A Web Application and Software Package for Analysis and Exploration of the FCDS Florida Cancer Registry Data

Garrick Aden-Buie1, Jordan Creed1, Travis Gerke1
  1. Gerke Lab, Cancer Epidemiology, Moffitt Cancer Center, Tampa, FL

FCDS Explorer
is a web app for exploring
interactive cancer statistics

fcds
is an R Package for
tidy data processing of
Florida Cancer Registry Data

https://gerkelab.github.io/fcds

Abstract

Understanding and adapting to the specific needs of cancer patients within Moffitt’s catchment area, and more broadly across Florida, is critical to achieving Moffitt’s IMPACT 2028 goals [1] and CCSG Center of Excellence status. To this end, the Florida Statewide Cancer Registry [2] provides a comprehensive database of cancer diagnoses throughout Florida via the Florida Cancer Data System (FCDS). The goal of this work is provide a consistent and broadly accessible interface to FCDS resources. We presents a web application and a software package, fcds, to clean and harmonize FCDS cancer registry data to facilitate integration of highly-specific regional cancer incidence data within research data analyses of Moffitt patient information.

R Package Highlights

The primary output of this work is software package for the R statistical computing environment [3] to facilitate analysis and use of FCDS cancer registry data with a clear and consistent data structure as advocated in Tidy Data by Hadley Wickham [4].

Import and Clean FCDS Cancer Statistics Data

library(tidyverse)
library(fcds)
fcds <- fcds_import("STAT_dataset_2018.dat")

# Load previously imported data
fcds <- fcds_load()

Subset and Count Cancer Incidence

fcds_moffitt_prostate <- fcds %>% 
  filter(cancer_site_group == "Prostate Gland", year > 1985) %>% 
  filter_age_groups(age_gt = 20) %>% 
  count_fcds(sex = "Male", moffitt_catchment = TRUE)
sex county_name year_group year age_group n
Male Charlotte 1986-1990 1988 50 - 54 1
Male Charlotte 1986-1990 1988 55 - 59 12
Male Charlotte 1986-1990 1988 60 - 64 57
Male Charlotte 1986-1990 1988 65 - 69 126
Male Charlotte 1986-1990 1988 70 - 74 137

Calculate Age-Adjusted Rates

fcds_moffitt_prostate <- fcds_moffitt_prostate %>% 
  complete_age_groups(age_gt = 20) %>% 
  age_adjust() %>% 
  mutate(n = n / 5, rate = rate / 5)
sex county_name year_group year n population rate
Male Charlotte 1986-1990 1988 107.0 38144 137.3968
Male Charlotte 1991-1995 1993 160.6 48412 160.8875
Male Charlotte 1996-2000 1998 212.4 53639 187.3476
Male Charlotte 2001-2005 2003 214.4 60336 174.3292
Male Charlotte 2006-2010 2008 226.4 64059 176.2285

Interactive and Static Maps

fcds_map(fcds_moffitt_prostate) +
  facet_wrap(~ year_group) +
  ggtitle("Prostate Cancer in Moffitt Catchment Area",
          "Age-Adjusted Rates, Men Aged 20+") +
  labs(fill = "Mean Yearly\nAge-Adjusted Rate") +
  theme(legend.position = "bottom")

FCDS Explorer

The fcds R package powers and enables an easy-to-use and accessible web application for Moffitt researchers. The web application, called FCDS Explorer, is built using the R Shiny framework [5] and provides features for summarizing and presenting the FCDS data.

Select Specific Cancer Types

View Incidence on Interactive Map

Summarize Cancer Incidence Tables

Reference ICD-O-3 Code List

References

[1] “IMPACT 2028 building for the next 10 years.” https://moffitt.org/media/9330/moffitt-annual-report-2018.pdf.

[2] “Florida cancer data system.” https://fcds.med.miami.edu/inc/welcome.shtml.

[3] R Core Team, “R: A language and environment for statistical computing.” R Foundation for Statistical Computing, Vienna, Austria, 2018 [Online]. Available: https://www.R-project.org/

[4] H. Wickham, “Tidy data,” Journal of Statistical Software, Articles, vol. 59, no. 10, pp. 1–23, 2014 [Online]. Available: https://www.jstatsoft.org/v059/i10

[5] W. Chang, J. Cheng, J. Allaire, Y. Xie, and J. McPherson, “Shiny: Web application framework for r.” 2018 [Online]. Available: https://CRAN.R-project.org/package=shiny