A National Survey of Stress Reactions After the Sept 11 2001 Terrorist Attacks | Correlation And Dependence | Psychological Trauma

Please download to get full document.

View again

of 6
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
Document Share
Document Tags
Document Transcript
  SPECIAL REPORT N Engl J Med, Vol. 345, No. 20 · November 15, 2001 ·  www.nejm.org · 1507 Special Report  A NATIONAL SURVEY OF STRESS REACTIONS AFTER THE SEPTEMBER 11, 2001,TERRORIST ATTACKS M ARK A. S CHUSTER , M.D., P H .D., B RADLEY D. S TEIN , M.D., M.P.H., L ISA H. J AYCOX , P H .D., R EBECCA L. C OLLINS , P H .D.,G RANT N. M ARSHALL , P H .D., M ARC N. E LLIOTT , P H .D., A NNIE J. Z HOU , M.S., D AVID E. K ANOUSE , P H .D.,J ANINA L. M ORRISON , A.B., AND S ANDRA H. B ERRY , M.A. A BSTRACT Background  People who are not present at a trau-matic event may experience stress reactions. We as-sessed the immediate mental health effects of theterrorist attacks on September 11, 2001. Methods  Using random-digit dialing three to fivedays after September 11, we interviewed a nationallyrepresentative sample of 560 U.S. adults about theirreactions to the terrorist attacks and their perceptionsof their children’s reactions. Results  Forty-four percent of the adults reportedone or more substantial symptoms of stress; 90 per-cent had one or more symptoms to at least some de-gree. Respondents throughout the country reportedstress symptoms. They coped by talking with others(98 percent), turning to religion (90 percent), partici-pating in group activities (60 percent), and makingdonations (36 percent). Eighty-four percent of par-ents reported that they or other adults in the house-hold had talked to their children about the attacks foran hour or more; 34 percent restricted their chil-dren’s television viewing. Thirty-five percent of chil-dren had one or more stress symptoms, and 47 per-cent were worried about their own safety or the safetyof loved ones. Conclusions  After the September 11 terrorist at-tacks, Americans across the country, including chil-dren, had substantial symptoms of stress. Even clini-cians who practice in regions that are far from therecent attacks should be prepared to assist peoplewith trauma-related symptoms of stress. (N Engl JMed 2001;345:1507-12.) Copyright © 2001 Massachusetts Medical Society. HE terrorist attacks against the United Stateson September 11, 2001, shook the nation.Television coverage was immediate, graphic,and pervasive. 1-3 Newscasts included remarkable vid-eo footage showing two airplanes crashing into the World Trade Center and the aftermath of four air-plane crashes. 2,3 People who are present at a traumaticevent often have symptoms of stress, but there is evi-dence that adults and children need not be present tohave stress symptoms, 4-6 especially if they considerthemselves similar to the victims. 4 The events on Sep- T tember 11 were widely described as attacks on Amer-ica, and most or all Americans may have identified with the victims or perceived the attacks as directedat themselves as well.The immediate mental health effects of a nationalcatastrophe experienced from afar — especially onethat carries the threat of further attacks — have rarely been examined. We surveyed a nationally represen-tative U.S. sample to determine the immediate reac-tions of adults to the attacks and their perceptionsof their children’s reactions. METHODS Data Collection  We used random-digit dialing within the United States. The in-terview period was three to five days after the attack — from Friday evening, September 14, at the end of the national day of mourningdeclared by President George W. Bush, through Sunday evening,September 16, just before the start of the workweek, when the pres-ident encouraged Americans to return to their normal activities. 7 Trained interviewers conducted computer-assisted telephone in-terviews in English; the median duration of the interviews was 28minutes. RAND’s institutional review board approved the study procedures. Sample  Adults (persons 19 years of age or older) who were at home when we called were eligible for the study; if two or more adults were athome, we randomly selected one to interview. We spoke with a totalof 768 selected adults. Of these persons, 73 percent (560) were in-terviewed, 24 percent refused to be interviewed, and 3 percentagreed to be interviewed later in the weekend but the interview didnot take place. Because of the extremely short time for this survey, we could not establish how many of the 3505 telephone numbers we called might eventually have yielded an eligible person or beenestablished as ineligible. At the end of the interview period, 683 tele-phone numbers were determined to be nonworking or businessnumbers; 182 were cell phones, pagers, fax machines, or other suchineligible numbers; 495 were unanswered after several attempts. As compared with the U.S. population represented in the March2001 Current Population Survey, 8 our sample slightly overrepre-sented women, non-Hispanic whites, and persons with higher lev-els of education and household income, which is typical of sam- From RAND, Santa Monica, Calif. (M.A.S., B.D.S., L.H.J., R.L.C.,G.N.M., M.N.E., A.J.Z., D.E.K., S.H.B.); and the Departments of Pediat-rics (M.A.S., J.L.M.), Health Services (M.A.S.), and Medicine (M.N.E.),University of California, Los Angeles. Address reprint requests to Dr.Schuster at RAND, 1700 Main St., Santa Monica, CA 90407-2138, or atschuster@rand.org. The New England Journal of MedicineDownloaded from nejm.org at UC SHARED JOURNAL COLLECTION on April 11, 2012. For personal use only. No other uses without permission.Copyright © 2001 Massachusetts Medical Society. All rights reserved.  1508 · N Engl J Med, Vol. 345, No. 20 · November 15, 2001 ·  www.nejm.org The New England Journal of Medicine ples selected by means of random-digit dialing. 9,10 As a sensitivity analysis, we repeated all analyses after weighting the sample to re-semble the population estimates from the Current Population Sur- vey, which neither reduced the total sampling error nor substantially altered the results.Respondents living with a child 5 to 18 years old were askedquestions about the child (or about a randomly selected child if there were two or more children at home); information was ob-tained for a total of 170 children. Although we did not ask  whether the respondent was the child’s parent, we use the term“parent” because data from the Current Population Survey suggestthat most adults in households with children are their parents. Instrument and Key Measures To assess reactions to the September 11 attacks, we selected anddeveloped questionnaire items on the basis of prior research andcurrent media reports. Except as otherwise noted, the questionsspecified a time frame of “since Tuesday”; questions about televi-sion viewing specified “on Tuesday.” To assess exposure to the at-tacks through television viewing, we asked respondents the amountof time (in hours, or in minutes if less than one hour) on Sep-tember 11 that they and their children watched television cover-age of the attacks.To assess stress in adults, we modified 5 questions about symp-toms from the 17-question Posttraumatic Stress Disorder Checklist 11 (Table 1). The symptoms were selected from those reported by 50 percent or more of the survivors of the Oklahoma City bomb-ing. 12 For the analysis, we defined a substantial stress symptom asone of the two highest of the five response options 13 (“quite a bit”or “extremely”). A substantial stress reaction was defined as one ormore substantial stress symptoms. For children, we modified fiveitems from the Diagnostic Interview Schedule for Children, VersionIV (parent’s version) 14 (Table 1). A stress reaction was defined asan affirmative response to at least one of the items.To determine the distance of the respondents from all threecrash sites, as well as from the takeoff and destination sites of theflights, we performed a geographic information system analysis,coding the location as the longitude and latitude for the centerof the ZIP Code area (or of the telephone-exchange area for the8 percent of respondents who provided no ZIP Code). We assessedthe relation between stress in adults and the distances from indi- vidual sites, as well as the relation between stress and the distancefrom the nearest crash site and from the nearest of any of the sites.The strongest association was with the distance from the WorldTrade Center. Therefore, that is the association we report in thisarticle. We also examined population density, a characteristic of lo-cation that we believed might be associated with differences in theperceived risk of terrorism and with reported stress. Statistical Analysis  We report the results of univariate analyses (means and percent-ages) and bivariate analyses (Pearson’s and Spearman’s tests of correlation, t-tests, and chi-square tests of homogeneity). Whereapplicable, transformations of variables were used to satisfy the as-sumptions of these tests. Data have been weighted to account formultiple telephone lines in a household; our question about thenumber of telephone lines did not exclude inactive and data-transfer lines, so the results of significance tests may be conserva-tive. We used the linearization method to estimate standard errorsand to correct statistical tests for weights. 15 The 95 percent sam-pling error for reported percentages was no more than 4.3 per-centage points for adults and no more than 7.7 percentage pointsfor children. No imputation of missing values was performed. RESULTS Adults Forty-four percent of the U.S. adults we surveyedreported at least one of five substantial stress symp- *For adults, substantial stress was defined as an answer of “quite a bit” or “extremely” on a five-point scale (“not at all,” “a little bit,” “moderately,” “quite a bit,” and “extremely”). For children,stress was defined as an answer of “yes” on a two-point scale (“yes,” “no”).†Respondents who answered some but not all of the questions about stress are included. T ABLE 1.  A  DULTS    WITH S UBSTANTIAL  S TRESS S  YMPTOMS    AND C HILDREN    WITH S TRESS  S  YMPTOMS    AND W  ORRIES . Q UESTION N O . OF R ESPONDENTS S UBSTANTIAL S TRESS * %  Adults Since Tuesday, have you been bothered by:Feeling very upset when something reminds you of what happened?Repeated, disturbing memories, thoughts, or dreams about whathappened?Having difficulty concentrating?Trouble falling or staying asleep?Feeling irritable or having angry outbursts? At least one of the above†55455755855555856030161411944 Children Since Tuesday, has your child been: Avoiding talking or hearing about what happened?Having trouble keeping his or her mind on things and concen-trating?Having trouble falling asleep or staying asleep?Losing his or her temper or being irritable?Having nightmares? At least one of the above16716716716716716718121010635Since Tuesday, has your child been worrying about his or her safety orthe safety of loved ones?16747 The New England Journal of MedicineDownloaded from nejm.org at UC SHARED JOURNAL COLLECTION on April 11, 2012. For personal use only. No other uses without permission.Copyright © 2001 Massachusetts Medical Society. All rights reserved.  SPECIAL REPORT N Engl J Med, Vol. 345, No. 20 · November 15, 2001 ·  www.nejm.org · 1509 toms since September 11, 2001 (Table 1); 68 per-cent experienced at least one symptom “moderately”and 90 percent experienced at least one symptom “alittle bit.” Stress reactions varied significantly accord-ing to sex, race or ethnic group, presence or absenceof prior emotional or mental health problems, dis-tance from the World Trade Center, and region of the country (Table 2).On September 11, adult respondents watched tele- vision coverage of the attacks for a mean of 8.1 hours;2 percent of respondents watched for less than 1 hour,15 percent for 1 to 3 hours, 34 percent for 4 to7 hours, 31 percent for 8 to 12 hours, and 18 percentfor 13 hours or more. Extensive television viewing was associated with a substantial stress reaction (Ta-ble 2). Adults responded to the attacks in various ways(Table 3). People with a substantial stress reaction were more likely than others to have talked at least“a medium amount” about their feelings (91 percent *For adults, substantial stress was defined as an answer of “quite a bit” or “extremely” to one or more of five questionsabout stress on a five-point scale (“not at all,” “a little bit,” “moderately,” “quite a bit,” and “extremely”). For children,stress was defined as an answer of “yes” to one or more of five questions about stress. P values were calculated with theuse of Spearman’s tests of correlation for ordered categories (population density, miles from World Trade Center, andhours of television viewing) and chi-square tests of homogeneity for unordered categories. Variables not included in thetable were not significant at the P<0.05 level (respondent’s age, respondent’s level of education, number of children 5 to18 years old, household income, sex of child, child’s age, and number of hours of television viewing by child on Septem-ber 11 about the attacks). To convert miles to kilometers, multiply by 1.609344; to convert square miles to square kilo-meters, multiply by 2.589988.†In a multivariate model that included all characteristics with significant bivariate associations (P<0.10), all variablesother than region and prior emotional or mental health problems were significantly associated with adult stress (P<0.05).‡Respondents were asked whether they had needed help for emotional or mental health problems, such as feeling sad,blue, anxious, or nervous, during the 12 months before the attacks. 16 §Regions were defined according to U.S. Census regions.¶The percentages were almost identical when adults who avoided television and other reminders of the attack wereomitted from the analysis (P=0.003). T ABLE 2. S TRESS R  EACTIONS A  CCORDING   TO   THE C HARACTERISTICS   OF   THE R  ESPONDENTS .* C HARACTERISTIC   OF R ESPONDENT A DULTS C HILDREN NO . OFRESPONDENTSSUBSTANTIAL STRESS   REACTION P  VALUE † NO . OFRESPONDENTSSTRESS   REACTION P  VALUE %% Total5604416735SexFemaleMale29822650370.006966441250.05Race or ethnic group White (non-Hispanic)Nonwhite4131064162<0.0011253533410.41Prior emotional or mental health problems‡ YesNo6648956420.051615053320.10Distance from World Trade Center«100 mi101–1000 mi»1001 mi44274242614836<0.0011393615226440.23Region§NortheastSouthMidwest West93169154144554642360.0530544835433025470.17Population density «100 persons/mi 2 101–300 persons/mi 2 301–1000 persons/mi 2 1001–2000 persons/mi 2 »2001 persons/mi 2 1221071441058239483847520.17353043302926383334460.21Hours of television viewing on September 11about the attacks¶0–3 hr4–7 hr8–12 hr»13 hr94185175102373946580.001———— The New England Journal of MedicineDownloaded from nejm.org at UC SHARED JOURNAL COLLECTION on April 11, 2012. For personal use only. No other uses without permission.Copyright © 2001 Massachusetts Medical Society. All rights reserved.  1510 · N Engl J Med, Vol. 345, No. 20 · November 15, 2001 ·  www.nejm.org The New England Journal of Medicine  vs. 83 percent, P=0.008), turned to religion (84 per-cent vs. 69 percent, P<0.001), made donations (42percent vs. 31 percent, P=0.01), and checked onthe safety of family members and friends (83 percent vs. 69 percent, P<0.001).Thirty-six percent of adults thought that terrorism was a “very serious” or “somewhat serious” problemin the area where they live and work. Forty-four per-cent thought terrorism would increase over the nextfive years, and 21 percent thought it would remainat the current level. Children Thirty-five percent of parents reported that theirchildren had at least one of five stress symptoms; 47percent reported that their children had been wor-rying about their own safety or the safety of lovedones (Table 1). Parents with a substantial stress reac-tion were more likely than others to report that theirchildren had symptoms of stress (50 percent vs. 22percent, P< 0.001).Children watched television coverage of the attacksfor a mean of 3.0 hours on September 11; 8 percentdid not watch any of the coverage, 33 percent watchedfor 1 hour or less, 36 percent watched for 2 to4 hours, and 23 percent watched for 5 hours ormore. Older children watched more (Pearson’s r=0.52, P<0.001); for example, 73 percent of children who were 5 to 8 years old watched for one hour orless, whereas 51 percent of those who were 17 or 18 years old watched for five hours or more. Thirty-four percent of parents tried to restrict (limit or pre- vent) their children’s viewing of the televised cover-age of the attacks; in this subgroup, the children watched an average of 2.3 hours of coverage, ascompared with 3.4 hours for other children (P=0.005). Parents were more likely to try to limit tele- vision viewing by younger children than by olderchildren (Spearman’s r=0.39, P<0.001). Parents who reported that their children were stressed weremore likely than others to restrict their children’stelevision viewing (45 percent vs. 29 percent, P=0.05); among children whose parents did not try torestrict television viewing, there was an associationbetween the number of hours of television viewingand the number of reported stress symptoms (Pear-son’s r=0.27, P=0.02). The response to the ques-tion about whether the child worried about his orher safety or the safety of others was not significantly associated with whether parents tried to restrict tele- vision viewing or with the number of hours of tele- vision viewing by children whose parents did not try to restrict viewing.One percent of parents reported that they (or oth-er adults in the household) did not speak with theirchildren about the attacks; 15 percent discussed theattacks for less than one hour, 48 percent for one tothree hours, 22 percent for four to eight hours, and14 percent for nine hours or more. The number of hours of discussion was higher for older childrenthan for younger children (Pearson’s r=0.27, P=0.001) and was associated with the number of hoursof television viewing (Pearson’s r=0.40, P<0.001).There was no significant association between the ex-tent of communication and the degree of stress symp-toms on the part of parents or children. DISCUSSION  A few days after the September 11 terrorist at-tacks, 44 percent of a nationally representative sam-ple of adults reported that they had had at least oneof five substantial stress symptoms since the attacks,and 90 percent reported at least low levels of stresssymptoms. Children also experienced stress: 35 per-cent had at least one of five stress symptoms after theattacks. Although the rates of stress reactions werehighest among subgroups previously found to have *Each question referred to the interval between September 11 and thedate of the interview (September 14, 15, or 16). Because of rounding, notall percentages total 100. T ABLE 3. C OPING B EHAVIOR     AND O THER  R  EACTIONS   BY  A  DULTS .* Q UESTION T OTAL N O . OF R ESPONDENTS R ESPONSE NOT AT    ALL  A    LITTLEBIT A    MEDIUM AMOUNT A LOT percent How much have you talked with someone about yourthoughts and feelingsabout what happened?5562123057How much have you turnedto prayer, religion, orspiritual feelings?55610153144How much have you partici-pated in a public or groupactivity in recognition of  what happened?55940262311How much have you avoidedactivities such as watchingTV because they remind you of what happened?5556120145  YES percent Have you donated blood ormoney or done any volun-teer work?55936Have you gotten any extrafood, gas, cash, or othersupplies you might need?55718Have you checked the safety of immediate family mem-bers and friends?55675Have you checked on some-one you thought might behurt or missing?55632 The New England Journal of MedicineDownloaded from nejm.org at UC SHARED JOURNAL COLLECTION on April 11, 2012. For personal use only. No other uses without permission.Copyright © 2001 Massachusetts Medical Society. All rights reserved.
Similar documents
Search Related
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks