Tropical cyclones rank among the top natural disasters with respect to human fatalities.1,2 Few studies have epidemiologically assessed cyclone-related injuries and fatalities in developing countries, despite the fact that developing countries have borne the heaviest burden in injuries and fatalities.2
On 10 August 2006, Typhoon Saomei struck southern China's Fujian and Zhejiang Provinces. One weather station in Fujian Province observed wind gust at 75.8 m/s, breaking the records in both provinces, and nearly reaching the all-time typhoon record of 78.9 m/s observed in Hong Kong on 1 September 1962.3 We evaluated the severity and types of injuries due to Saomei, and their related risk factors.
METHODS
We conducted our study in Longhua Village, a coastal village in Fujian Province, Southeast China (27.189N, 120.363E) (Fig. 1). This village was directly in Saomei's path, and was declared to be a severe disaster area. Of the village's 3000 residents, about half did not have a permanent residency permit but lived in similar and intermingling housing units with the regular residents.
FIGURE 1.:
Geographic location (in relation to China) and sketch map (drawn based on Google map) of Longhua Village (27.189N, 120.363E), Fujian Province, China.
We defined a typhoon-related injury case as a Longhua Village resident who sustained at least one of the following types of injuries between 12:00 noon on 10 August and 12:00 noon on 11 August 2006, based on the ICD-10 codes for unintentional injuries4: cut/pierce (W25–W29, W45), drowning (W65–W74, V90, V92), fall/jump/push (W00–W19), all transport (V01–V89, V91, V93–V99), struck by or against (W20–W22, W50–W52), and suffocation (W75–W84).
We conducted a complete census of all Longhua Village residents by door-to-door visits. We collected information on deaths through proxy interviews of family members. We classified injuries into 5 severity categories: minor (not requiring a doctor's visit), intermediate (needing treatment and missing up to 1 day of work), severe (missing more than 1 day of work or being hospitalized for up to 9 days), very severe (needing a surgery or being hospitalized for more than 9 days), and disabling (sustaining permanent disabilities).
We then conducted a 1:2 matched case-control study to evaluate risk factors for injury. Cases were the 98 persons with more than minor injuries who agreed to participate in the study. Two controls per case (n = 196) were selected among uninjured neighbors, matched on row of the house plus sex and age. We collected information regarding personal, familial, and societal risk factors for injury. We used conditional logistic regression to control for confounding, accounting for the 1:2 matched study design.5,6 We used SPSS (version 13.0, SPSS Inc., Chicago, IL) and EpiInfo (version 3.5.1, CDC, Atlanta, GA) for data analysis.
RESULTS
Saomei made landfall at 17:25 local time (9:25 GMT) on 10 August 2006. The wind speed peaked twice, at approximately 18:00 and 20:00 (Fig. 2). Of the 3108 Longhua Village residents, 3002 (97%; 1543 male, 1459 female) participated in the census. We identified 143 injured persons, of whom 136 met the case definition for typhoon-related injury (injury rate: 4.5%). Seven persons died, all fishermen (death rate: 2.3 per 1000); 6 of these deaths occurred when the fishing vessels capsized.
FIGURE 2.:
Epidemic curve of injury (black bars) in relation to rainfall and wind speed during Typhoon Saomei between 12:00 noon 10 August 2006 and 12:00 noon 11 August 2006.
The first injuries occurred at 16:00 local time, and peaked at 18:00 with the first peak of wind speed. Of the 129 surviving cases, 47% were injured during the initial 30–90-minute window after Saomei made landfall. The majority of all injuries (87%) were caused by flying sharp objects such as window glass and doors, or falling blunt objects such as bricks. The most common circumstance for injury was during an attempt to close a door or window (40%). Intermediate to severe injuries accounted for 73% of all cases (Table 1). Men, persons aged 40–69 years, fishermen, and persons with fewer years of schooling were at greater risk for injury (Table 2). In this village, houses are lined up uniformly in 6 rows, with the first row directly facing the sea (Fig. 1). The injury rate among dwellers of first-row houses was 18 times as high as dwellers of the last row (Table 2).
TABLE 1: Main Causes, Circumstances, and Severity of Injuries (n = 129) Caused by Typhoon Saomei: Longhua Village, Fujian Province, China, 2006
TABLE 2: Injury Attack Rate due to Typhoon Saomei, by Age, Sex, Occupation, Education, and Location of House (Based on Census Data)
Of the 109 persons with more than minor injuries, 98 (90%) agreed to participate in the case-control study. Univariate analysis showed that failure to receive early warnings before the storm, failure to reinforce doors or windows, failure to take protective measures, staying in a windowed room, staying near a window or door, staying in a damaged room, and staying outside of home during the storm were all risk factors for injury. In the conditional logistic regression analysis, failure to reinforce doors or windows prior to the storm, staying near a window or door, and staying in a damaged room remained as important risk factors after controlling for the confounding effects of each other (Table 3).
TABLE 3: Risk Factors for Injury Based on a Matched Case-Control Study
DISCUSSION
Typhoon Somei, with wind velocity nearly reaching the all-time record, caused high rates of injury and death in Longhua Village. Injury due to Saomei was associated with wind speed, location of the residence, and inadequate preventive measures. The death rate (2.3 per 1000) was 530 times as high as that in the US due to Hurricane Andrew (0.0044 per 1000), during which most of the injuries were also caused by high winds.7 However these 2 rates are not entirely comparable due to many differences in the 2 events.
Injuries that occur during tropical cyclones have been well-documented in developed countries.1,2 However, despite China's being one of the worst-affected countries in the world,1 we are aware of only one previously published Chinese study on typhoon-related injuries.8 In that study, risk was increased among those staying outside of home and those who had not received a warning. We identified several preventable risk factors for injury, including failure to reinforce doors and windows before the storm, staying in a damaged room, and staying near a window or door during the storm. These findings may help guide interventions in China and other developing countries.
Effective warning and timely evacuation have been key in reducing injury morbidity and mortality from tropical cyclones in developed countries.1,2 In Longhua Village, a typhoon warning was issued before Saomei made landfall. However, similar to the “overwarning fatigue” phenomenon observed in developed countries,2 most of the Longhua Village residents did not take the warning seriously. There had been warnings in the past for storms that turned out not to be severe. Also, few families owned vehicles, making it difficult to evacuate.
All 7 fatalities were fishermen; 6 were drowning deaths. This finding was similar to those observed elsewhere.1 We were told by the decedents’ family members that during a typhoon, fishermen routinely run their fishing vessel engines at full speed against the wind. However, because Saomei's high wind velocity, this strategy failed to prevent the fishing vessels from capsizing.
During a typhoon, there can be a relatively calm period as typhoon's eye passes through. The duration of this period depends on the tightness of the eye and the typhoon's speed. Past experience led Longhua Village residents to believe that the calm period would last for at least an hour; therefore, many residents went out to check on damages to their properties. However, Saomei's calm period lasted for less than an hour and many residents were caught by surprise. Nearly half of the injured cases occurred during the 30–90-minute window after Saomei made landfall, coinciding with the peaks of the storm just before and after the eye.
Nearly 60% of the injury cases lived in the first row of houses directly facing the sea; the injury rate among residents of those houses was markedly higher than residents of the houses behind. This is probably because houses close to the sea effectively served as a shield for the houses behind. This finding suggests that reinforcing houses directly on the seashore to withstand typhoons may help reduce the damage caused by typhoons.
In a previous study, persons aged ≥60 years had the highest risk of injury.9 In our study persons aged 40–69 years had the highest injury rate. Presumably, this age group was most affected in Longhua Village because these people were household heads who took the responsibility to close the windows and doors during the storm and to check on property damages during the brief calm period, while neglecting their own safety.
Our study was conducted in only one village. Also, we did not evaluate potential differences between the permanent and nonpermanent residents of the village. Therefore, caution should be used in interpreting our findings and generalizing them to other settings.
ACKNOWLEDGMENTS
We are indebted to staff at Fujian Provincial CDC and Fuding City CDC for their assistance during this investigation. We also thank James Mendlein and Robert Fontaine, US CDC, who provided many valuable suggestions and helped to revise this article.
REFERENCES
1. Shultz JM, Russell J, Espinel Z. Epidemiology of tropical cyclones: the dynamics of disaster, disease, and development.
Epidemiol Rev. 2005;27:21–35.
2. Noji EK.
The Public Health Consequences of Disasters. New York: Oxford University Press; 1997;9.
3. Duan Y. Typhoons Bilis and Saomai: Why the impacts were so severe. Available at:
http://www.wmo.ch/pages/mediacentre/bilis_en.html. Accessed January 16, 2008.
4. World Health Organization. International statistical classification of diseases and related health problems (10th revision, version for 2007). Available at:
http://www.who.int/classifications/apps/icd/icd10online/. Accessed January 28, 2007.
5. Schlesselman JJ, ed.
Case-Control Studies: Design, Conduct, Analysis. New York: Oxford University Press; 1982.
6. Hosmer DW Jr, Lemeshow S.
Applied Logistic Regression. 2nd ed. New York: John Wiley; 2000.
7. Combs DL, Parrish RG, McNabb SJ, et al. Deaths related to Hurricane Andrew in Florida and Louisiana, 1992.
Int J Epidemiol. 1996;25:537–544.
8. Gong ZY, Chai CL, Tu CY, et al. A field epidemiological study on the risk factors of injury caused by typhoon [in Chinese].
Zhonghua Liu Xing Bing Xue Za Zhi. 2006;27:773–776.
9. Eidson M, Lybarger JA, Parsons JE, Maccormack JN, Freeman JI. Risk factors for tornado injuries.
Int J Epidemiol. 1990;19:1051–1056.
