Reactions of Oklahoma City bombing survivors to media coverage of the September 11, 2001, attacks

Available online at www.sciencedirect.com

ScienceDirect Comprehensive Psychiatry 65 (2016) 70 – 78 www.elsevier.com/locate/comppsych

Reactions of Oklahoma City bombing survivors to media coverage of the September 11, 2001, attacks Betty Pfefferbaum a,⁎, Pascal Nitiéma a , Rose L. Pfefferbaum b, c , J. Brian Houston b , Phebe Tucker a , Haekyung Jeon-Slaughter d , Carol S. North d, e, f a

Department of Psychiatry and Behavioral Sciences, College of Medicine, University of Oklahoma Health Sciences Center, Oklahoma City, OK, USA b Department of Communication, University of Missouri, Columbia, MO, USA c Phoenix Community College, Phoenix, AZ, USA d Department of Psychiatry, University of Texas Southwestern Medical Center, Dallas, TX, USA e Department of Surgery, University of Texas Southwestern Medical Center, Dallas, TX, USA f Metrocare Services, Dallas, TX USA

Abstract Objective: This study explored the effects of media coverage of a terrorist incident in individuals remote from the location of a major attack who had directly experienced a prior terrorist incident. Method: Directly-exposed survivors of the 1995 Oklahoma City bombing, initially studied six months after the incident, and indirectlyaffected Oklahoma City community residents were assessed two to seven months after the September 11, 2001, attacks. Survivors were assessed for a diagnosis of bombing-related posttraumatic stress disorder (PTSD) at index and follow up, and emotional reactions and September 11 media behavior were assessed in all participants. Results: Among the three investigated forms of media (television, radio, and newspaper), only television viewing was associated with 9/11-related posttraumatic stress reactions. Exposure to the Oklahoma City bombing was associated with greater arousal in relation to the September 11 attacks, and among survivors, having developed bombing-related PTSD was associated with higher scores on all three September 11 posttraumatic stress response clusters (intrusion, avoidance, and arousal). Although time spent watching television coverage of the September 11 attacks and fearrelated discontinuation of media contact were not associated with Oklahoma City bombing exposure, discontinuing September 11 media contact due to fear was associated with avoidance/numbing in the full sample and in the analysis restricted to the bombing survivors. Conclusion: Surviving a prior terrorist incident and developing PTSD in relation to that incident may predispose individuals to adverse reactions to media coverage of a future terrorist attack. © 2015 Elsevier Inc. All rights reserved.

1. Introduction Media coverage brings events such as the September 11 attacks into the homes of millions and can affect individuals indirectly and remotely [1]. Even though contact with media coverage does not constitute trauma exposure for the diagnosis of posttraumatic stress disorder (PTSD) and is specifically excluded as a form of exposure in the diagnostic criteria in the fifth edition of the Diagnostic and Statistical Manual [2], television coverage of the September 11 attacks ⁎ Corresponding author at: Department of Psychiatry and Behavioral Sciences, College of Medicine, University of Oklahoma Health Sciences Center, P.O. Box 26901, WP-3217, Oklahoma City, Oklahoma, USA, 73126-0901. E-mail address: [email protected] (B. Pfefferbaum). http://dx.doi.org/10.1016/j.comppsych.2015.09.010 0010-440X/© 2015 Elsevier Inc. All rights reserved.

raised concern as a potential source of emotional distress in New York City residents [3–6] and across the country [7–9]. Neria and Sullivan [10] have suggested that pre-existing vulnerability (e.g., prior trauma, psychiatric history) is important in the development of posttraumatic responses in relation to media reports of disasters. In a longitudinal study of children exposed first to Hurricane Katrina and then to Hurricane Gustav, Weems and colleagues [11] found that the deleterious consequences of viewing extensive disaster media coverage were greatest for youth with high levels of pre-existing PTSD symptoms and that in youth with increased pre-existing PTSD symptoms, those who viewed more disaster media coverage were more likely to remain symptomatic one month post event. Ahern and colleagues [5] found that viewing television coverage of the attacks was

B. Pfefferbaum et al. / Comprehensive Psychiatry 65 (2016) 70–78

related to “probable” PTSD in New York City residents who were directly affected by the attacks and those who had a history of prior trauma. In a national study of the September 11 attacks, Silver and colleagues [9] found that in addition to exposure, participants' emotional and physical status pre September 11 explained posttraumatic stress symptoms and global distress over time. To date, however, little is known about the effects of media coverage of terrorist events in individuals remote from the location of a major attack who have directly experienced a prior terrorist incident, an issue of growing concern as the threat of terrorism increases. The current study—fielded two months after the September 11 attacks—took advantage of a seven-year follow-up investigation of Oklahoma City bombing survivors to compare their reactions to media coverage of the September 11 attacks with those of Oklahoma City residents who were not directly exposed to the bomb blast. The study also provided the opportunity to examine differences in reactions to media coverage among survivors who did and did not develop bombing-related PTSD. 2. Methods 2.1. Procedures and participants The University of Oklahoma Health Sciences Center and Washington University School of Medicine institutional review boards approved this seven-year follow-up study. Potential participants for the follow-up study were contacted by letter and/or telephone. All participants provided written informed consent and were paid $75 to compensate for their time and effort. Study participants included two groups distinguished by exposure to the 1995 Oklahoma City bombing, directly-exposed survivors (survivors) and indirectly-affected community participants (community residents). Of 182 survivors first evaluated in the index study [12], which was conducted approximately six months after the bombing, 99 completed both the diagnostic interview and self-report questionnaires for the follow-up investigation approximately seven years after the bombing and two to seven months after the September 11, 2001, attacks [13]. Among bombing survivors assessed at baseline, 83 (45.6%) did not participate in this study of September 11 media behavior and posttraumatic stress reactions, including 69 survivors lost to follow up and 14 who participated in the seven-year follow up but did not complete the September 11-related survey. The 83 survivors who did not participate in the September 11 component of the study were not statistically different from the 99 survivors who took part in the September 11 survey in gender (female gender: 53.0% versus 50.5%; χ 2 = 0.11; df = 1, p = 0.7361), mean age at baseline (43.2 ± 10.6 years versus 42.6 ± 12.5 years, t = 0.34, df = 180, p = 0.7364), race/ethnicity (proportion of Caucasians: 85.5% versus 91.9%, χ 2 = 1.88; df = 1, p = 0.1707), or proportion of individuals diagnosed with bombing-related PTSD at baseline (37.5% versus 31.1%, χ 2 = 0.73; df = 1, p = 0.3921).

71

Sixty one unexposed Oklahoma City community residents constituted the community comparison group that was recruited and assessed as part of the seven-year follow-up study. These participants resided in the Oklahoma City metropolitan area at the time of the bombing, were not in the immediate vicinity of the blast, did not have friends or relatives killed in the bombing, and were not rescue or recovery workers in the disaster. Recruitment of the community group used two approaches. First, participating survivors were asked to identify a peer from the community who met the community group inclusion criteria. Additional recruitment was through word of mouth and advertisement. The study sample included 99 bombing survivors and 61 unexposed community participants. 2.2. Measures Participants were administered the Diagnostic Interview Schedule for DSM-IV (DIS-IV) [14] to obtain demographic information and assess full diagnostic criteria for a post-disaster (i.e., at any time after the bombing) diagnosis of bombingrelated PTSD. Bombing survivors were assessed at both index and follow up, but the community participants were assessed only in the follow-up study. All participants in the seven-year follow-up study also completed the self-administered Disaster Supplement Questionnaire [15] assessing emotional reactions and media behavior related to the September 11 attacks. For the diagnosis of bombing-related PTSD, Diagnostic and Statistical Manual-IV-Text Revision [16] PTSD criteria A through F were required in relation to the Oklahoma City bombing. Criterion A required direct exposure to the bombing, which all bombing survivors met. Survivors who reported one or more of five bombing-related criterion B (intrusion) symptoms, three or more of seven criterion C (avoidance and numbing) symptoms, and two or more of five criterion D (hyperarousal) symptoms related to the bombing, and who also reported that the symptoms lasted more than one month (criterion E) and created clinically significant emotional distress or interfered with functioning (criterion F), met bombingrelated PTSD criteria. By definition, the community group participants did not have any PTSD-qualifying exposures to the bombing; thus, none met criteria for bombing-related PTSD. The 99 Oklahoma City bombing survivors were grouped according to their bombing-related PTSD diagnosis status. Survivors who met DSM-IV criteria for bombing-related PTSD at any time after the disaster and were still symptomatic at the seven-year follow up were grouped under the label of “Current PTSD.” Survivors who met bombing-related PTSD criteria at some time after the disaster but had no current PTSD symptoms at the follow up were grouped under the label of “Remitted PTSD.” Finally, survivors who did not develop bombing-related PTSD were grouped under the label of “Never PTSD,” and those who did develop bombing-related PTSD were grouped under the label of “Ever PTSD.” A modified version of the Impact Event Scale-Revised (IES-R) [17,18] used in studies of the Oklahoma City bombing

72

B. Pfefferbaum et al. / Comprehensive Psychiatry 65 (2016) 70–78

[19–22] was added to the Disaster Supplement Questionnaire for the seven-year follow-up study to obtain posttraumatic stress response data specific to the September 11 attacks (9/11 IES-R). The 9/11 IES-R consisted of 22 items from the version of the IES-R described by Weiss and Marmar [18] with a modified scale of 1 for “not at all” to 4 for “often.” The possible range of 9/11 IES-R scores was between 22 and 88. The 9/11 IES-R total score and its three subscales—intrusion (8 items), avoidance/numbing (8 items), and hyperarousal (6 items) subscales—were analyzed separately to measure the participants' posttraumatic stress responses to the September 11 attacks. The reliability statistics (Cronbach's alpha) of the three subscales were very good in the study sample (arousal: α = 0.90; avoidance/numbing: α = 0.92; intrusion: α = 0.94). Participants were asked to recall the number of hours per day they watched television, listened to radio, or read newspaper coverage (one question for each media form) of the September 11 attacks during the first week after the attacks. They also were asked with a single question if they quit watching, listening to, or reading September 11 media coverage because of fear. 2.3. Data analysis Proportions of variables of interest in bombing survivors and controls were compared using Bayesian estimation with vague priors Beta(1, 1). Bayesian simple linear regression was used to compare the means of continuous variables when two groups were evaluated. Bayesian multivariable linear models were fitted to assess the relationship between 9/11 IES-R scores and selected variables. Diffuse priors were used for the model parameters, that is, a normal distribution N(0, 10E6) for the model coefficients, and an inverse gamma distribution (shape = 3/10; scale = 10/3) for the variances. Posterior distributions of the parameters estimates were generated using the Metropolis–Hastings algorithm (number of iterations = 100,000). Simulations were thinned at a rate of 1/2 to minimize autocorrelation among the generated values, leaving 50,000 values to perform statistical inferences about the model parameters. These simulated posterior distributions were summarized with the distribution mean and its 95% Bayesian credible interval (BCI). Separate models were built for the 9/11 IES-R total score and for each of the subscales (intrusion, avoidance/numbing, and arousal). Two Bayesian logistic regression models (one for the full sample with a covariate for the bombing exposure group, and one for the Oklahoma City bombing survivors with a covariate for PTSD) with diffuse priors were also used to identify the factors statistically associated with discontinuation of September 11 media contact. Collinearity among predictors in the logistic regression models was assessed with the weighted values of these predictors, and it was found not to be of concern, with the computed variance inflated factors (VIF) ranging from 1.1 to 7.5. SAS 9.3 (SAS Institute, Cary, NC) was used for most of the statistical analyses.

3. Results The results are presented in three sections. First is a description of the analysis of the full sample followed in the second section by a description of the analysis of bombing survivors. The third section presents an analysis examining the discontinuation of 9/11-related media contact due to fear. 3.1. Full sample The analysis of the full sample included an examination of the descriptive data, a comparison of survivor and community groups, an exploration of the effect of timing of the assessment, and the use of linear regression models to assess the relationship between 9/11 IES-R scores and hours spent in contact with each form of September 11 media coverage adjusting for demographics and other variables of interest. 3.1.1. Demographics A total of 99 bombing survivors and 61 comparison participants were included in this study. Fifty (50.5%) of the 99 survivors and 33 (54.1%) of the 61 participants in the comparison group were female, with no statistical gender difference (95% BCI for proportion difference = −19.0%; 12.0%). The great majority of the study sample was Caucasian (n = 85, 85.9% for bombing survivors; n = 52, 85.2% for the comparison group) with no statistical difference between the two exposure groups (95% BCI for proportion difference = −9.9%; 13.0%). The age of the participants ranged from 26 to 80 years for the survivors and from 25 to 67 years for the comparison group. The mean age in the survivor group (49.6 years; SD = 10.6) was statistically higher than that of the comparison group (mean = 46.2 years; SD = 9.3), with a mean difference of 3.4 years (95% BCI = 0.1; 6.6). 3.1.2. September 11-related posttraumatic stress response scores The mean 9/11 IES-R total and all three subscale (intrusion, avoidance/numbing, arousal) scores were each statistically higher in bombing survivors than in the comparison group. See Table 1. 3.1.3. September 11 media behavior Table 1 presents the hours per day of contact with September 11 media coverage for each of the media forms (television, radio, newspaper) during the first week after the disaster. In both survivors and comparison participants, television was the media form most used by respondents, followed by radio and then newspaper. Twenty-eight (17.5%) participants reported that they had discontinued contact with September 11 media coverage because of fear. There were no statistical differences between the bombing survivors and comparison participants in average time spent watching television, listening to radio, or reading newspaper coverage of the September 11 attacks during the first week after the attacks. There was no statistical difference between survivors (n = 19, 19.2%) and community participants (n =

B. Pfefferbaum et al. / Comprehensive Psychiatry 65 (2016) 70–78

73

Table 1 September 11 IES-R scores, media contact, and effect of time interval before assessment on measured variables.

9/11 IES-R score Intrusion Avoidance/numbing Arousal Total Hours per day Television Radio Newspaper

Full sample (n = 160), mean (SD)

Bombing survivor group (n = 99), mean (SD)

Community comparison group (n = 61), mean (SD)

15.6 (6.7) 13.8 (6.4) 10.1 (4.6) 39.2 (16.4)

16.9 (7.1) 14.7 (6.5) 11.1 (5.2) 42.0 (17.3)

13.7 (5.6) 12.4 (6.1) 8.6 (3.1) 34.7 (13.6)

4.3 (3.2) 1.9 (2.2) 1.1 (1.0)

4.0 (2.9) 1.8 (2.3) 1.1 (0.8)

4.6 (3.5) 2.2 (2.2) 1.1 (1.3)

Survivor comparison groups: mean difference (95% BCI)

3.1 2.4 2.6 7.4

(1.0; 5.3)⁎ (0.2; 4.5)⁎ (1.0; 4.0)⁎ (2.1; 12.9)⁎

−0.6 (−1.6; 0.5) −0.4 (−1.2; 0.4) 0.0 (−0.4; 0.4)

Full sample (n = 160), coefficient estimate of simple linear model of length of time on variable of interest (95% BCI)a −0.018 (−0.029; −0.008)⁎ −0.011 (−0.020; −0.001)⁎ −0.010 (−0.017;-0.003)⁎ −0.035 (−0.061; −0.009)⁎ 0.003 (−0.002; 0.008) 0.003 (−0.001; 0.006) −0.001 (−0.002; 0.001)

⁎ Statistically different from 0. a Effect of length of time interval (in days) between September 11, 2001, and date participants (n = 160) were assessed; the precision of the numbers was extended to three decimals given their relatively small values.

9, 14.8%) in the proportions reporting fear-related discontinuation of media contact (95% BCI for proportion difference = −7.9%; 16.0%). 3.1.4. Effect of timing of the assessment The time interval between September 11, 2001, and the study assessments ranged from 60 to 415 days (mean = 192.3; SD = 93.1) for the bombing survivors, and from 64 to 499 days (mean = 291.3; SD = 88.1) for the community participants. The average time interval was statistically longer in the community group than the bombing survivors, with a mean difference of 99.0 days (95% BCI = 69.4; 128.5). In the full sample, the amounts of time in contact with television, radio, and newspaper coverage of the September 11 attacks were not statistically associated with the length of time between September 11, 2001, and the assessments. See Table 1. Hence, the amount of time participants reported that they had spent in contact with each form of September 11 media coverage in the first week after the attacks was independent of the time elapsed between the attacks and the assessment. The 9/11 IES-R total and all three subscale scores were each associated with a shorter duration of time between the September 11 attacks and the assessments (i.e., those evaluated earlier had statistically higher scores). See Table 1.

3.1.5. Multivariable models of September 11-related posttraumatic symptoms scores on selected covariates Bayesian linear regression models (one model for the total IES-R score and one for each of the three IES-R subscales) were fitted to assess the relationship between 9/11 IES-R scores and hours spent in contact with each form (television, radio, newspaper) of media coverage of the September 11 attacks, adjusting for gender, age, time elapsed before the assessment, Oklahoma City bombing exposure group, and whether the respondent discontinued contact with media coverage of the attacks due to fear. With all media forms included in each of the models, television viewing was statistically associated with higher 9/11 IES-R total (β = 0.97; 95% BCI = 0.06; 1.89) and intrusion subscale (β = 0.47; 95% BCI = 0.11; 0.81) scores. Listening to radio and reading newspaper coverage of the attacks were not associated with 9/11 IES-R total or any subscale scores. Based on the values of the β estimates in the models, fear-based discontinuation of media contact was statistically associated with 9/11 IES-R total and all three subscale scores. The 9/11 IES-R arousal score was higher in Oklahoma City bombing survivors than in the community comparison participants. Females scored statistically higher than males on the 9/11 IES-R total and the intrusion and

Table 2 Multivariable models examining the relationship between September 11 media contact and 9/11 IES-R scores (β estimates with 95% BCI). Independent variables

Bombing survivor (reference = community comparison group) Watching television (hours/day) Listening to radio (hours/day) Reading newspaper (hours/day) Discontinued contact with media coverage (reference = no) Gender (reference = male) Age (years) Time (days) between 9/11/01 and assessment ⁎ Statistically different from 0.

Dependent variable (one model for each) 9/11 IES-R Total

9/11 IES-R Intrusion

9/11 IES-R Avoidance/Numbing

9/11 IES-R Arousal

5.93 (−0.74; 12.69) 0.97 (0.06; 1.89)⁎ 0.71 (−0.74; 2.11) −1.55 (−4.66; 1.53) 10.96 (3.75; 18.84)⁎ 6.61 (0.13; 12.76)⁎ 0.10 (−0.21; 0.40) −0.03 (−0.07; 0.00)

1.66 (−0.90; 4.43) 0.47 (0.11; 0.81)⁎ 0.21 (−0.34; 0.79) −0.66 (−1.83; 0.51) 3.68 (0.75; 6.63)⁎ 2.54 (0.27; 5.01)⁎ 0.06 (−0.05; 0.18) −0.02 (−0.03; 0.00)

2.19 (−0.68; 4.90) 0.20 (−0.16; 0.58) 0.48 (−0.09; 1.05) −0.64 (−1.79; 0.56) 5.12 (2.21; 8.07)⁎ 2.36 (−0.18; 4.69) 0.03 (−0.09; 0.15) −0.01 (−0.02; 0.00)

2.14 (0.28; 3.99)⁎ 0.22 (−0.03; 0.46) 0.15 (−0.22; 0.56) −0.45 (−1.29; 0.38) 2.34 (0.32; 4.36)⁎ 1.78 (0.14; 3.51)⁎ 0.02 (−0.06; 0.10) −0.01 (−0.02; 0.00)

74

B. Pfefferbaum et al. / Comprehensive Psychiatry 65 (2016) 70–78

arousal subscale scores. See Table 2. Participant age was not statistically associated with the 9/11 IES-R total or subscale scores. Finally, although bivariate analyses showed that 9/11 IES-R scores decreased as the time interval between the September 11 attacks and the assessment date of the participants increased, this was not the case in the multivariable models. See Table 2. 3.2. Oklahoma City bombing survivors The analysis of Oklahoma City bombing survivors compared those who did and did not develop PTSD over the course of the follow-up period. 3.2.1. Oklahoma City bombing-related PTSD Forty (40.4%) of the 99 bombing survivors developed bombing-related PTSD at some time during the seven-year follow-up period (“Ever PTSD”) while 59 (59.6%) survivors did not develop bombing-related PTSD as assessed either at baseline or at follow up (“Never PTSD”). Nineteen (19.2%) survivors with bombing-related PTSD had bombing-related PTSD at the seven-year follow up (“Current PTSD”). Another 21 (21.2%) survivors previously met criteria for bombing-related PTSD but no longer met PTSD diagnostic criteria at follow up (“Remitted PTSD”). 3.2.2. Survivors with versus those without bombing-related PTSD In bivariate analyses, 9/11 IES-R total and all three subscale mean scores were higher in survivors who developed bombing-related PTSD compared to survivors who did not. Additionally, the 9/11 IES total and intrusion and arousal subscales scores, but not the avoidance/numbing subscale score, were statistically higher in survivors with current bombing-related PTSD compared to those with remitted bombing-related PTSD. See Table 3. Multivariable analyses, with one model for each 9/11-IES-R subscale and the total score, adjusting for demographics (age and gender), September 11 media behavior (September 11 television, radio, and newspaper contact and fear-related discontinuation of media contact), bombing-related PTSD, and time elapsed before assessment, confirmed the statistical associations observed in

the bivariate analyses. In these multivariable analyses, bombing-related PTSD was still statistically and positively associated with 9/11 IES-R intrusion (β = 5.54, 95% BCI = 2.40; 8.67), avoidance/numbing (β = 7.23, 95% BCI = 4.62; 9.87), arousal (β = 4.21, 95% BCI = 1.75; 6.61), and total (β = 17.27; 95% BCI = 9.03; 27.51) scores. Hours of contact with media coverage of the September 11 attacks in the first week after the event did not differ statistically between Oklahoma City bombing survivors who did and did not develop bombing-related PTSD. See Table 3. The proportion of respondents who discontinued September 11 media contact because of fear was higher in survivors who developed bombing-related PTSD (32.5%) compared to survivors who did not develop bombing-related PTSD (10.2%) (95% BCI for proportion difference = 5.3%; 38.0%). The proportion of survivors with current PTSD who discontinued September 11 media contact because of fear (31.1%) was not different from that of survivors with remitted PTSD (33.3%) (95% BCI for proportion difference = −29.0%; 26.0%). 3.3. Multivariable models of discontinuing September 11 media contact In a model with the full sample, fear-related discontinuation of September 11 media contact was statistically associated with the 9/11 IES-R avoidance/numbing subscale score after controlling for gender, age, Oklahoma City bombing exposure group, and 9/11 IES-R intrusion and arousal subscale scores. September 11 IES-R intrusion and arousal subscale scores were not statistically associated with fear-related discontinuation of media contact independent of the other variables in the model. In this model, Oklahoma City bombing exposure was not statistically associated with fear-related discontinuation of September 11 media contact. None of the other covariates included in the model was statistically associated with fear-related discontinuation of September 11 media contact. See Table 4. Similarly, in the model including only the Oklahoma City bombing survivors, fear-related discontinuation of September 11 media contact was statistically associated with the 9/11

Table 3 September 11 IES-R scores and contact with September 11 media coverage in bombing survivors according to their bombing-related PTSD status. Never PTSD (n = 59) 9/11-related IES-R scores Intrusion 13.8 (5.9) Avoidance/Numbing 11.7 (4.7) Arousal 8.9 (3.9) Total score 34.1 (13.1) Media time Television 3.8 (2.7) Radio 1.6 (2.3) Newspaper 1.2 (1.0) Total 6.0 (3.0) ⁎ Statistically different from 0.

Ever PTSD (including Remitted Remitted and Current PTSD PTSD) (n = 40) (n = 21)

Current PTSD (n = 19)

21.4 (6.5) 18.9 (6.5) 14.4 (5.1) 54.1 (16.0) 4.5 (3.3) 2.2 (2.2) 1.1 (0.5) 6.9 (3.7)

23.9 (6.0) 7.6 (5.1; 10.1)⁎ 20.9 (5.0) 7.2 (4.9; 9.4)⁎ 16.2 (5.4) 5.4 (3.5; 7.3)⁎ 60.9 (14.2) 20.1 (13.6; 26.3)⁎ 4.4 (3.0) 0.7 (−0.5; 2.1) 2.0 (2.1) 0.6 (−0.4; 1.6) 1.0 (0.6) −0.1 (−0.5; 0.2) 6.8 (3.1) 0.9 (−0.6; 2.5)

19.3 (6.2) 17.2 (7.1) 13.0 (4.6) 49.1 (15.7) 4.6 (3.7) 2.4 (2.4) 1.1 (0.5) 7.1 (4.4)

Ever PTSD v Never Current PTSD v Remitted PTSD difference PTSD difference (95% BCI) (95% BCI) 4.7 (0.7; 8.9)⁎ 3.8 (−0.2; 7.9) 3.2 (0.1; 6.7)⁎ 11.9 (1.7; 22.6)⁎ −0.2 (−2.5; 2.2) −0.4 (−2.1; 1.4) −0.1 (−0.6; 0.4) −0.3 (−3.2; 2.7)

B. Pfefferbaum et al. / Comprehensive Psychiatry 65 (2016) 70–78

75

Table 4 Multivariable models examining the relationship between discontinuation of September 11 media contact and 9/11 IES-R scores (β estimates with 95% BCI). Dependent variable: fear-related discontinuation of September 11 media contact Independent variables Gender (reference = male) Age (years) 9/11 IES-R Intrusion 9/11 IES-R Avoidance/Numbing 9/11 IES-R Arousal Exposure to the Oklahoma City bombing (reference = community comparison group) Oklahoma City bombing-related PTSD (current or remitted; reference = no PTSD) Time (days) between 9/11/01 and assessment

Full sample (n = 160)

Bombing survivors (n = 99)

0.85 (−0.22; 1.93) 0.01 (−0.04; 0.06) −0.16 (−0.36; 0.03) 0.19 (0.07; 0.32)⁎ 0.07 (−0.17; 0.31) −0.07 (−1.23; 1.12) – 0.00 (−0.01; 0.01)

−0.14 (−1.65; 1.35) −0.06 (−0.13; 0.01) −0.07 (−0.31; 0.16) 0.19 (0.03; 0.36)⁎ −0.01 (−0.32; 0.30) – 0.59 (−1.01; 2.17) 0.00 (−0.01; 0.01)

In the full sample, exposure to Oklahoma City bombing (survivor versus controls) was included as an independent variable, while in the survivor sample, PTSD status (Ever PTSD versus Never PTSD) was used as independent variable. ⁎ Statistically different from 0.

IES-R avoidance/numbing subscale score after controlling for gender, age, bombing-related PTSD, and 9/11 IES-R intrusion and arousal subscale scores. September 11 IES-R intrusion and arousal subscale scores were not statistically associated with fear-related discontinuation of media contact independent of the other variables in the model. In this model, bombingrelated PTSD was not statistically associated with fear-related discontinuation of September 11 media contact. See Table 4. 4. Discussion This seven-year follow-up study of directly-exposed Oklahoma City bombing survivors, fielded in the fall of 2001, two months after the September 11 attacks, offered the opportunity to assess media behavior and posttraumatic stress reactions associated with the September 11 attacks in remotely-affected individuals who had survived direct exposure to the Oklahoma City bombing seven years earlier. This follow-up study also included a comparison group of Oklahoma City residents who did not directly experience the Oklahoma City bombing. Forty percent of the bombing survivors in this analysis developed bombing-related PTSD, with slightly more than one-half (52.5%, 21/40) of these cases remitted at seven years. By definition, none of the unexposed Oklahoma City community comparison group developed PTSD related to the Oklahoma City bombing, and neither the Oklahoma City bombing survivors nor the Oklahoma City community comparison group were candidates for a diagnosis of September 11-related PTSD based on their lack of qualifying September 11 trauma exposures [2]. This study sought to examine the effect of prior direct exposure to terrorism on Oklahoma City bombing survivors' reactions to media coverage of a new terrorist event with which they had no direct involvement. 4.1. September 11-related posttraumatic stress response scores This study found exposure to the Oklahoma City bombing to be associated with greater arousal in relation to the September 11 attacks. Also, among survivors, having devel-

oped bombing-related PTSD was associated with higher scores on all three September 11 posttraumatic stress response cluster (intrusion, avoidance, and arousal) measures. Prior trauma and pre-existing psychopathology are associated with greater risk for PTSD among individuals exposed to disasters [23,24]. In a sample of directly-exposed trauma survivors, however, Breslau and colleagues [25] found PTSD risk to be associated with prior trauma exposure only if the individual had developed PTSD in relation to the prior event. The current study examining similar relationships in a sample whose experience with a subsequent terrorist attack was remote found that prior exposure to terrorism in the Oklahoma City bombing was associated only with the arousal subscale of the 9/11 IES-R while bombingrelated PTSD was associated with the total 9/11 IES-R score and all 9/11 IES-R subscale scores. Research has identified avoidance and numbing responses as indicative of the core psychopathology of PTSD, with intrusion and arousal symptoms by themselves being generally unassociated with indicators of psychopathology or impairment in functioning [12,26–28]. For example, in a latent class analysis of data from two community samples of participants who were exposed to one or more PTSD diagnosis-qualifying events, emotional numbing distinguished the group with pervasive disturbance from those with no or intermediate disturbance perhaps serving as a marker for PTSD [26]. Others also have observed that criterion C (avoidance and numbing) is a marker for PTSD [29] including North and colleagues [12] in the index Oklahoma City bombing study from which the current sample was derived. Summarizing the literature, Whitman and colleagues [29] noted that, in the absence of meeting criterion C (avoidance and numbing), criterion B (intrusion) and criterion D (hyperarousal) symptoms do not signify psychopathology, and that the majority of individuals who satisfy criterion C (avoidance and numbing) meet full diagnostic criteria. In their paper on validating the diagnosis of PTSD, North and colleagues [30] concluded that avoidance and numbing symptoms emerge as “the core pathology” for identifying PTSD (p. 39). Thus, the arousal experienced by individuals with direct exposure to a prior terrorist incident relative to those without prior terrorism exposure may be a normative response

76

B. Pfefferbaum et al. / Comprehensive Psychiatry 65 (2016) 70–78

rather than indicative of psychopathology. For survivors who develop PTSD related to a previous terrorist experience, the reaction to a new event may be more problematic given the potentially heightened vulnerability associated with the disorder and the finding in this study that bombing-related PTSD was associated with all 9/11 IES-R subscale scores.

scores. Prior studies also have failed to find an association between listening to disaster radio [32] or reading disaster newspaper [33,34] coverage and the psychological outcomes measured. Thus, consistent with a recent review of disaster media coverage, no definitive conclusions can be reached about the association between contact with coverage through these other two media forms and psychological outcomes [35].

4.2. September 11 media behavior The amount of time spent in contact with media coverage (television, radio, and newspaper) of the September 11 attacks was not associated with exposure to the Oklahoma City bombing, development of bombing-related PTSD, or non-remission of bombing-related PTSD. Fear-based discontinuation of media contact was not associated with bombing exposure, but it was associated with the diagnosis of bombing-related PTSD among survivors and with non-remission of PTSD among survivors diagnosed with bombing-related PTSD. Thus, greater time spent in contact with September 11 media coverage was not a result of, or even a marker of, prior direct exposure to terrorism or development of PTSD related to it, but fear-related discontinuation of September 11 media contact may have been a product of PTSD related to the prior terrorist incident. 4.3. Relationship between September 11 media behavior and posttraumatic stress response scores Hours of contact with September 11 television coverage was associated with a higher 9/11 IES-R total score in the full sample but, of the 9/11 IES-R subscales, only the intrusion subscale score was statistically associated with hours of contact with September 11 television coverage after adjusting for other participant characteristics. The association between hours of September 11 television viewing and 9/11 IES-R intrusion is not necessarily causal, however. It is possible that television viewing of disaster coverage may lead to intrusive reactions especially given the graphic imagery commonly contained in television coverage, but it also is possible that those with more intense intrusive reactions are drawn to television coverage. Moreover, this could represent a reinforcing or spiraling process in which contact with media coverage of terrorism intensifies intrusive reactions that in turn drive attention to more terrorism media coverage [31]. After adjusting for gender, age, intrusion, avoidance/ numbing, arousal, and bombing exposure (in the full sample model) and bombing-related PTSD (in the model restricted to survivors), discontinuing September 11 media contact due to fear was statistically associated only with higher September 11 avoidance/numbing scores. Fear-related discontinuation of media contact may itself constitute avoidance/numbing behavior in which participants either reached or surpassed an intolerable threshold of fear or decided to discontinue contact to avoid becoming frightened or distressed. In the current study, listening to radio and reading newspaper coverage of the September 11 attacks were not associated with the 9/11 IES-R total or any 9/11 IES-R subscale

4.4. Limitations Limitations of this study include the high proportion (45.6%) of bombing survivors who were included in the index study and did not complete the study instruments examined in this report. As reported in the Methods section, however, the survivors who were not included in this analysis did not differ statistically from those who did participate in gender, age, race/ethnicity, or bombing-related PTSD at baseline. In addition, the comparison group was a convenience sample and may not be representative of Oklahoma City adult residents who had not been exposed to the bombing. Smith and colleagues [36] found high rates of bombing exposure and higher posttraumatic stress responses in residents of Oklahoma City several months and 1.5 years post event relative to residents of a community that did not experience a bombing (Indianapolis). This suggests that while not measured in this study, many participants in the community comparison group may have been indirectly exposed to the bombing and may have experienced substantial posttraumatic stress in response. Accuracy of the provided information is another potential limitation of the study especially given that the time interval between the September 11 attacks and the assessment was relatively long for some participants, particularly for the community comparison group. Because of the time frame for assessment, participants' reactions may have been misreported. As expected, 9/11 IES-R scores were higher in those assessed closer in time to the attacks, though the relationship between 9/11 IES-R scores and time since event was not statistically different from zero in the multivariable models controlling for demographics, hours of television viewing, and discontinuing contact with media coverage of the attacks due to fear, probably because of insufficient statistical power. The literature is inconclusive about the effects of time on retrospective recall of symptoms of acute stress disorder [37] and psychological distress [38] in individuals directly exposed to an event. To our knowledge, the issue has not been examined in remotely-affected individuals whose contact with an event is through media coverage. Moreover, while the lack of association between respondents' reported media behavior and time since the event does not necessarily establish the accuracy of the information, it suggests that the variation in the amount of media contact reported by the respondents was independent of the time interval. Another potential source of bias is the possibility that the posttraumatic stress responses that participants attributed to September 11 instead reflected their bombing reactions or reactions to another traumatic event. Although the 9/11

B. Pfefferbaum et al. / Comprehensive Psychiatry 65 (2016) 70–78

IES-R queries were anchored in the September 11 attacks, participants may not have been able to distinguish the source of some reactions such as sleep disturbance or arousal. Unfortunately, contact with internet coverage was not examined in this study, although the internet provided dramatic coverage of the September 11 attacks. Internet coverage would seem to possess many of the characteristics associated with television coverage (e.g., unedited with compelling images, spontaneous emotional reactions that may be replayed repeatedly) that have been purported to increase television effects [33]. Few studies have examined contact with disaster internet coverage and adverse emotional outcomes [e.g., 33,39,40]. Thus, future research should explore this association. The current study did not examine details of the media coverage or of participants' media behaviors. Future studies might examine, for example, the specific content of media coverage (e.g., dramatic images of victims, information about perpetrators, safety precautions), various media forms (e.g., television, radio, newspaper, internet, social media), motivations for media contact (e.g., seeking information, passive contact), duration of media contact, reasons for discontinuing contact, a variety of outcomes in diverse samples, and context (i.e., environment in which media contact occurs) to yield a more comprehensive understanding of media effects [see e.g., 35]. There may be positive outcomes associated with media contact which were not measured in this study. For example, media coverage may serve positive functions by providing a major conduit of information including risk communication. The current study did not identify the point in time when participants who discontinued contact with media coverage actually suspended contact and how that might have affected their viewing time. Moreover, participants' motivations for discontinuing media coverage, other than fear, were not studied.

5. Conclusions In this study of Oklahoma City bombing survivors and an unexposed comparison group of Oklahoma City residents all of whom were geographically remote from, and hence not directly exposed to, the September 11 attacks, bombing survivors had higher September 11 arousal reactions than the comparison group after adjusting for other participant characteristics. Only amount of television coverage, but not radio or newspaper coverage, of the attacks was associated with September 11 posttraumatic stress reactions in bombingexposed individuals remote from the incident, and only in the form of intrusion and not avoidance/numbing or arousal reactions. Given that of the three PTSD symptom cluster groups, avoidance/numbing responses are key indicators of psychopathology, the association of the 9/11 IES-R arousal subscale score with bombing exposure in the absence of avoidance/numbing may be a normative response. Although time spent watching television coverage of the September 11

77

attacks and fear-related discontinuation of media contact were not associated with Oklahoma City bombing exposure, discontinuing September 11 media contact due to fear was associated with avoidance/numbing in the full sample and in the analysis restricted to the bombing survivors. Despite similar contact with September 11 television coverage, bombing survivors who developed bombing-related PTSD had stronger September 11 posttraumatic stress reactions than those who did not develop PTSD suggesting greater vulnerability associated with PTSD related to a prior incident in remotely-exposed samples. Additional research is needed to fully understand the relationships between disaster media coverage and adverse outcomes in various populations. Acknowledgment This work was supported by the Oklahoma City National Memorial Institute for the Prevention of Terrorism and the Office of Justice Programs, National Institute of Justice, U.S. Department of Justice (B. Pfefferbaum) and by the National Institute of Mental Health (C. North). None of the authors of this manuscript have any actual or potential conflicts of interest. Points of view in this document are those of the authors and do not represent the official position of Metrocare Services; the National Institutes of Mental Health; the Office of Justice Programs, National Institute of Justice, Department of Justice; the Oklahoma City National Memorial Institute for the Prevention of Terrorism; Phoenix Community College; the U.S. government; the University of Missouri; the University of Oklahoma Health Sciences Center; or the University of Texas Southwestern Medical Center. References [1] North CS. Approaching disaster mental health research after the 9/11 World Trade Center terrorist attacks. Psychiatr Clin North Am 2004;27(3):589-602. [2] American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 5th ed. Washington, DC: The American Psychiatric Association; 2013. [3] Ahern J, Galea S, Resnick H, Kilpatrick D, Bucuvalas M, Gold J, et al. Television images and psychological symptoms after the September 11 terrorist attacks. Psychiatry 2002;65(4):289-300. [4] Ahern J, Galea S, Resnick H, Vlahov D. Television images and probable posttraumatic stress disorder after September 11. The role of background characteristics, event exposures, and perievent panic. J Nerv Ment Dis 2004;192(3):217-26. [5] Ahern J, Galea S, Resnick H, Vlahov D. Television watching and mental health in the general population of New York City after September 11. J Aggression Maltreatment Trauma 2004;9(1/2):109-24. [6] Bernstein KT, Ahern J, Tracy M, Boscarino JA, Vlahov D, Galea S. Television watching and the risk of incident probable posttraumatic stress disorder. A prospective evaluation. J Nerv Ment Dis 2007;195(1):41-7. [7] Schlenger WE, Caddell JM, Ebert L, Jordan BK, Rourke KM, Wilson D, et al. Psychological reactions to terrorist attacks. Findings from the National Study of Americans' Reactions to September 11. JAMA 2002;288(5):581-8. [8] Schuster MA, Stein BD, Jaycox LH, Collins RL, Marshall GN, Elliott MN, et al. A national survey of stress reactions after the September 11, 2001, terrorist attacks. N Engl J Med 2001;345(20):1507-12.

78

B. Pfefferbaum et al. / Comprehensive Psychiatry 65 (2016) 70–78

[9] Silver RC, Holman EA, McIntosh DN, Poulin M, Gil-Rivas V. Nationwide longitudinal study of psychological responses to September 11. JAMA 2002;288(10):1235-44. [10] Neria Y, Sullivan GM. Understanding the mental health effects of indirect exposure to mass trauma through the media. JAMA 2011;306(12):1374-5. [11] Weems CF, Scott BG, Banks DM, Graham RA. Is TV traumatic for all youths? The role of preexisting posttraumatic-stress symptoms in the link between disaster coverage and stress. Psychol Sci 2012;23(11):1293-7. [12] North CS, Nixon SJ, Shariat S, Mallonee S, McMillen JC, Spitznagel EL, et al. Psychiatric disorders among survivors of the Oklahoma City bombing. JAMA 1999;282(8):755-62. [13] North CS, Pfefferbaum B, Kawasaki A, Lee S, Spitznagel EL. Psychosocial adjustment of directly exposed survivors 7 years after the Oklahoma City bombing. Compr Psychiatry 2011;52(1):1-8. [14] Robins LN, Cottler LB, Compton WM, Bucholz K, North CS, Rourke KM. Diagnostic Interview Schedule for the DSM-IV (DIS-IV). St. Louis, MO: Washington University; 2000. [15] North CS, Pfefferbaum B. The Diagnostic Interview Schedule/Disaster Supplement Questionnaire. St. Louis, MO: Washington University; 2002. [16] American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th ed. Washington, DC: The American Psychiatric Association; 2000 [text rev.]. [17] Horowitz M, Wilner N, Alvarez W. Impact of Event Scale: a measure of subjective stress. Psychosom Med 1979;41(3):209-18. [18] Weiss DS, Marmar CR. The Impact of Event Scale–Revised. In: Wilson JP, & Keane TM, editors. Assessing psychological trauma and PTSD. New York: The Guilford Press; 1997. p. 399-411. [19] Pfefferbaum B, Call JA, Lensgraf SJ, Miller PD, Flynn BW, Doughty DE, et al. Traumatic grief in a convenience sample of victims seeking support services after a terrorist incident. Ann Clin Psychiatry 2001;13(1):19-24. [20] Pfefferbaum B, Pfefferbaum RL, Gurwitch RH, Doughty DE, Pynoos RS, Foy DW, et al. Teachers' psychological reactions 7 weeks after the 1995 Oklahoma City bombing. Am J Orthopsychiatry 2004;74(3):263-71. [21] Tucker P, Dickson W, Pfefferbaum B, McDonald NB, Allen G. Traumatic reactions as predictors of posttraumatic stress six months after the Oklahoma City bombing. Psychiatr Serv 1997;48(9):1191-4. [22] Tucker PM, Pfefferbaum B, North CS, Kent A, Burgin CE, Parker DE, et al. Physiologic reactivity despite emotional resilience several years after direct exposure to terrorism. Am J Psychiatry 2007;164(2):230-5. [23] Norris FH, Friedman MJ, Watson PJ, Byrne CM, Diaz E, Kaniasty K. 60,000 disaster victims speak: part I. An empirical review of the empirical literature, 1981–2001. Psychiatry 2002;65(3):207-39. [24] North CS, Oliver J, Pandya A. Examining a comprehensive model of disaster-related posttraumatic stress disorder in systematically studied survivors of 10 disasters. Am J Public Health 2012;102(10):e40-8. [25] Breslau N, Peterson EL, Schultz LR. A second look at prior trauma and the posttraumatic stress disorder effects of subsequent trauma. A prospective epidemiological study. Arch Gen Psychiatry 2008;65(4):431-7.

[26] Breslau N, Reboussin BA, Anthony JC, Storr CL. The structure of posttraumatic stress disorder. Latent class analysis in 2 community samples. Arch Gen Psychiatry 2005;62(12):1343-51. [27] Maes M, Delmeire L, Schotte C, Janca A, Creten T, Mylle J, et al. Epidemiologic and phenomenological aspects of post-traumatic stress disorder: DSM-III-R diagnosis and diagnostic criteria not validated. Psychiatry Res 1998;81(2):179-93. [28] Norris FH. Epidemiology of trauma: frequency and impact of different potentially traumatic events on different demographic groups. J Consult Clin Psychol 1992;60(3):409-18. [29] Whitman JB, North CS, Downs DL, Spitznagel EL. A prospective study of the onset of PTSD symptoms in the first month after trauma exposure. Ann Clin Psychiatry 2013;25(2):e8-e17. [30] North CS, Suris AM, Davis M, Smith RP. Toward validation of the diagnosis of posttraumatic stress disorder. Am J Psychiatry 2009;166(1):34-41. [31] Slater MD. Reinforcing spirals: the mutual influence of media selectivity and media effects and their impact on individual behavior and social identity. Commun Theory 2007;17(3):281-303. [32] Propper RE, Stickgold R, Keeley R, Christman SD. Is television traumatic? Dreams, stress, and media exposure in the aftermath of September 11, 2001. Psychol Sci 2007;18(4):334-40. [33] Cho J, Boyle MP, Keum H, Shevy MD, McLeod DM, Shah DV, et al. Media, terrorism, and emotionality: Emotional differences in media content and public reactions to the September 11th terrorist attacks. J Broad Elect Media 2003;47(3):309-27. [34] Huddy L, Feldman S, Lahav G, Taber C. Fear and terrorism: psychological reactions to 9/11. In: Norris P, Kern M, & Jest M, editors. Framing terrorism: The news media, the government, and the public. New York: Routledge; 2003. p. 255-78. [35] Pfefferbaum B, Newman E, Nelson SD, Nitiéma P, Pfefferbaum RL, Rahman A. Disaster media coverage and psychological outcomes: descriptive findings in the extant research. Curr Psychiatry Rep 2014;16:464, http://dx.doi.org/10.1007/x11920-014-0464-x. [36] Smith DW, Christiansen EH, Vincent R, Hann NE. Population effects of the bombing of Oklahoma City. J Oklahoma State Med Assoc 1999;92(4):193-8. [37] Harvey AG, Bryant RA. Memory for acute stress disorder symptoms. A two-year prospective study. J Nerv Ment Dis 2000;188(9):602-7. [38] Brennan AM, Stewart HA, Jamhour N, Businelle MS, Gouvier WD. An examination of the retrospective recall of psychological distress. J Forensic Neuropsychol 2005;4(4):99-110. [39] Bui E, Rodgers RF, Herbert C, Franko DL, Simon NM, Birmes P, et al. The impact of internet coverage of the March 2011 Japan earthquake on sleep and posttraumatic stress symptoms: an international perspective. Am J Psychiatry 2012;169(2):221-2. [40] Goodwin R, Palgi Y, Hamama-Raz Y, Ben-Ezra M. In the eye of the storm or the bullseye of the media: Social media use during Hurricane Sandy as a predictor of post-traumatic stress. J Psychiatr Res 2013;47(8):1099-100.