Public health concerns from Katrina, especially the potential for infectious disease, are significant. Surveillance conducted in the shelters in Houston following Katrina indicated significant numbers of individuals with respiratory and diarrheal disease, including those caused by Norovirus and possibly Shigella. While a significant amount of research exists documenting the public health impacts of natural disasters, very little research exists that examines the correlation of those health impacts with measures of the hazard and resultant damage to the built and natural environment. The purpose of this study is to examine the correlation between health outcomes for people in the impacted areas and hazardous conditions to which they may have been exposed including: 1) structural damage which presents a potential for physical injuries; 2) exposure to environmental contaminants in the water and air, including toxic substances and infectious agents; 3) length of time exposed to the elements prior to evacuation; 4) time with limited access to food and/or water; and 5) length of time without access to medical care, including access to necessary medications.
Primary data collection will focus on data related to the population's exposure to potentially hazardous conditions. Information on the flood levels, wind speed, and the nature of contamination will be obtained from secondary sources as well as a contemporaneous environmental study by one of the co-investigators. HAZUS software will be used to estimate building damage on a census tract level (hurricane) or a census block level (flood) for the counties of interest. In-person interviews will be conducted in four selected impacted communities in January/February 2006 including Biloxi, MS; Gulfport, MS; Slidell, LA; and if possible New Orleans, LA. A predictive risk model will be developed to elucidate relationships between the various independent variables and the dependent variables (health outcomes). As the variables are interrelated, multivariate analyses will be utilized to operationalize the model. Appropriate analytic software will be used to take the sampling methodology into consideration and minimize the effect on the variance. The resulting model will provide essential information which can be used to improve decision-making regarding mitigation, preparedness, and response activities that can reduce the potential for short and long-term health burdens as a result of hurricanes, and it may allow for extrapolation to other types of disaster as well.
