Psychiatric Disorders Among Survivors of the Oklahoma City Bombing

ORIGINAL CONTRIBUTION

Carol S. North, MD, MPE Context Disasters expose unselected populations to traumatic events and can be used Sara Jo Nixon, PhD to study the mental health effects. The Oklahoma City, Okla, bombing is particularly significant for the study of mental health sequelae of trauma because its extreme magni- Sheryll Shariat, MPH tude and scope have been predicted to render profound psychiatric effects on survivors. Sue Mallonee, RN, MPH Objective To measure the psychiatric impact of the bombing of the Alfred P. Mur- J. Curtis McMillen, PhD rah Federal Building in Oklahoma City on survivors of the direct blast, specifically examining rates of posttraumatic stress disorder (PTSD), diagnostic comorbidity, func- Edward L. Spitznagel, PhD tional impairment, and predictors of postdisaster psychopathology. Elizabeth M. Smith, PhD† Design, Setting, and Participants Of 255 eligible adult survivors selected from a confidential registry, 182 (71%) were assessed systematically by interviews approxi- mately 6 months after the disaster, between August and December 1995. ISASTERS OFFER UNIQUE OP- portunities to study mental Main Outcome Measures Diagnosis of 8 psychiatric disorders, demographic data, health effects of traumatic level of functioning, treatment, exposure to the event, involvement of family and friends, events in unselected popula- and physical injuries, as ascertained by the Diagnostic Interview Schedule/Disaster Supple- tions.D Because most disasters strike ran- ment. domly, studies of disasters circumvent Results Forty-five percent of the subjects had a postdisaster psychiatric disorder and 34.3% had PTSD. Predictors included disaster exposure, female sex (for any postdi- the limitations of research on trauma to 2 saster diagnosis, 55% vs 34% for men; ␹1 = 8.27; P = .004), and predisaster psychi- individuals in the community, where ␹ 2 atric disorder (for PTSD, 45% vs 26% for those without predisaster disorder; 1 = 6.86; risk for traumatic events is con- P = .009). Onset of PTSD was swift, with 76% reporting same-day onset. The rela- founded with vulnerability to psycho- tively uncommon avoidance and numbing symptoms virtually dictated the diagnosis 1 pathology. The extreme magnitude and of PTSD (94% meeting avoidance and numbing criteria had full PTSD diagnosis) and intensity of the Oklahoma City, Okla, were further associated with psychiatric comorbidity, functional impairment, and treat- bombing made it a particularly signifi- ment received. Intrusive reexperience and hyperarousal symptoms were nearly uni- cant subject for the study of mental versal, but by themselves were generally unassociated with other psychopathology or health effects of trauma because of the impairment in functioning. profound effects anticipated among its Conclusions Our data suggest that a focus on avoidance and numbing symptoms survivors, including persons with no could have provided an effective screening procedure for PTSD and could have iden- predisaster psychiatric history.2-6 tified most psychiatric cases early in the acute postdisaster period. Psychiatric comor- The bombing of the Alfred P. Mur- bidity further identified those with functional disability and treatment need. The nearly rah Federal Building in Oklahoma City universal yet distressing intrusive reexperience and hyperarousal symptoms in the majority of nonpsychiatrically ill persons may be addressed by nonmedical interventions on April 19, 1995, was the most se- of reassurance and support. vere incident of terrorism ever experi- JAMA. 1999;282:755-762 www.jama.com enced on American soil.7,8 The death count totaled 167, including 19 children; the number of persons injured to- cluded documenting rates of postdi- Author Affiliations: School of Medicine, Depart- ment of Psychiatry (Drs North and Smith), George War- taled 684. The fatality rate inside the saster psychopathology, examining ren Brown School of Social Work (Dr McMillen), and Murrah Building was 46%, and 93% of functional impact, and identifying pre- Department of Mathematics and Statistics (Dr Spitzna- survivors who were in the building were dictors of these difficulties to help guide gel), Washington University, St Louis, Mo; and De- 9 partment of Psychiatry and Behavioral Sciences, Uni- injured. The explosion demolished or mental health intervention workers in versity of Oklahoma Health Sciences (Dr Nixon), and damaged more than 800 building struc- future disasters. We anticipated that the the Oklahoma State Department of Health (Mss Shariat and Mallonee), Oklahoma City. tures, with an estimated property dam- scope and severity of this event would †Dr Smith died March 7, 1997. age of $625 million. elicit higher rates of psychopathology Corresponding Author and Reprints: Carol S. North, MD, MPE, Department of Psychiatry, Washington Uni- We studied direct survivors of the than previous disasters studied using versity School of Medicine, 4940 Children’s Pl, St Louis, blast. Our research objectives in- similar research methods. MO 63110 (e-mail: [email protected]).

Downloaded From: https://jamanetwork.com/ on 10/02/2021 PSYCHIATRIC IMPACT ON OKLAHOMA CITY BOMBING SURVIVORS

METHODS Journal Record buildings) or in less able during the design of longitudinal Subjects damaged buildings, or outdoors. The disaster studies.11,12 Diagnostic infor- Declaration of all bombing-associated remaining 18% were more than 184 m mation was obtained for 8 psychiatric injuries and illnesses as reportable cases from the detonation point. disorders: posttraumatic stress disor- by the Oklahoma State Department of The 35 individuals (16% of those lo- der (PTSD), major depression, panic Health commissioner led to the devel- cated and eligible) who did not partici- disorder, generalized anxiety disor- opment of a confidential registry of sur- pate did not differ from study partici- der, somatization disorder, alcohol use vivors, from which the study sample pants in age, sex, injury rates, or disorder, drug use disorder, and anti- was drawn. The registry contained 1092 medical treatment received. Signifi- social personality disorder. The inter- names of survivors directly exposed to cantly more participants than nonpar- view also documented demographic the blast based on their proximity to the ticipants were in the most heavily dam- data, level of functioning, and treat- ␹ 2 Murrah Building. Persons exposed only aged buildings (79% vs 55%; 1 = 9.69, ment received. The Disaster Supple- indirectly through search and rescue or P = .002), indicating that persons with ment elicited subjects’ disaster-related clean-up efforts or by bereavement less intense exposure may have been experience including exposure to the alone were not listed. Further detail on less likely to participate in the study. event, involvement of family and the development of the Oklahoma State The sample was representative of the friends, and physical injuries. All in- Department of Health registry is pro- health department’s registry popula- terviews were administered by mem- vided in an earlier publication.8 tion with respect to sex and age. Com- bers of the Washington University di- Participation in the study was lim- pared with the registry population, sig- saster research team who received ited to subjects at least 18 years old. nificantly more study subjects had been formal training to administer the DIS. Those too severely injured to partici- in the most heavily damaged build- Sixty-three percent of the interviews ␹ 2 Ͻ pate were excluded as ineligible. To com- ings (79% vs 66%; 1 = 11.25, P .001) were conducted in person, but due to mence with interviewing with minimal and specifically in the Murrah build- logistics in the field, 25% were con- ␹ 2 Ͻ delay, the first 20 registry members to ing (31% vs 18%; 1 = 15.88, P .001). ducted by telephone, and another 12% complete and return a preliminary health The nonrandomly selected subjects initiated in person were completed by department survey of their demograph- did not differ from the 162 selected by telephone. No associations of any rel- ics, exposure to the blast, injuries, and randomization in demographics, pre- evant variables with telephone inter- medical treatment were selected for this disaster psychiatric disorder, or any di- views were identified in the data. study. Subsequently, using an SAS com- agnosis made after the disaster. A higher puter program (SAS 6.12; SAS Institute proportion of them, however, were in Data Analysis Inc, Cary, NC), names of 242 addi- the Murrah building (50% vs 27% of the Because individuals could have had ␹ 2 tional eligible registry entries were ran- others; 1 = 4.45, P = .04). Removal of PTSD resulting from other traumatic domly selected from the registry, which the 20 nonrandomly selected subjects events besides the bombing, diagnoses included 1 of the 20 nonrandomized from the sample effectively reduces the and symptoms of bombing-associated subjects. Of the 261 subjects thus se- proportion of Murrah building occu- PTSD were tabulated separately from lected, 3 had left the country, 1 did not pancy from 31% to 27%, but the dif- those associated with other traumas. It speak English, and 2 had died in the in- ference from the registry remains sig- is well established that traumatic events ␹ 2 terim, precluding their participation. Of nificant ( 1 = 9.68, P = .002). experienced by individuals in the com- the 255 remaining subjects, 32 (13%) Approval for the study was ob- munity disproportionately strike per- could not be contacted, 35 (14%) re- tained from the Washington Univer- sons with proclivities to psychopathol- fused to be interviewed, 6 (2%) were not sity School of Medicine Institutional Re- ogy, suggesting that PTSD following interviewed with no reason available, view Board, St Louis, Mo. All subjects sporadic traumas to individuals in the and 182 (71%) were interviewed. provided written informed consent community may represent a somewhat Thirty-one percent of the study sub- prior to participating. different phenomenon from the PTSD jects were within 46 m of the bomb (dis- arising from a community-wide disas- tance selected for 99% of all deaths oc- Interviews ter such as the bombing.1 Therefore, 4 curring within this radius) at the instant An average of 6 months (range, be- cases of PTSD unrelated to the bomb- of detonation. All of these subjects were tween August and December 1995, 4-8 ing were excluded from calculation of located in the Murrah building, ex- months for most) after the event, sub- postdisaster PTSD prevalence. cept for 1 who was in the Athenian jects were interviewed using the Diag- Associations between categorical vari- building, which stood across the street nostic Interview Schedule (DIS)/ ables were tested using ␹2 analyses, sub- and was in the direct path of the blast. Disaster Supplement, which is based on stituting Fisher exact tests when ex- Another 51% were 46 to 184 m from the Diagnostic and Statistical Manual of pected numbers in cells were less than the point of detonation in heavily dam- Mental Disorders, Revised Third Edition 10 5. Linear regression analyses were per- aged (YMCA, Water Resources, and (DSM-III-R), the operating criteria avail- formed to compare numeric variables.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 PSYCHIATRIC IMPACT ON OKLAHOMA CITY BOMBING SURVIVORS

For comparisons of means on re- mined incident and recurrent or per- cohol and drug use disorders were re- peated measures, McNemar tests were sistent disorders by assessing whether ported as inactive after the disaster. For performed. Statistical significance was the individual had met criteria for the all diagnoses except generalized anxi- set at ␣Ͻ.05. same disorder at any time before the ety disorder, postdisaster occurrence of bombing. Overall, nearly half the the disorder was statistically associ- RESULTS sample met criteria for 1 or more psy- ated with predisaster history of the same Demographics and chiatric diagnoses after the disaster, disorder. Disaster Experience with more than one third qualifying for Fifty-seven percent of subjects with TABLE 1 displays the demographics of a diagnosis of PTSD specific to the bombing-related PTSD had a predisas- the sample, which had roughly equal bombing. Women had at least twice the ter lifetime history of psychiatric ill- sex representation and was largely rate of PTSD as men (45% vs 23%, re- ness. Subjects with a predisaster disor- ␹ 2 white. Eighty-seven percent of study spectively; 1 = 9.44; P = .002), major der were more likely than others to participants reported injuries sus- depression (32% vs 13%, respectively; ␹ 2 tained in the blast, and 77% overall 1 = 9.82; P = .002), and generalized had required medical intervention, anxiety disorder (9% vs 0%, respec- Table 1. Demographics of Study including hospitalization (20%) and tively, P = .007). Women were more Participants (N = 182) No. (%) of surgery (15%). The most prevalent likely to qualify for any postdisaster di- Characteristic Subjects ␹ 2 injuries were lacerations (76%), fol- agnosis (55% vs 34% for men; 1 = 8.27; Sex lowed by contusions (50%), skin- P = .004). No subjects met criteria for Male 88 (48.4) Female 94 (51.6) embedded glass or metal shards somatization disorder or antisocial per- Age, y (46%), hearing loss (34%), and smoke sonality disorder. 18-29 19 (10.4) or dust inhalation (23%). Table 2 also shows that 15% of the 30-44 84 (46.2) 45-64 74 (40.7) Eighty-two percent of survivors re- sample had experienced PTSD at some Ն65 5 (2.7) ported witnessing others being in- time before the bombing, and 43% had Mean (SD), y [range] 43.0 (11.5) [19-89] Race jured or killed at the bombing scene, any predisaster lifetime diagnosis. Sev- Non-Hispanic white 162 (89.0) and 46% recalled thinking they were go- enty-four percent of the subjects who Black 16 (8.8) Hispanic 2 (1.1) ing to die during the event. Forty- experienced PTSD had not experi- Other 2 (1.1) three percent reported loss of a family enced it before the bombing, and 56% Education Less than high school 7 (3.8) member or friend in the bombing, and of subjects who experienced major de- High school graduate 97 (53.3) 92% personally knew someone in- pression had no history of it before the College graduate 45 (24.7) jured or killed. bombing. Preexisting major depres- Postgraduate 33 (18.1) Mean (SD), y [range] 14.3 (2.2) [8-17] sion was especially likely to persist or Marital status PTSD and Other recur after the bombing (78% of pre- Married 115 (63.2) Psychiatric Diagnoses Widowed 8 (4.4) disaster cases). Incident postdisaster Separated 4 (2.2) TABLE 2 displays rates of predisaster and substance use disorders were not ob- Divorced 37 (20.3) postdisaster disorders. We deter- served. The majority of predisaster al- Single (never married) 10 (9.9)

Table 2. Predisaster and Postdisaster Diagnostic Disorders* No. (%) of Subjects With Postdisaster Disorder†

Generalized Alcohol Drug Any Major Panic Anxiety Use Use Non-PTSD Any No No. (%) of Subjects PTSD‡ Depression Disorder Disorder Disorder Disorder Diagnosis Diagnosis‡ Diagnosis With Predisaster Disorder† 62 (34.3) 41 (22.5) 12 (6.6) 8 (4.4) 17 (9.4) 4 (2.2) 55 (30.2) 82 (45.1) 100 (54.9) Subjects With Predisaster Disorder and Postdisaster Disorder PTSD, 27 (15.0) 16§ 10 1 4 4 1 13 23§ 4§ Major depression, 23 (12.6) 15 18§ 5 4 4 1 18 18§ 5§ Panic disorder, 5 (2.8) 4 5 4§ 2 1 1 5 5§ 0§ Generalized anxiety disorder, 5 (2.8) 4 4 2 3 1 1 4 5§ 0§ Alcohol use disorder, 48 (28.7) 18 12 5 1 17§ 2 23 30§ 18§ Drug use disorder, 17 (9.4) 7 7 2 2 4 4§ 10 11§ 6§ Any predisaster disorder, 79 (43.4) 35 27 9 5 17 4 39 52 27 No disorder, 103 (56.6) 27 14 3 3 0 0 16 30 73 *Columns and rows do not sum because each subject could have more than 1 diagnosis. †Percentages represent the total number of the 182 subjects. ‡Includes only cases with bombing-related posttraumatic stress disorder (PTSD). §PϽ.05.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 PSYCHIATRIC IMPACT ON OKLAHOMA CITY BOMBING SURVIVORS

experience bombing-related PTSD (45% who were diagnosed as having postdi- ter the bombing, its course was chronic. vs 26% for those without a predisaster saster PTSD also had postdisaster Eighty-nine percent of subjects with ␹ 2 disorder; 1 = 6.86; P = .009). All post- comorbidity vs 30% of those with a pre- bombing-related PTSD reported that disaster disorders were significantly as- disaster disorder who did not experi- they were still symptomatic during the ␹ 2 Ͻ sociated with history of predisaster psy- ence PTSD ( 1 = 14.98; P .001). Sixty month before the interview (which was chopathology. Sixty-three percent of percent of the bombing survivors had at least 3 months after the bombing), de- subjects with any active postdisaster experienced a psychiatric disorder at fining their PTSD as chronic.10 Sixty- psychiatric disorder had a predisaster some time in their lives either before three percent of the entire sample, in- diagnosis; ie, more than one third of or after the bombing. cluding those not meeting full PTSD those with a postdisaster disorder had Onset of PTSD was acute. Of 62 sub- criteria, acknowledged having experi- never had a psychiatric disorder be- jects with bombing-related PTSD, 47 enced some disaster-related PTSD symp- fore the bombing. Conversely, 66% of (76%) reported immediate (same day) toms within the past month. subjects who had at least 1 predisaster onset, another 11 (94% cumulative) in psychiatric disorder had an active dis- the first week, 3 more by the end of the Functional Impairment, Coping, order afterward, compared with 29% of first month (98% cumulative), and only and Treatment those who had no psychiatric history 1 more between 1 and 6 months. Due The effects of PTSD on occupational ␹ 2 Ͻ ( 1 = 24.32; P .001). to the timing of the index interviews, and social functioning reported by the In 63% of the cases, PTSD was ac- subjects had little or no time to de- subjects suggest the clinical impor- companied by postdisaster comorbid- velop delayed PTSD, defined in the DSM- tance of this disorder (TABLE 3). More ity, occurring most often in 55% of the III-R10 as onset more than 6 months af- than half of subjects with PTSD alone subjects with PTSD who also were di- ter the traumatic event. Comparing onset and the vast majority of those with co- agnosed as having major depression. information with 44 PTSD-producing morbid PTSD reported that their PTSD Only 9% of the sample subjects had a traumas at some other time in their lives symptoms interfered with their activi- non-PTSD postdisaster diagnosis in the (for which there had been ample oppor- ties; similar numbers in each group absence of PTSD. Only 4% of subjects tunity to observe timing of onset well be- were dissatisfied with their work per- without any predisaster disorder and no yond 6 months), 32 (73%) of the cases formance after the disaster. Negative PTSD after the bombing had any non- reportedly began the same day, and an- changes in personal relationships as a PTSD diagnosis afterward compared other 10 (95% cumulative) the same result of the bombing were acknowl- with 48% of those with no predisaster week. Two delayed-onset cases (5% of edged by 75% of subjects with PTSD disorder but who had bombing- the total PTSD) were reported: 1 begin- compared with 27% of those without Ͻ ␹ 2 Ͻ related PTSD (Fisher exact P .001). ning between 6 and 12 months, and 1 this diagnosis ( 1 = 20.53; P .001). Di- Conversely, 74% of the subjects who more than 3 years afterward. Even agnostic comorbidity with PTSD was had preexisting psychopathology and though PTSD onset was very acute af- specifically associated with effects on

Table 3. Functional Indicators of Posttraumatic Stress Disorder (PTSD) and Coping Measures by Diagnostic Subgroups* No. (%) of Subjects in Diagnostic Groups After Disaster

PTSD and Non-PTSD PTSD Only Comorbid Diagnosis Diagnosis Only No Diagnosis Total Sample (n = 23) (n = 39) (n = 16) (n = 98) (N = 176) Functional Indicators Functional interference 12/23 (52.2)† 34/39 (87.2)† 4/15 (26.7) 14/85 (16.5)‡ 64/162 (39.5) Dissatisfaction with work performance 10/19 (52.6) 21/27 (77.8)† 4/13 (30.8) 34/86 (39.5)§ 69/145 (47.6) Permanently worsened relationships with Spouse or significant other 2/22 (9.1) 10/32 (31.3)† 3/14 (21.4) 4/87 (4.6)‡ 19/155 (12.3) Other household members 1/11 (9.1) 7/30 (23.3)࿣ 0/12 (0.0) 1/49 (2.0)§ 9/102 (8.8) Other relatives or friends 4/9 (44.4)¶ 13/30 (43.3)† 3/12 (25.0) 5/50 (10.0)‡ 25/101 (24.8) Coping Measures Took medication 10/23 (43.5) 27/37 (73.0)† 4/15 (26.7) 22/89 (24.7)‡ 63/164 (38.4) Drank alcohol 3/23 (13.0) 12/37 (32.4)† 6/15 (40.0)† 5/89 (5.6)‡ 26/164 (15.9) Turned to family and friends 21/23 (91.3) 34/37 (91.9) 14/15 (93.3) 80/88 (90.9) 149/163 (91.4) *Subjects with nonbombing specific PTSD and subjects for whom postdisaster diagnsosis data were not available are excluded from this analysis. Numerators indicate the number of subjects reporting the presence of each item. Denominators indicate the total number of subjects in the diagnostic grouping who responded to each item. †PϽ.001, compared with subjects with no diagnosis. ‡PϽ.001, compared with subjects with a diagnosis. §PϽ.05, compared with subjects with a diagnosis. ࿣PϽ.01, compared with subjects with no diagnosis. ¶PϽ.05, compared with subjects with no diagnosis.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 PSYCHIATRIC IMPACT ON OKLAHOMA CITY BOMBING SURVIVORS

relationships with spouses and other sician (5%) or pastor (3%). The highest usual human experience and that would household members. Nearly 40% of all rates of professional mental health ser- be markedly distressing to almost any- the survivors used medication to cope, vices utilization were among those with one.”10 The subject must then present including about 25% of those who did comorbid PTSD (72%). Subjects who with at least 1 of the symptoms from the not experience any postdisaster psy- had PTSD but who did not have a post- intrusive and reexperience category chiatric disorder, and 73% of those with disaster comorbid disorder did not use (group B), have at least 3 symptoms from comorbid PTSD (Table 3). Only those mental health services more often than the avoidance and numbing category persons with PTSD that was compli- those who had a comorbid disorder (group C), and at least 2 of the symp- cated by comorbidity were using medi- (39% vs 29%, respectively, P = .32; toms from the hyperarousal category ␹2 cation or alcohol as a coping mecha- 1 = 0.978). (group D). These symptoms must last nism. Regardless of diagnostic status, for at least 1 month and must be severe turning to others for support was a PTSD Symptom Groups enough to cause subjective distress or nearly universal response. FIGURE 1 shows the rates of each of the functional impairment. PTSD symp- Mental health treatment was abun- PTSD symptoms that are arranged by toms were nearly universal: only 7 sub- dant. Sixty-nine percent of the survi- DSM-III-R symptom groups: group B (in- jects (4%) reported no bombing- vors received some kind of mental trusive reexperience), group C (avoid- related PTSD symptoms. The 2 most health intervention after the disaster; ance and numbing), and group D (hy- commonly experienced symptoms were 40% had participated in debriefings and perarousal). To be diagnosed as having in the hyperarousal category: difficulty 41% had sought professional mental PTSD, according to DSM-III-R criteria, concentrating (78%) and exaggerated health treatment, but only 16% had a subject must first be exposed to a trau- startle response (77%). The 3 least ex- been treated by a psychiatrist. Few in- matic stressor, which the DSM-III-R de- perienced symptoms were in the avoid- dividuals had received mental health in- fines as criterion A, and were exposed ance and numbing category: sense of tervention from their primary care phy- to an “event that is outside the range of foreshortened future (19%), restricted

Figure 1. Individual Posttraumatic Stress Disorder (PTSD) Symptoms and Criterion Groups

Intrusive Memories Criterion Group B, Dreams or Nightmares Intrusive Reexperience Symptoms Flashbacks Upset by Reminders Meets Criterion Group B

Avoids Thoughts or Feelings Avoids Reminders Psychogenic Amnesia Criterion Group C, Avoidance and Numbing Loss of Interest Symptoms Detachment or Estrangement Restricted Range of Affect Sense of Shortened Future Meets Criterion Group C

Insomnia Irritability or Anger Criterion Group D, Difficulty Concentrating Hyperarousal Symptoms Hypervigilance Jumpy or Easily Startled Physiologic Reactivity Meets Criterion Group D

0 102030405060708090 % of Subjects

Because the Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition10 defines criterion A as exposure to an “event that is outside the range of usual human experience and that would be markedly distressing to almost anyone” as its criterion and does not list symptoms, criterion A was not included in this figure. The criterion groups are delineated with letters B, C, and D. The lighter bar in each group represents the percentage of persons meeting criterion for each group.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 PSYCHIATRIC IMPACT ON OKLAHOMA CITY BOMBING SURVIVORS

range of affect (13%), and psychogenic groups alone (when avoidance and Variables associated with a non-PTSD amnesia (12%). numbing criteria were not met). The disorder after the bombing included fe- The vast majority of survivors ful- avoidance and numbing criterion group male sex (39% women vs 21% men; ␹ 2 filled criteria for intrusive reexperi- was also associated with receiving treat- 1 = 7.71, P = .006) and number of di- ence and hyperarousal categories. Only ment, whereas intrusion and hyper- saster-related injuries (4.6 [4.2] vs 3.4 one third of the total subjects fulfilled arousal in its absence were not. As seen [2.7]; t = 2.35; P = .02). Postdisaster ma- the avoidance and numbing criteria. in Figure 2, the avoidance and numb- jor depression was not more prevalent The avoidance and numbing criteria ing group (and to a much smaller ex- among those who had lost a friend or were highly specific for the diagnosis tent, hyperarousal) was associated with relative in the disaster, nor was the num- of PTSD: 94% of the subjects who had reports of functional interference. The ber of depressive symptoms higher in this fulfilled avoidance and numbing crite- avoidance and numbing criterion group group. Controlling for the confounding ria met full PTSD criteria related to the was also associated with dissatisfac- effects of sex on education and marital bombing. By DSM-III-R requirement, all tion with work performance. status (women having less education and subjects meeting criteria for PTSD ful- being more often divorced or separated filled the avoidance and numbing cri- Other Predictors compared with men), these 2 variables teria (100% sensitivity). Subjects with postdisaster PTSD re- were not associated with PTSD or other FIGURE 2 shows that the avoidance ported a mean (SD) of 5.7 (4.2) inju- postdisaster psychopathology. and numbing criterion group was sig- ries, compared with 3.1 (2.4) injuries nificantly associated with predisaster among others (Wilcoxon z = 3.14, COMMENT psychopathology and with postdisas- P= .002). Those reporting injury or death The Oklahoma City bombing pro- ter comorbidity, associations gener- of a family member or friend in the vided a rare opportunity to study men- ally not observed in conjunction with bombing had higher rates of PTSD than tal health effects resulting from a se- ␹ 2 intrusion and hyperarousal symptom others (43% vs 25%; 1 = 5.02, P = .03). verely traumatic event in an essentially

Figure 2. Diagnosis, Treatment, and Functional Indicators Associated With Posttraumatic Stress Disorder (PTSD) Criterion Groups

80 Comorbid Postdisaster Predisaster Receiving Mental Health ∗ Diagnosis Disorder Treatment 70 ∗ † 60

20 % With Diagnosis or Treatment 10

Group B Group D Group C Group B Group D Group C Group B Group D Group C Criterion Group C Not Met Criterion Group C Not Met Criterion Group C Not Met

Interference With ∗ Dissatisfaction With Work 70 Activities Performance ∗ 60

50 PTSD Criterion Group 40 Does Not Meet Criteria 30 † Meets Criteria

% With Impairment 20

Group B Group D Group C Group B Group D Group C Criterion Group C Not Met Criterion Group C Not Met

In this figure, criteron group categories are not mutually exclusive, with the exception of categories group B and group D, which exclude those who met criteron group C. An asterisk indicates PϽ.001; a dagger, PϽ.01.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 PSYCHIATRIC IMPACT ON OKLAHOMA CITY BOMBING SURVIVORS

unselected population. This study docu- the most commonly associated disor- The relatively uncommon postdisas- mented extensive psychopathology in der, and most preexisting depression re- ter avoidance and numbing symp- a highly exposed sample of direct vic- curred or persisted in the period after the toms were virtually tantamount to the tims of the blast. bombing. No new cases of substance diagnosis. Avoidance and numbing abuse were observed, consistent with symptoms were associated with preex- Postdisaster Psychopathology previous findings20-23,25 pertaining to new isting and comorbid postdisaster psy- Nearly half the bombing survivors stud- postdisaster alcohol use disorders after chopathology, functional impair- ied had an active postdisaster psychi- other events studied by this team. ment, use of medication and alcohol to atric disorder, and full criteria for PTSD This study found several predictors of cope, and treatment received—unlike were met by one third of the survi- bombing-related PTSD: degree of disas- the more prevalent intrusive reexperi- vors. PTSD symptoms were nearly uni- ter exposure (represented by number of ence and hyperarousal symptoms only, versal, especially symptoms of intru- injuries), female sex, preexisting psy- which did not show these associa- sive reexperience and hyperarousal. chopathology, and secondary expo- tions. These observations confirm this An explosion in a Norwegian paint sure through loved ones (injury and team’s previously published findings in factory studied by Weisæth13,14 repre- death). Physical injuries and involve- studies of an earthquake in North- sented a similar type of disaster, al- ment of loved ones may represent spe- ridge, Calif,24 and a mass murder epi- though it was considerably smaller in cific mechanisms for generation of sode at a cafeteria in Killeen, Tex.34 scope and magnitude, with far less mor- psychiatric sequelae of disasters. The tality and morbidity (6 fatalities, 2 in- predominance in postdisaster psycho- Implications for Mental Health capacitating injuries, and another 21 mi- pathology in women has been reported Intervention and Policy nor injuries among 125 survivors). The in previous disaster studies4,13,14,23,26-29 and in the Postdisaster Setting high-exposure group had a 43% rate of was not unexpected because the disor- Because virtually all the cases of PTSD PTSD at 10 weeks after the bombing, di- ders classically observed after disasters— started acutely after the bombing, the minishing to 36% by 7 months. Only a depression and anxiety disorders—are most efficient plan would be to expe- few other disaster studies of nonmili- more prevalent among women in the dite large-scale efforts to identify sur- tary populations have reported higher general population. Preexisting psycho- vivors with psychiatric illness as soon PTSD rates: 44% after the Buffalo Creek pathology has also been identified as a as possible. Because most individuals Dam break and floods,15 53% after Aus- robust predictor of PTSD by previous with any psychiatric disorder had PTSD, tralian bushfires,16 54% after an air- studies of this team4,20-24 and oth- focusing on PTSD could identify most plane crash landing,17 and 50% to 100% ers.13,14,30-32 The 43% rate of lifetime pre- cases for triage to psychiatric care. after a plane crash into a shopping mall.18 disaster psychiatric illness in the Okla- Shortages of resources encountered in Differences in research methods, such homa sample does not exceed the acute disaster settings make it impor- as use of unstructured interviews and expected general population lifetime tant to focus attention on those at great- self-report scales known to be associ- rates of 48% in a large population as- est risk for PTSD. This study found ated with higher estimates of psycho- sessed with structured interviews33 and highest risk among women, individu- pathology,19 unfortunately preclude is not significantly higher than rates als with more direct and indirect disas- meaningful comparisons. of preexisting illness in other disaster ter exposure (defined by the number of Comparison across disasters is pos- sites studied by this disaster research personal injuries and secondary expo- sible within the Washington University team.20,23,24,34 sure through loved ones), and sub- research database on several different di- jects with a predisaster psychiatric his- saster events studied using uniform Observations on PTSD tory. The data indicate that PTSD may methods. The 34% rate of PTSD after the This study provided important observa- be readily and efficiently identifiable bombing was the highest of all the di- tions on the character and early course with truncated assessment for avoid- sasters studied to date by this team.4,20-24 of PTSD following a particularly severe ance and numbing criteria only. Rates of PTSD in these other studies were disaster. Symptom onset was rather im- Once PTSD is identified, as sug- 2% following a tornado,22 28% after a mediate—usually the same day—and gested by the comorbidity data col- mass shooting episode,23 and 29% after few other cases developed after the first lected from the Oklahoma bombing, cli- a plane crash into a hotel.20 month. This rapidity of onset is consis- nicians would be well advised to The degrees of both occupational and tent with other traumatic events sub- continue searching for other psychopa- social impairment associated with PTSD jects had experienced and with find- thology, a finding verified elsewhere in after the bombing demonstrate the clini- ings of 2 other studies.24,34 In the the literature.20,24,34 These data suggest cal importance of this disorder. These Weisæth13,14 paint factory explosion that subjects with comorbidity will be functional effects of PTSD appeared to study, 114 of 117 symptomatic subjects significantly more impaired by their psy- be mediated in large part by its psychi- reported symptom onset within 5 hours, chopathology. The chronicity of PTSD atric comorbidity. Major depression was and the remaining 3 within 32 days. identified in this study (with 9 of 10

Downloaded From: https://jamanetwork.com/ on 10/02/2021 PSYCHIATRIC IMPACT ON OKLAHOMA CITY BOMBING SURVIVORS

cases still symptomatic at interview an health professionals with nonmedical assessment at a single time, however, lim- average of 6 months after the disaster, interventions such as public and work- its findings to description of only the first fulfilling DSM-IV35 criteria of at least 3 place debriefings. The therapeutic tools 6 months after the event. Further study months’ duration for chronicity) indi- for these uncomplicted intrusion and is needed to chart the course of postdi- cates that availability of ongoing treat- hyperarousal syndromes will be edu- saster psychiatric disorders over a longer ment of PTSD is essential. cation, general support, and reassur- period , to observe for the development In the absence of avoidance and ance that the symptoms are normal and of delayed cases of PTSD, and to iden- numbing, the nearly ubiquitous intru- not evidence of impending psychiat- tify predictors of chronicity vs recov- sive reexperience and hyperarousal ric illness. ery, including potential effects of treat- symptoms were associated with little to ment. Naturalistic observation studies no functional impairment or psychiat- Research Limitations such as this one suffer from confound- ric comorbidity. This suggests differ- and Future Work ing of outcomes with seeking treat- ent management strategies for these Because the study sample was slightly ment whose benefits cannot be assessed normative yet distressing symptoms skewed toward proximity to the blast, under the available design. Additional from the professional assessment and the findings may reflect a mildly elevated study with uniformly applied methods intervention generally advised for estimate of psychiatric impact of the across various disaster events will allow avoidance and numbing responses. The disaster population as a whole. Two merging the data to generate statistical nonpathological nature of intrusion and major strengths of the study are its ran- power for untangling disaster-specific hyperarousal symptoms uncompli- dom sampling that maximized the gen- confounders, modeling complex hypoth- cated by avoidance and numbing sug- eral representativeness of the registry eses, and generalizing across events.36 gests that after major psychiatric ill- population and the structured research Funding/Support: This research was supported by re- ness is ruled out, these symptoms may interview that generated psychiatric diag- search grant MH40025 from the National Institute of be managed by nonphysician mental noses. The cross-sectional nature of Mental Health, Bethesda, Md.

REFERENCES 1. North CS. Human response to violent trauma. Bail- industrial disaster. In: Pichot P, Berner P, Wolf R, Thau of disaster on previously assessed mental health. In: lieres Clin Psychiatry. 1995;1:225-245. K, eds. Psychiatry: The State of the Art. New York, Shore JH, ed. Disaster Stress Studies: New Methods 2. Hocking F. Psychiatric aspects of extreme environ- NY: Plenum Press; 1985:299-307. and Findings. Washington, DC: American Psychiatric mental stress. Dis Nerv Syst. 1970;31:542-545. 14. Weisæth L. The stressors and the post-traumatic Press; 1986:22-48. 3. Breslau N, Davis GC. Posttraumatic stress disorder stress syndrome after an industrial disaster. Acta Psy- 26. Moore HE, Friedsam HJ. Reported emotional stress in an urban population of young adults: risk factors for chiatr Scand. 1989;80:25-37. following a disaster. Soc Forces. 1959;38:135-138. chronicity. Am J Psychiatry. 1992;149:671-675. 15. Grace MC, Green BL, Lindy JL, et al. The Buffalo 27. Kasl SV, Chisholm RE, Eskenazi B. The impact of 4. Smith EM, North CS, Spitznagel EL. Post- Creek Disaster: a 14-year follow-up. In: Wilson JP, the accident at Three Mile Island on the behavior and traumatic stress in survivors of three disasters. In: Allen Raphael B, eds. International Handbook of Trau- well-being of nuclear workers. Am J Public Health. RD, ed. Handbook of Post Disaster Interventions [spe- matic Stress Syndromes. New York, NY: Plenum; 1993: 1981;71:472-495. cial issue]. J Soc Behav Pers. 1993;8:353-368. 441-449. 28. Lopez-Ibor JJ Jr, Canas SF, Rodriguez-Gamazo M. 5. Feinstein A, Dolan R. Predictors of post-traumatic 16. McFarlane AC. Posttraumatic morbidity of a di- Psychological aspects of the toxic oil syndrome catas- stress disorder following physical trauma: an exami- saster: a study of cases presenting for psychiatric treat- trophe. Br J Psychiatry. 1985;147:352-365. nation of the stressor criterion. Psychol Med. 1991; ment. J Nerv Ment Dis. 1986;147:4-13. 29. Steinglass P, Gerrity E. Natural disasters and post- 21:85-91. 17. Sloan P. Posttraumatic stress in survivors of an air- traumatic stress disorder: short-term vs long-term re- 6. Shore JH, Tatum EL, Vollmer WM. The Mount St plane crash-landing: a clinical and exploratory re- covery in two disaster-affected communities. J Appl Helens stress response syndrome. In: Shore JH, ed. Di- search intervention. J Trauma Stress. 1988;1:211- Soc Psychol. 1990;20:1746-1765. saster Stress Studies: New Methods and Findings. 229. 30. Bromet EJ, Parkinson DK, Schulberg HC. Mental Washington, DC: American Psychiatric Press; 1986: 18. Newman JP, Foreman C. In: Program and ab- health of residents near the Three Mile Island reac- 77-97. stracts of the Annual Meeting of the International So- tor: a comparative study of selected groups. J Prev Psy- 7. Nixon SJ, Vincent R, Krug RS, Pfefferbaum B. Struc- ciety for Traumatic Stress Studies; 1987; Baltimore, Md. chiatry. 1982;1:225-276. ture and organization of research efforts following the 19. Rubonis AV, Bickman L. Psychological impair- 31. McFarlane AC. The aetiology of post-traumatic bombing of the Murrah Building. J Pers Interpers Loss. ment in the wake of disaster: the disaster- morbidity: predisposing, precipitating and perpetuat- 1998;3:99-115. psychopathology relationship. Psychol Bull. 1991; ing factors. Br J Psychiatry. 1989;154:221-228. 8. Mallonee S, Shariat S, Stennies G, Waxweiler R, 109:384-399. 32. Ramsay R. Post-traumatic stress disorder: a Hogan D, Jordan F. Physical injuries and fatalities re- 20. Smith EM, North CS, McCool RE, Shea JM. Acute new clinical entity? J Psychosom Res. 1990;34:355- sulting from the Oklahoma City bombing. JAMA. postdisaster psychiatric disorders: identification of per- 365. 1996;276:382-387. sons at risk. Am J Psychiatry. 1990;147:202-206. 33. Kessler RC, McGonagle KA, Zhao S, et al. Life- 9. 1995 Oklahoma City Bombing Injuries Database. 21. North CS, Smith EM, McCool RE, Shea JM. Short- time and 12-month prevalence of DSM-III-R psychi- Oklahoma City, Okla: Oklahoma Department of term psychopathology in eyewitnesses to mass mur- atric disorders in the United States. Arch Gen Psychia- Health; 1999. der. Hosp Commun Psychiatry. 1989;40:1293- try. 1994;51:8-19. 10. American Psychiatric Association. Diagnostic and 1295. 34. North CS, Smith EM, Spitznagel EL. One-year fol- Statistical Manual of Mental Disorders, Third Re- 22. North CS, Smith EM, McCool RE, Lightcap PE. low-up of survivors of a mass shooting. Am J Psy- vised Edition. Washington, DC: American Psychiat- Acute post-disaster coping and adjustment. J Trauma chiatry. 1997;154:1696-1702. ric Association; 1987:247-251. Stress. 1989;2:353-360. 35. American Psychiatric Associatin. Diagnostic and 11. Robins LN, Helzer JE, Croughan J, Williams JBW, 23. North CS, Smith EM, Spitznagel EL. Posttrau- Statistical Manual of Mental Disorders, Fourth Edi- Spitzer RL. NIMH Diagnostic Interview Schedule. Ver- matic stress disorder in survivors of a mass shooting. tion. Washington, DC: American Psychiatric Associa- sion 3. Bethesda, Md: National Institute of Mental Am J Psychiatry. 1994;151:82-88. tin; 1994:427-429. Health; 1981. 24. McMillen JC, North CS, Smith EM. What parts of 36. Kessler RC. PTSD in a national survey of the United 12. Robins LN, Smith EM. The Diagnostic Interview PTSD are normal: intrusion, avoidance, or arousal? data States: results and future directions. In: Programs and Schedule/Disaster Supplement. St Louis, Mo: Wash- from the Northridge, Calif, earthquake. J Trauma Stress. abstracts of the Annual Meeting of the International ington University School of Medicine; 1983. In press. Society for Traumatic Stress Studies; November 10, 13. Weisæth L. Post-traumatic stress disorder after an 25. Robins LN, Fishbach RL, Smith EM, et al. Impact 1997. Montreal, Quebec.

Downloaded From: https://jamanetwork.com/ on 10/02/2021