Epidemiologic Challenges in Post-Disaster Population

The post-Katrina New Orleans environment is a major concern of returning residents. The Healthy Home Project is a study assessing the indoor environment of New Orleans households (N=100). Given the combination of mass migration, extensive demolitions, and low occupancy rates, standard probability sampling was not feasible. Challenges included lack of an accurate sampling frame and methods to delineate non-response from vacancy. We developed a sampling methodology to obtain a representative sample of all occupied households in New Orleans.

 US Census 2000 data and post-Katrina occupancy estimates were combined to develop sampling parameters. Stratified random sampling was employed to ensure representation of all areas of the city, including areas heavily flooded. Ten New Orleans planning districts were included as strata. Required sample size per stratum was proportional to post-Katrina occupancy. Households were oversampled to account for non-response and vacancies. The following formula was used.  

                                                     

A total of 1,638 addresses were drawn randomly from a pre-Katrina sampling frame to represent all occupied housing in New Orleans.  To preserve the integrity of the random selection process, addresses were picked chronologically for recruitment. Each household's occupancy was determined based on a 21-item occupancy criteria.

Probability sampling in a post-disaster environment poses unique challenges. Although non-probability sampling methods are acceptable, we developed a probabilistic method to obtain a representative sample. This approach may help ensure greater representation of New Orleans households.