Impacts of Hurricane-Related Flooding on Time to Initial Cancer-Directed Treatment in North Carolina

Hurricanes can disrupt healthcare delivery and patients' access to timely cancer care. Such disruptions may contribute to adverse cancer outcomes, yet few studies have examined hurricane-related delays in treatment initiation.

We conducted a quasi-experimental, observational study using a controlled interrupted time series to evaluate the association between hurricane exposure and time to initial treatment among patients diagnosed with breast, colorectal, or lung cancer in North Carolina (NC) following Hurricanes Matthew (2016) and Florence (2018). The analysis included 59,406 patients between November 2015 and August 2019 in the NC Central Cancer Registry. Hurricane-exposed areas were defined using (i) presidential disaster declarations (county level) and (ii) flood extent data [Flood Extents Archive ((FLDEX)] at county and ZIP-code levels. The monthly median time to initial cancer-directed treatment within 12 months of diagnosis was modeled using AutoRegressive Integrated Moving Average (ARIMA) to estimate immediate (level) and sustained (trend) effects.

No significant immediate or sustained changes in the median time to treatment were observed using Federal Emergency Management Agency-based exposure definitions. However, using ZIP code-level FLDEX data, patients with colorectal cancer in high-flood exposure areas (≥0.1% and ≥0.5% flooded) experienced significant immediate median delays of 10.39 [95% confidence interval (CI) = 2.8-18] and 10.97 days (95% CI = 2.1-19.9), respectively.

Although no significant effects on time to treatment were observed using traditional measures of hurricane hazards, more granular flood extent data revealed treatment delays among patients with colorectal cancer living in areas with greater flooding extent.

These findings provide support for further investigation of the use of ZIP code-level flood extent data to evaluate hurricane impacts on continuity of cancer care.