Hurricanes and tropical storms can result in injury and death, either directly or indirectly. Directly caused injuries include trauma, drowning, asphyxiation, and electrocution.Reference Issa, Ramadugu and Mulay1-Reference Ragan, Schulte and Nelson5 Indirectly caused injuries include exacerbation of medical conditions, infection, poisoning (eg, carbon monoxide, gasoline), bites and stings, vehicular crashes, and preparation or repair injuries (eg, falls, electrocution).Reference Issa, Ramadugu and Mulay1-Reference Sullivent, West and Noe14 Injuries may occur before the storm makes landfall.Reference Issa, Ramadugu and Mulay1 One study reported that 12% of deaths associated with Hurricane Irma were related to preparation and repair injuries.Reference Issa, Ramadugu and Mulay1
In hurricane-prone regions, hurricane or storm shutters provide protection for glass windows and doors against windborne debris that occur during hurricanes. If windows or doors break, in addition to the broken glass that may cause injuries, sudden pressurization of the interior of a building may result in structural or nonstructural damage. Although shutters will not eliminate the possibility of wind-driven rain entering a building, it will improve the building’s resistance and possibly reduce the risk of broken glass. A wide variety of hurricane shutter types are available and may be made of wood, metal, and polycarbonate. Shutters may be recommended or required in hurricane-prone areas.15
Injuries may occur when putting up and taking down hurricane shutters. When Hurricane Frances struck in 2004, doctors in Palm Beach County and Broward County, Florida, reported that the most common injuries were sustained by people trying to put up and take down hurricane shutters and plywood.Reference Pacenti16 Such injuries tend to not be life-threatening and consist of lacerations to hands, arms, and legs, including lacerations to tendons (because shutters may have sharp edges and people do not wear work gloves); falls from ladders; and back injuries.Reference Pacenti16
A review of PubMed and the Internet failed to identify any studies published in peer-review journals examining injuries associated with hurricane shutters. The objective of this study was to describe hurricane shutter-related injuries managed at emergency departments (EDs). In particular, a comparison was made between the temporal pattern of injuries and the landfall of selected hurricanes and tropical storms in the United States (US).
METHODS
The data source for this study was the National Electronic Injury Surveillance System (NEISS) operated by the US Consumer Product Safety Commission (CPSC). The NEISS collects data on consumer product-related injuries from the EDs of approximately 100 hospitals as a probabilistic sample of the more than 5000 US hospitals with EDs. The NEISS is a stratified sample based on ED size and geographic location. Information collected include the patient’s age, sex, race, ethnicity, injury diagnosis, body parts affected, and location where the injury occurred, among others.17 NEISS data are publicly available and de-identified; therefore, the study is exempt from institutional review board approval.
Cases were hurricane shutter-related injuries in the NEISS database that occurred during 2001-2017. The Narrative_1 and Narrative_2 text fields were searched for any mention of the terms “shutter” and “hurricane,” “storm,” “metal,” “aluminum,” or “plywood.” The resulting records were then reviewed to determine whether the injury appeared to be related to the shutter. All records that mention “shutter” in the Narrative_1 and Narrative_2 text fields were not included because some of the shutters involved in injuries might not be hurricane shutters but regular shutters.
The distribution by selected variables was determined for the resulting cases as well as a weighted estimate calculated by the sum of the numbers in the Weight number field. It should be noted that the CPSC considers an estimate unstable and potentially unreliable when the estimate is less than 1200, the number of records used is less than 20, or the coefficient of variation (CV) exceeds 33%.17
The geographic distribution (ie, state or county) of the cases could not be examined because such information was not available in the public NEISS database. The distribution of the cases by year was determined. For any years where there seemed to be a higher number of cases, the distribution of cases by month and day was examined to identify any apparent clusters of cases. These clusters were then compared with the reported dates of landfall of hurricanes and tropical storms in the United States. Because of the small number of cases in small time-frames, weighted estimates were not calculated for the temporal analyses.
RESULTS
During 2001-2017, a total of 329 hurricane shutter-related injuries occurred and were reported to the NEISS, resulting in a national estimate of 19,335 injuries during that time period. Figure 1 shows the annual number of hurricane shutter-related injuries. While few injuries were reported during most years, higher numbers were reported during 2005 (14.6% of total), 2008 (10.9% of total), and 2017 (33.4% of total). When the monthly number of injuries were examined for these 3 years, the majority of the injuries were reported during October 2005 (n = 36), August 2008 (n = 30), and September 2017 (n = 103); these 3 months accounted for 169 (51.4%) of the total injuries during 2001-2017. Table 1 presents the number of hurricane shutter-related injuries during these 3 months and selected others when major hurricanes made landfall in the United States. Hurricane Wilma made landfall in October 2005, Tropical Storm Fay in August 2008, and Hurricane Irma in September 2017; all 3 made initial landfall in Florida. Fifteen (4.6%) hurricane shutter-related injuries in total were reported in the months of the 5 other storms examined (Hurricane Ivan in September 2004, Hurricane Katrina in August 2005, Hurricane Ike in September 2009, Hurricane Sandy in October 2012, Hurricane Harvey in August 2017). Figure 2 provides the daily number of hurricane shutter-related injuries for the 10 days before landfall, the day of landfall, and the 10 days after landfall for Hurricane Wilma, Tropical Storm Fay, and Hurricane Irma. The number of reported hurricane shutter-related injuries began to increase days before landfall and continued for days after landfall for all 3 storms.
FIGURE 1 Annual Number of Hurricane Shutter-Related Injuries Treated in US EDs, NEISS.
TABLE 1 Number of Hurricane Shutter-Related Injuries Treated in US EDs, NEISS, During Selected Months With Storms
Abbreviations: ED, emergency department; NEISS, National Electronic Injury Surveillance System; US, United States.
FIGURE 2 Daily Number of Hurricane Shutter-Related Injuries, NEISS.
Table 2 shows the distribution of hurricane shutter-related injuries by patient demographics and the circumstances of the injury. The majority of the patients were adults with the highest proportions being age 60-69 years followed by 30-39 years. Most of the patients were male. Almost all of the injuries occurred at the patient’s home.
TABLE 2 Demographics and Circumstances of Hurricane Shutter-Related Injuries Treated in US EDs, NEISS, 2001-2017
a Estimate = sum of the Weight field.
b The US Consumer Product Safety Commission considers an estimate unstable and potentially unreliable when the number of records used is less than 20 or the estimate is less than 1200. Therefore, the estimate for this subgroup should be considered unstable and potentially unreliable.
Abbreviations: ED, emergency department; NEISS, National Electronic Injury Surveillance System; US, United States.
The most frequently reported injuries were laceration followed by sprain or strain, fracture, and contusions or abrasions. Lower extremities were the most commonly affected body part followed by the upper extremity and head or neck. The majority of patients were treated or examined at the ED and released; 1 death was reported (Table 3).
TABLE 3 Type and Management of Hurricane shutter-related injuries treated in US EDs, NEISS, 2001-2017
a Estimate = sum of the Weight field
b The US Consumer Product Safety Commission considers an estimate unstable and potentially unreliable when the number of records used is less than 20 or the estimate is less than 1200. Therefore, the estimate for this subgroup should be considered unstable and potentially unreliable.
Abbreviations: ED, emergency department; NEISS, National Electronic Injury Surveillance System; US, United States.
DISCUSSION
Hurricane shutter-related injuries may occur with hurricanes and tropical storms.Reference Pacenti16 Of the 329 hurricane shutter-related injuries identified during 2001-2017, over half were reported during 3 specific months (October 2005, August 2008, September 2017) with over 30% reported during 1 of these months (September 2017). During these 3 months, storms made landfall in the United States: Hurricane Wilma in October 2005, Tropical Storm Fay in August 2008, and Hurricane Irma in September 2017. When the daily number of injuries was examined in relation to the landfall of these 3 storms (Figure 2), the number of reported injuries began to increase days before landfall and continued for days after landfall for all 3 storms.
These observations suggest that individuals are at increased risk of hurricane shutter-related injuries around the time hurricanes and tropical storms make landfall. Moreover, injuries might occur in the days before the storm makes landfall, as hurricane shutters are installed, and continue for days afterward, as the shutters are removed. It should be emphasized that this relationship between injuries and storms is based solely on examination of the dates of the injuries and storms. The NEISS public database does not contain information on the geographic location of the hospital EDs where the patients were treated. Thus, some of the injuries may be close to the landfall of a storm in time but widely separated geographically.
Also, hurricanes and tropical storms differed in the number of hurricane shutter-related injuries that occurred in the month they made landfall, with few injuries reported with some of the storms, such as Hurricane Ivan, Hurricane Katrina, Hurricane Ike, Hurricane Sandy, and Hurricane Harvey (Table 1). Thus, it might be difficult to predict the number of hurricane shutter-related injuries that might occur with a particular storm. Hurricanes and tropical storms may differ in the degree of hurricane shutter-related injuries that occur due to differences in the characteristics of the storm (eg, wind strength) and where the storm made landfall. There may be geographic differences in the requirement or use of hurricane shutters. Also, for those storms with few reported injuries, the hospital EDs where these hurricane shutter injuries may have been treated may not have been included in the NEISS.
It may be noted that these patterns were based on a relatively small number of reported hurricane shutter injuries. Although the 3 months with the highest number of hurricane shutter injuries (October 2005, August 2008, September 2017) accounted for over half of all such injuries, there were only a total of 169 injuries. This may be considered small when the storms that made landfall during these months (Hurricane Wilma, Tropical Storm Fay, Hurricane Irma, respectively) affected millions of people, often in multiple states. However, as indicated in the Methods section, these injuries are from hospital EDs that are a probabilistic sample of all US hospital EDs and are not injuries that were treated at all US hospital EDs. The injuries in the NEISS database can be used to calculate an estimate of the total number of injuries. In this instance, the 169 injuries result in an estimated 9775 hurricane shutter injuries during these 3 months. Furthermore, all of the individuals affected by a storm may not be at equal risk of a hurricane shutter injury. Individuals further inland may not be expected or required to put up hurricane shutters.
The majority of the patients were male, and most were adults, with the highest proportions being age 60-69 years followed by 30-39 years. Almost all of the injuries occurred at the patient’s home. This might be expected if the injuries occurred during installation or removal of hurricane shutters from the home.
The injuries were most often laceration, sprain or strain, fracture, and contusions or abrasions. The affected body part was most often the lower extremity followed by the upper extremity and head or neck. Most of patients were treated or examined at the ED and released. A news article suggested that hurricane shutter-related injuries tended not to be life-threatening and consist of lacerations, falls, and back injuries.Reference Pacenti16
The results of this study do not suggest ways to prevent hurricane shutter injuries. However, precautions have been recommended by others to minimize the risk of hurricane shutter-related injuries. Work gloves should be worn to avoid lacerations and splinters. Shutters should be installed before the storm winds increase. More than 1 person should be involved in the installation or removal of the shutters to assist in the handling of the heavy shutters and securing the ladder. Persons should be careful of their posture and bend with their knees to avoid back injuries. Any ladders should be checked before use. Plenty of time should be allotted for the installation and removal of shutters.18
There are limitations to this study. Cases were identified by searching for “shutter” and “hurricane,” “storm,” “metal,” “aluminum,” or “plywood” in the electronic record narrative. Hurricane shutter-related injuries where these terms were not documented in the narrative would not have been included in this study. Also, some of the injuries included in the study might not have been related to hurricane shutters but to other types of shutters, such as regular shutters. Only those injuries treated at an ED were included. The number of hurricane shutter-related injuries not seen at an ED is unknown. In addition, there were relatively few cases. This is particularly important when calculating estimates, which may be unstable and potentially unreliable when small numbers of cases are being examined. As mentioned previously, the geographic location (eg, state or county) of the ED where the patient was treated was not available in the NEISS public database. This limited the ability to more definitively link particular injuries to particular storms. This study was limited to hospital EDs in the United States. It is unclear whether other countries affected by such storms may also experience hurricane shutter-related injuries, and if they do, whether such injuries would follow a similar pattern. Studies of hurricane shutter-related injuries in other countries would be useful to add to information on this topic.
CONCLUSIONS
Over half of the hurricane shutter-related injuries reported to EDs appeared to occur in association with hurricanes and tropical storms. Most patients were male and adults. The majority of injuries occurred at home. The most frequently reported injuries were laceration followed by sprain or strain, fracture, and contusions or abrasions. Lower extremities were the most commonly affected body part followed by the upper extremity and head or neck. The majority of patients were treated or examined at the ED and released.
