Jiangchao3 / tropical_cyclones_hospitalizations_nat_comms

Code used for analysis and plotting for Nature Communications paper 2021

Tropical cyclone exposure is associated with increased hospitalization rates in older adults

Parks, R.M., Anderson, G.B., Nethery, R.C. et al. Tropical cyclone exposure is associated with increased hospitalization rates in older adults. Nat Commun 12, 1545 (2021). https://doi.org/10.1038/s41467-021-21777-1

Code used for analysis and plotting of Figures for Parks et al. Nature Communications paper 2021

Art by Amy Wolfe

Model for main analyses in paper

model/model.R

Figure 1. Tropical cyclone exposure days. Number of days with tropical cyclone exposure by county for 1999 – 2014.

figures/figure_1.R

Figure 2. Annual Medicare hospitalizations by cause. Number of Medicare hospitalizations by year and cause of hospitalization for counties with at least one tropical cyclone exposure for 1999 – 2014.

figures/figure_2.R

Figure 3. Percentage change in hospitalization rates with tropical cyclone exposure by cause of hospitalization and lag time. Lag time is measured in days after tropical cyclone exposure. Dots show the point estimates and error bars represent Bonferroni-corrected 95% confidence intervals.

figures/figure_3.R

Figure 4. Average percentage change in hospitalization rates with tropical cyclone exposure by cause and sub-cause of hospitalization. Average percentage change in hospitalization rates is across studied lag period (0 to 7 days after tropical cyclone exposure). Dots show the point estimates and error bars represent 95% confidence intervals.

figures/figure_4.R

Figure 5. Percentage change in hospitalization rates with tropical cyclone exposure by cause of hospitalization, intensity of local wind exposure and lag time. Lag time is measured in days after tropical cyclone exposure. Dots show the point estimates and error bars represent Bonferroni-corrected 95% confidence intervals.

figures/figure_5.R

Figure 6. Percentage change in hospitalization rates with tropical cyclone exposure by cause of hospitalization, type of hospital admission and lag time. Lag time is measured in days after tropical cyclone exposure. Dots show the point estimates and error bars represent Bonferroni-corrected 95% confidence intervals.

figures/figure_6.R

Figure 7. Change in hospitalizations for the US Medicare population, for an expected number of tropical cyclone exposures by county over a decade. The top row shows the break down by cause covering the day of tropical cyclone exposure to seven days afterwards, with black bars representing Bonferroni-corrected 95% confidence interval. The bottom row shows the break down by lag days after the exposure.

figures/figure_7.R