these children paralleled what Project Liberty counselors recorded for the adults they served. Sadness, tearfulness, fear, anxiety, concentration difficulties, anger, irritability, intrusive thoughts and images, and difficulty sleeping were the most common event reactions reported for children as well as adults in this and other samples ( 2 , 13 ).

There were, however, some important differences in event reactions between younger and older children. Compared with reports about older children (12 to 17 years), counselors reported that elementary school-age children exhibited relatively more isolation and withdrawal, anxious and fearful reactions, and concentration difficulties. Older children were more similar to adults and more likely than younger children to exhibit several avoidance and "numbing" reactions across domains and to abuse substances. Additionally, the incidence of possible major depressive disorder and PTSD appeared to increase with age. These outcomes are consistent with literature indicating that younger children are less likely to manifest psychic numbing ( 5 ) or avoidance symptoms associated with PTSD ( 6 ).

Given the growing body of evidence linking childhood trauma to a large number of later-life psychological ( 13 , 14 , 15 ) and physical ( 16 , 17 , 18 ) disorders, accurate assessment of vulnerability to traumatic stress is critical. Cognitive-behavioral interventions are proving effective in ameliorating the effects of traumatic stress on children ( 19 , 20 , 21 , 22 , 23 , 24 ). These interventions typically meld four basic components: psychoeducation about trauma, sustained exposure to trauma related cues or memories, coping skills to manage fearfulness, and parent training ( 7 ). Little is known, however, about how to best match varying doses of these components to different age categories ( 7 , 22 ). Although accurate assessment is central to identifying children in need of early intervention ( 25 ), it is also important in helping to develop specific age-relevant interventions.

This investigation has several limitations. Race and ethnicity were coded as a single variable on the basis of observation by Project Liberty service providers, rather than from the service recipient's self-assessment. Also, because services were provided anonymously, it was not possible to determine how many different children had follow-up visits or what the average number of follow-up visits was for any given child.

Conclusions

Information from Project Liberty service logs indicates that by January 2003 community-based agencies succeeded in providing nearly 40,000 children with individual counseling both within the schools and in other community settings. Compared with adults, children were more likely to participate in more than one counseling session. A majority of follow-up sessions were provided in the schools. However, after the analyses controlled for setting, children were still more likely than adults to receive follow-up services. Event reaction patterns differed as a function of children's age. Elementary school-age children manifested relatively more isolation and withdrawal, anxious and fearful reactions, and concentration difficulties than older children. Knowledge of the reaction patterns shown by younger and older children may be useful in refining treatments to help reduce the psychological impact of children's trauma after terrorist incidents and other disasters.

Acknowledgments

This evaluation was funded by grant FEMA-1391-DR-NY (titled "Project Liberty: Crisis Counseling Assistance and Training Program") to New York State from the Federal Emergency Management Agency. The Center for Mental Health Services of the Substance Abuse and Mental Health Services Administration administered the grant. The authors express their appreciation to Katherine M. Shear, M.D.