Purpose: Floods cause significant issues for citizens, authorities, and government agencies. Hurricane Harvey (HH) caused havoc in Texas in August 2017, as the third major flood to affect the city of Houston in 3 years. As clinics and medical resources are taxed, patients may rely on emergency departments (ED) and urgent care centers (UC). Infectious disease prevalence has been reported to increase after flooding events, but no study to date has specifically described the nature of pediatric visits. This study aims to describe the geospatial distribution of pediatric patients’ residences in the Houston area, and characterize their utilization of EDs and UCs before and after HH while investigating a potential link between exposure to flooding and increased pediatric disease burden after HH. Methods: This is a retrospective cross-sectional study of patients seen at all Texas Children’s Hospital (TCH) EDs and UCs from July 30, 2017 - September 23, 2017. We extracted patient demographic information, including home address, and the first 5 discharge diagnoses. We grouped ICD9 diagnosis codes using a published diagnosis grouping system (DGS). Using ArcMap 10.6 (Esri, Redlands, CA), we used overlay analysis techniques to determine the location of patients associated with each diagnosis in the period before (July 30th-August 25th 2017) and after (August 26th-September 23rd 2017) HH, and assess the likelihood that their residences may have flooded (based on FEMA-designated floodplain data). Results: There were 17,705 visits to EDs and UCs during our study period. A total of 32,299 diagnoses were collected, of which 32,126 were associated with patients residing in Texas. Most diagnoses were associated with patients residing in Harris County (73%), Fort Bend County (9.7%), and Montgomery County (6.3%). Despite a period of non-operation in the 6 days following HH, there was a significant increase in UC diagnoses in the period after HH (9306) compared to the period preceding HH (6744) (p<0.001). In the period before HH, 25.4% of diagnoses made in EDs were associated with patients living inside FEMA floodplains; as opposed to 27.8% in the post-HH period (Figure 1). Similarly, before HH, 21% of diagnoses made in UCs were associated with patients living inside FEMA floodplains; as opposed to 23.2% in the post-HH period (Figure 2). Conclusions: Floods have a significant impact on child health as shown in studies of the general population. This study uses novel geospatial techniques to provide some new insight into exposure to flooding and increased risk of illness in the pediatric population, and adds to the body of literature supporting the link between health and environmental exposures. Future work will examine the relationship between increased disease specific prevalence and residence in floodplains.

Figure 1

Maps of ED patients' residences pre- and post- Hurricane Harvey

A: Map of ED patients' residences between July 30th- August 25th 2017 B: Map of ED patients' residences between August 26th-September 30th 2017 Green dots indicate residences located outside FEMA flood plains. Red dots indicated residences located inside FEMA flood plains.

Figure 1

Maps of ED patients' residences pre- and post- Hurricane Harvey

A: Map of ED patients' residences between July 30th- August 25th 2017 B: Map of ED patients' residences between August 26th-September 30th 2017 Green dots indicate residences located outside FEMA flood plains. Red dots indicated residences located inside FEMA flood plains.

Close modal

A: Map of UC patients' residences between July 30th- August 25th 2017 B: Map of UC patients' residences between August 26th-September 30th 2017 Green dots indicate residences located outside FEMA flood plains. Red dots indicated residences located inside FEMA flood plains.

A: Map of UC patients' residences between July 30th- August 25th 2017 B: Map of UC patients' residences between August 26th-September 30th 2017 Green dots indicate residences located outside FEMA flood plains. Red dots indicated residences located inside FEMA flood plains.

Close modal