BACKGROUND: The objective of this study is to evaluate the effects of Hurricane Sandy on hospital admissions due to cardiovascular diseases.
METHODS:
This was a retrospective cohort design using a Poisson regression analysis to assess the effects of Hurricane Sandy on cardiovascular hospitalizations in the impacted area of New York State. Hospitalization data (2001-2012) were extracted from a statewide hospital discharge reporting database. Cardiovascular disease (CVD) diagnoses were extracted based on ICD9 codes: 401-404 (hypertensive disease), 410-414 (ischemic heart disease), 427 (cardiac dysrhythmias), 428 (heart failure), 436-438 (acute, late, and other cerebrovascular diseases). The exposures were defined by three time periods: 1) immediate Hurricane impact period (10/28/12 – 11/9/12); 2) three months after Hurricane Sandy, and 3) one year after Hurricane Sandy. Our analysis evaluated changes in the CVD hospitalization rates in each of the three time periods by comparing the rates during the same periods in preceding years, as well as comparing the rates in the storm zone with the non-storm zone areas. The final regression models were stratified by county and adjusted for long-term temporal trends and short-term day of the week effects.
RESULTS: After adjusting for long-term trends, a significant decline in rates of CVD was found in both storm and non-storm zones when comparing the same time periods during Hurricane Sandy with those in previous years. However, there were some exceptions. For example, Queens County had a significantly higher hospitalization rate in the storm zone vs. non-storm zone area (1.28 p: <0.01). Similarly in the 12-month analysis post Hurricane Sandy, Westchester County had a significantly higher hospitalization rate in the storm zone vs. non-storm zone area (1.23 p: <0.01)
CONCLUSIONS: The overall rate of CVD-related hospitalizations declined during Hurricane Sandy in all study areas, and more so in the non-storm zone areas. Although Hurricane Sandy was considered a “super storm” affecting large areas of the state, we did not find evidence that people living in geographic areas prone to storm surges had significantly higher risks for CVD than people living further inland.