Figures

Authors

Affiliations

Hannah Rosenthal

Albert Einstein College of Medicine

Vagish Hemmige

Montefiore Medical Center/ Albert Einstein College of Medicine

The code in this script generates figures for the manuscript.

Source code

The full R script is available at:

• R/figures.R

This R script file is itself reliant on the following helper files:

• R/setup.R
• R/functions.R

Final flowchart

The following code uses the flowchart package to create a STROBE flowchart for the study.

#Final STROBE flowchart
patients_fc_matched%>%
  fc_draw()%>%
  fc_export("figures/flowchart.svg",
            width=20,
            height=10)

Beeswarm plot of total inflated costs

The following code uses the ggbeeswarm add-on to ggplot2 to compare cases and controls in terms of total inflated costs post-index. No attempt is made to adjust for censoring or death.

#GGplot that is beeswarm of adjusted total costs
cost_inflated%>%
  mutate(IN_CLM_365d_cost_adjusted_total=ifelse(IN_CLM_365d_cost_adjusted_total==0, 1, IN_CLM_365d_cost_adjusted_total))%>%
  ggplot()+
  geom_quasirandom(mapping=aes(x=patient_type, y=IN_CLM_365d_cost_adjusted_total), alpha=0.1)+
  scale_y_log10()+
  theme_classic()

Month-by-month plots

In order to plot monthly costs predicted from the longitudinal models generated in the R/models.R script, we use the plot_mean_monthly_costs() function from the R/functions.R script to create the ggplot objects, and then save as .svg files.

#Loop that plots the different outcomes longitudinally

for (outcome in c("grand_total_cost_month", "IN_CLM_month_total", "PS_REV_month_total", "IN_CLM_month_groupedHomeHealth",
                  "IN_CLM_month_groupedHospice","IN_CLM_month_groupedNonclaimauxiliary","IN_CLM_month_groupedDialysis",
                  "IN_CLM_month_groupedOutpatient","IN_CLM_month_groupedInpatient","IN_CLM_month_groupedSkilledNursingFacility")){
  
  
  plot_mean_monthly_costs(fit[[outcome]][["glmmTMB"]][["tweedie"]])
  ggsave(filename = paste0("figures/Longitudinal ", outcome," glmmTMB", " ", "tweedie.svg"),
         width=14,
         height=7
         )
  
  plot_mean_monthly_costs(fit[[outcome]][["glmmTMB"]][["linear"]])
  ggsave(filename = paste0("figures/Longitudinal ", outcome," glmmTMB", " ", "linear.svg"),
         width=14,
         height=7
  )
  
  
  plot_mean_monthly_costs(fit[[outcome]][["gee"]][["linear"]])
  ggsave(filename = paste0("figures/Longitudinal ", outcome," gee", " ", "linear.svg"),
         width=14,
         height=7
  )
  
  
  plot_mean_monthly_costs(fit[[outcome]][["glmmTMB"]][["log"]])
  ggsave(filename = paste0("figures/Longitudinal ", outcome," glmmTMB", " ", "log.svg"),
         width=14,
         height=7
  )
  
  
  plot_mean_monthly_costs(fit[[outcome]][["gee"]][["log"]])
  ggsave(filename = paste0("figures/Longitudinal ", outcome," gee", " ", "log.svg"),
         width=14,
         height=7
  )
  
  
}

Other portions of the analysis

• Setup: Defines global paths, data sources, cohort inclusion criteria, and analysis-wide constants.
• Functions: Reusable helper functions for cohort construction, matching, costing, and modeling.
• Tables: Summary tables and regression outputs generated from the final models.
• About: methods, assumptions, and disclosures