9-11

Thoughts on Communal Bereavement
Following the World Trade Center Tragedy

Winter, 2002
by Terry Hartig, Institute for Housing and Urban Research, Uppsala University
and Ralph Catalano, School of Public Health, University of California, Berkeley

Within days of the stunningly effective attacks on the World Trade Center and Pentagon, John Hughes, the editor of the Journal of Health and Social Behavior, contacted us to ask if we wanted to add some text to an article just going into production (Catalano & Hartig, 2001). In the article, we had reported an increased incidence of very low birthweight (VLBW) in the months following the murder of Swedish Prime Minister Olof Palme in 1986. We also found a smaller though still significant increase in the months after the ferry M/S Estonia sank during a storm on the Baltic in 1994. The two events marked instances of what we referred to as communal bereavement, “the widespread experience of distress among persons who never met the deceased.” Like us, John Hughes had recognized the close parallels between what we had described in the article and what we saw unfolding in the aftermath of the attacks. Would we, he wanted to know, like to add some text to the article to flag the parallels?

The events in Sweden differed from each other in a fundamental respect: the number of lives lost. One person dead on a sidewalk would hardly seem comparable to 501 Swedes and hundreds of others entombed in a sunken ship. Why should these seemingly so dissimilar events have similar effects on health? In formulating an account, we departed from an epidemiologic literature that links somatic illness with the “deep and poignant” distress of personal bereavement (e.g., Clayton, 2000). As with other major stressors, personal bereavement appears to increase corticosteroid production and compromise immune function (Irwin, Daniels, & Weiner, 1987; Zisook, Shuchter, Irwin, & Darko, 1994). We proposed that in communal bereavement distress arises not so much from the permanent and possibly irreplaceable loss of tangible and intangible benefits enjoyed from a close personal relationship, but rather more from a sense that institutions essential to the normal functioning of the community had failed.

The loss of a single person who played a highly visible and difficult to fill role in an important institution could distress the community because the institution might not function as needed for an unknown length of time. The circumstances of that person’s death might aggravate distress by arousing concerns about the identity of responsible parties and the extent to which they may continue to present a threat. Similar concerns may arise when the death or deaths follow from a catastrophic failure of trusted institutions to perform essential tasks, such as ensuring the secure operation of transportation systems or other components of a society’s infrastructure (cf. Baum et al., 1992).

Moreover, we thought, mutual aid arrangements would suffer under the circumstances of communal bereavement. The number of persons experiencing distress would be relatively large, implying that unusually many members of social networks would need physical and psychological resources for coping. "Surplus" support would be lacking for the vulnerable. The fraction of distressed persons who become ill would be higher than at other times.

The attack on the WTC was beyond our imagination when we speculated that people who had never met the victim(s) could nonetheless experience deep and poignant distress when facing terrible evidence that institutions essential to the normal functioning of the community had failed. In doing our research, however, we were aware that events like the Palme murder and loss of the Estonia could - and probably would - occur again, and that they could have implications for the public’s health.

To test our hypothesis, we obtained and analyzed aggregate data for the quarterly incidence of VLBW births in Sweden during the years 1973 to 1995. Why did we use the incidence of VLBW as our outcome? For one, pregnant women adapting to stressors have an elevated risk of delivering very low weight infants (e.g., Hedegaard et al., 1996; Hobel et al., 1999; Lockwood, 1999). Mothers who have such babies - they weigh less than 1500 grams, or 3.5 pounds - almost always deliver them preterm (cf. Crouse & Cassidy, 1994; Solis, Pullum, & Frisbie, 2000). It appears that the timing of delivery is sensitive to elevated corticosteroid levels and compromised immune function, two stress reactions shown to attend personal bereavement, as we mentioned earlier. Beyond the evidence of a plausible biological mechanism, we had other reasons for focusing on VLBW. A VLBW birth often involves significant and persistent suffering and economic costs, for the child, the family and society, in part because of developmental deficiencies (e.g., Wise, Wampler, & Barfield, 1995). Also, we had previously used the incidence of VLBW to test the communal health effects of unemployment among males (Catalano, Hansen, & Hartig 1999).

In our interrupted time-series analysis of the VLBW data, we compared the incidence of VLBW in the quarters after the Palme murder and the Estonia catastrophe to the incidence that we would have expected after removing autocorrelation due to trends, seasonality, and other sources. After applying still other statistical controls, we allowed ourselves a good measure of confidence in the effects that we uncovered: about 35 excess VLBW births in the months following the Palme murder (a 21% increase over the expected value of 166), and about 25 more VLBW births than expected (a 15% increase) in the months following the Estonia catastrophe. It appears that, in the aftermath of each of these two events, some Swedish women prematurely concluded their pregnancies, delivering a very low weight baby.

When we have told colleagues what we found, we have sometimes noticed a skeptical raising of eyebrows. All well and good: we have a professional responsibility to regard research results with skepticism. We then must point out to them that rival explanations must fit within the constraints imposed by our statistical controls while at the same time positing an event that could coincide with the assassination of Olof Palme or the sinking of the Estonia. We could not think of such an alternative explanation.

Taking up John Hughes’ offer, we added some text to our article expressing our view that the September 11 attacks qualified as an instance of communal bereavement. The journal’s publisher then promptly sent out a press release, which the major news services just as promptly picked up. Their interest put us in an uncomfortable position. We greatly appreciated the recognition given the work and the interest in the potential practical value of our findings, but the circumstances that called so much attention to it crowded out a sense of satisfaction that we might otherwise have enjoyed. Further complicating our personal reactions, we were aware that the reporting of our results fed into a media construction of communal bereavement. Toward the end of our paper, we had written the following:

“We have assumed that the media precipitate communal bereavement by alerting the community to events that genuinely induce a feeling of distress in the population. The way the media describe an event, however, may affect how many people are distressed and how distressed they feel. If this were the case, it would suggest that the media may be acting irresponsibly by reporting deaths in sensational ways simply to attract readers or viewers. If additional research replicates findings such as ours, discussions of journalistic ethics might expand to include the health effects of sensational reporting of deaths.”

To their credit, the journalists with whom we had contact were aware of this issue, just as they were aware that our findings had possible preventive utility that deserved wider consideration. Still, we had a sense that the findings had been hijacked by the recent events.

This was not an idle concern. In a November 15 article in the New England Journal of Medicine, researchers from RAND and UCLA reported on post-attack stress reactions registered in a large telephone survey of adult Americans (Schuster et al., 2001). Their hypothesis resembled ours: “people who are not present at a traumatic event may experience stress reactions” (p. 1507). Of the stress reactions studied, such as difficulty concentrating and intrusive thoughts, 44% of the people surveyed reported that they experienced at least one to a substantial degree (i.e., “quite a bit” or “extremely”). The researchers found few studies of traumarelated stress symptoms in community-based samples of Americans not suffering from a psychiatric disorder, and those that they did find differed methodologically from their own study. Still, they concluded that their study indicated a much higher prevalence of event-related stress symptoms than shown previously. Relevant to our formulation of the phenomenon as one of a community rather than individual impact, the prevalence of substantial reactions varied with distance from the WTC, a variable we might view as a crude proxy for the strength of an individual’s identification or emotional involvement with the community affected by the attacks. Of those living within 100 miles of the WTC, 61% reported some substantial stress reaction, in contrast to 36% of respondents living more than 1001 miles away. Relevant to our concern about the media’s role in feeding communal bereavement, the researchers described a positive association between the prevalence of substantial stress reactions and the number of hours spent on September 11 watching televised accounts of the attacks. Certainly, we should ask questions about these findings. Did those who watched more TV have a family member or close friend in or near the WTC or Pentagon? What kind of person watches more than 13 hours of television in a day – even a day such as September 11? Although we cannot say with certainty that characteristics of the televised account explain this association, we cannot get away from a basic fact; in such an event, turning to the media for information is for many people one of only a few available coping strategies. They deserve careful treatment.

In the coming months and years we expect to see a substantial amount of attention given to stress-related health outcomes in the aftermath of the September 11 attacks. Seeing our results replicated will give us little joy.

References

Baum, A., Fleming, I., Israel, A., & O'Keefe, M. (1992). Symptoms of chronic stress following a natural disaster and discovery of a human-made hazard. Environment and Behavior, 24, 347-365.

Catalano, R., Hansen, H-T., & Hartig, T. (1999). The ecological effect of unemployment on the incidence of very low birthweight in Norway and Sweden. Journal of Health and Social Behavior, 40, 422-428.

Catalano, R., & Hartig, T. (2001). Communal bereavement and the incidence of very low birthweight in Sweden. Journal of Health and Social Behavior, 42.

Clayton, P. (2000). Bereavement. In G. Fink (Ed.), Encyclopedia of stress (Vol. 1; Pp 304-311). London: Academic Press.

Crouse, D., & Cassady, G. (1994). The small-forgestational- age infant. In G. Avery, M. Fletcher, & M. MacDonald (Eds.), Neonatology: Pathophysiology and management of the newborn (4th ed.; pp. 369-387). Philadelphia, PA: Lippincott. Hedegaard, M., Henriksen, T., Secher, N., Hatch, M., & Sabroe, S. (1996). Do stressful life events affect duration of gestation and risk of preterm delivery? Epidemiology, 7, 339- 345.

Hobel, C., Dunkel-Schetter, C., Roesch, S., Castro, L., & Arora, C. (1999). Maternal plasma corticotropin-releasing hormone associated with stress at 20 weeks’ gestation in pregnancies ending in preterm delivery. American Journal of Obsteterics and Gynecology, 180, 257-264.

Irwin, M., Daniels, M., & Weiner, H. (1987). Immune and neuroendocrine changes during bereavement. Psychiatric Clinics of North America, 10, 449-465.

Lockwood, C. (1999). Stress-associated preterm delivery: The role of corticotropin-releasing hormone. American Journal of Obstetrics and Gynecology, 180, 264-266. Solis, P., Pullum, S. G., & Frisbie, W. P. (2000). Demographic models of birth outcomes and infant mortality: An alternative measurement approach. Demography, 37, 489- 498.

Schuster, M. A., Stein, B. D., Jaycox, L. H., Collins, R. L., Marshall, G. N., Elliott, M. N., Zhou, A. J., Kanouse, D. E.,

Morrison, J. L., & Berry, S. H. (2001). A national survey of stress reactions after the September 11, 2001, terrorist attacks. New England Journal of Medicine, 345, 1507-1512.

Wise, P., Wampler, N., & Barfield, W. (1995). The importance of extreme prematurity and low birthweight to United States neonatal mortality patterns: Implications for prenatal care and women’s health. Journal of the American Medical Women's Association, 50, 152-55.

Zisook, S., Shuchter, S., Irwin, M., & Darko, D. (1994). Bereavement, depression, and immune function. Psychiatry Research, 52, 1-10.