Introduction
Shortly after 8:00 pm on Monday, October 29th 2012, Hurricane Sandy touched down in New York City (NYC; New York USA) causing massive destruction, paralyzing the city, and destroying lives. Staten Island, one of NYC’s five boroughs, suffered significant devastation after the storm. Fifty-three deaths were reported in New York StateReference Greenstein, Chacko, Ardolic and Berwald 1 secondary to Hurricane Sandy; 24 of those deaths occurred on Staten Island. 2
In 2010, Staten Island had a reported population of 468,730. 3 More recently, the census bureau estimated the 2013 population at 472,621. 4 This population is served primarily by three hospitals located on Staten Island: Staten Island University Hospital North (SIUH-N), Staten Island University Hospital South (SIUH-S), and Richmond University Medical Center (RUMC). Hurricane evacuation zones at the time of the storm put SIUH-N in evacuation Zone A and at most significant risk of the three hospitals for flooding. The RUMC, located approximately five miles north of SIUH-N, was not in a hurricane evacuation zone due to its higher altitude.
Since the storm, the flood zones of NYC have been redefined to a 6-level scale to provide greater flexibility in targeting areas to evacuate in advance of a predicted storm. 5 In the revised system, SIUH-N is now in Flood Zone 2 and SIUH-S is in Flood Zone 5 (Zone 1 being the greatest risk). However, given the degree of devastation to the island from Hurricane Sandy, both sites and their patients proved to be vulnerable to this natural disaster, regardless of zone assignment.
Research has shown that considerable damage and loss of life can be averted in at-risk areas from advanced preparation in communication procedures, evacuation planning, and resource allocation. However, research is limited in describing how natural disasters of this magnitude affect emergency departments (EDs).
The experience of the ED at SIUH-N, a large academic medical center seeing approximately 90,000 visits per year, is described through an analysis of patient demographics and ED diagnoses immediately before, during, and after the event. These data can help predict future expectations and preparations for EDs in areas contending with natural disasters of Hurricane Sandy’s magnitude.
Materials and Methods
This was a retrospective, descriptive study approved by the Institutional Review Board at Staten Island University Hospital. Data were obtained from the Emergency Department Information Systems (EDIS; Allscripts EDIS V.7.2.1; Raleigh, North Carolina USA), which is used for documentation, order entry, and patient tracking. There were times during the study period when the computer system was unavailable due to storm-related outages. During these times, paper downtime charts were used and later scanned into the EDIS.
Patient demographics, ED census, diagnoses, admission rates, and mode of transport were reviewed for the days immediately prior to, during, and after the storm. Data were compared to the previous years’, matched to same days of the same calendar week. Specifically, data from October 26, 2012 through November 2, 2012 were collected and compared to the same days of the week for the period of October 28, 2011 through November 4, 2011. Data were analyzed to look at patient characteristics on a daily and hourly basis.
The purpose of data collection was to identify trends in patient volume, types of conditions treated (inferred from diagnoses), and admission rates in the pre- and post-storm periods. This was compared with a control group of visits on similar days from the prior year. This control period was selected to account for the seasonal variation in complaints in the geographic area.
Results
The average daily census for the ED in 2012 was 253 patients. As seen in Table 1, there was a substantial drop in patient volume for all age groups on the day of landfall (Day 0) with a total daily volume of 114 patients (-55%). Of note, based on storm surge predictions, there were mandatory evacuations for Zone A on October 28, 2012 (Day -1). The trend to normalization of average daily census was slower in the pediatric population, which had a baseline average daily census of 58 patients. Patients aged 65 years and older showed a considerable increase in volume in the days following the storm. No significant gender differences were observed. Overall, the daily volume exhibited a precipitous decline on the days leading up to the storm, followed by relatively quick return to usual volumes (Figure 1).
Table 1 Demographics of Hurricane Sandy Visits
Figure 1 Number of Patients Seen per Day.
Primary diagnoses for patients seen on the day of the storm and the day following the storm were also evaluated. These were compared to the diagnoses on the same days of the week for the last week of October 2011. The ten most common primary diagnoses are shown in Figure 2. Chest pain and abdominal pain diagnoses were common primary diagnoses when compared against similar calendar dates in 2011. However, there were a notably larger percentage of patient visits for medication refills in 2012 at 5.8% versus 0.4% in 2011 (P<.05). Lacerations and cold exposure also were increased substantially in 2012 at 7.6% of patient visits versus 2.8% (P<.05) and 3.8% versus 0.0% (P<.05), respectively. This is likely related to the destruction of homes and flooding from the storm, in addition to the lack of availability of local clinics and pharmacies. Conversely, atraumatic back pain and motor vehicle accidents, two of the most common primary diagnoses in 2011, were nearly absent primary diagnoses in 2012.
Figure 2 Admission Rates per Day.
The average admission rate from the ED at SIUH-N in 2012 was 27%. In the days leading up to the storm, a large drop in admissions was seen, with a nadir seen on the day following the storm at five percent (Figure 3). Review of admitted patients revealed atypical admissions for home care services, such as need for supplemental oxygen or ventilators. Such equipment was unobtainable due to power outages and road closures impacting the delivery of supplies needed to meet the patients’ chronic needs.
Figure 3 Top Diagnoses during Hurricane Sandy. Abbreviations: COPD, chronic obstructive pulmonary disease; SOB, shortness of breath.
Lastly, SIUH-N ED typically sees 18% of its patient volume arriving by Emergency Medical Services (EMS). This is consistent with what the ED saw on Day 2 and three days prior to the storm. However, on the day following the storm, only two percent of patients arrived by ambulance. A significant drop in EMS utilizations can be observed during Day 0 and Day +1 (Figure 4). This likely is attributed to the lack of accessibility from flooding, roadway destruction, and loss of electricity and telephone lines.
Figure 4 Patient Modes of Arrival during Hurricane Sandy. Abbreviation: EMS, Emergency Medical Services.
Discussion
This study was performed to describe and analyze the distribution of ED visits as a result of Hurricane Sandy. Comparing the patients in the days surrounding the storm to the previous year provides information that can be critical for future disaster planning. Understanding patient trends during natural disasters can help EDs prepare for changes in volume and facilitate proper preparation of personnel and medical supplies.
Overall, there was a decline in patient census leading up to the storm with a surge in patient volume afterwards. As a result of the advanced warning from the anticipated storm, SIUH was able to make appropriate preparations. The inpatient units facilitated discharges for patients who were near discharge in an effort to decrease census. Inpatient staffing was maintained at its normal levels on the day leading up to, and on the day of, the storm. Due to anticipated potential dangerous travel conditions, staff was housed in the hospital. In the ED, a stay team and a recovery team were assembled. One unanticipated complication was the delay in getting the recovery team to the hospital due to bridge and other road closures along with gas shortages. In recent years, the ED at SIUH-N had the opportunity to modify staffing for inclement weather. Hurricane Irene (2011) and several blizzards resulted in the utilization of stay teams and recovery teams. However, in the case of Hurricane Sandy, the stay team had a unique experience relative to the other occurrences, notably the longer need for the stay team and the associated emotional burden.
These data indicate the surge in census and admissions to the hospital only arose in the aftermath of the storm. Staffing did not mirror this influx due to the difficulties of getting staff in after the storm. This trend can be applied in future disaster-related health care needs as staffing should be adjusted to absorb this influx of patients quickly and efficiently. Mandatory full staffing for the day of storm, in this case, was not as urgent as the need in the hours following the storm. Early notification, media coverage, and mandatory evacuations led to a major reduction in morbidity and mortality during the storm and likely lead to the lower than expected patient volume during the storm and the upturn in volume immediately following.
Flooding, destruction of roads, and poor EMS access for patients during the storm potentially limited access for those in need of immediate care. Several members of the ED staff participated with a local mobile medical unit performing home visits in the community. The aforementioned issues were reported by patients in addition to other unanticipated obstacles. Some patients reported not seeking care after the storm for fear of missing the Federal Emergency Management Agency (FEMA; Washington, DC USA) or insurance representatives’ visits to their home.
Another important finding in this study was the change in primary diagnoses for patients from the baseline characteristics. It is clear that personnel, staffing, and medical supply needs change to reflect the presenting conditions. At SIUH-N, a large area for patients who simply needed supplemental oxygen, but did not require formal hospital admission, was set up. Increasing ancillary staff to handle the arrival of urgent-care complaints, such as laceration, medication refills, and falls, would have been helpful. Having a better understanding of patient types following such an incident will help EDs prepare for future natural disasters.
Limitations
This study was limited by its retrospective nature. In addition, generalization of this study may be limited by the unique geographic location of Staten Island. Staten Island is positioned at the center of New York Bight, a sharp bend in the shoreline between New Jersey (USA) and Long Island (New York USA). This sharp bend leaves little room for the surging waters and cyclonic winds to disperse, as a result, causing larger than normal storm surges. This is one of the primary reasons why the New York Metro area is considered a high danger zone for storm-related surges. Lastly, the advance notification, forecasting of the storm, and mandatory evacuations in Zone A prior to landfall likely influenced the volume and demographics of patients, as well as the conditions treated, in the days surrounding Hurricane Sandy.
Conclusions
The daily ED census saw a large decline in the days leading up to the storm, most significantly on the day of landfall. Additionally, the types of conditions treated varied from the baseline, resulting in a drop in hospital admissions. Predicting clinical needs in the face of a natural disaster is difficult. Emergency department patients stand to benefit as data, such as those presented here, can help predict future expectations and preparations for EDs facing natural disasters.
