In GerkeLab/fcds: Process Data from the Florida Cancer Data System

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Installation

# install.packages("devtools")
devtools::install_github("GerkeLab/fcds")

Setup

library(dplyr)
library(fcds)

The fcds package doesn't include data from the FCDS Florida Cancer Registry. Please visit the FCDS Data Request page to request the most recent Florida Statewide Cancer Registry data.

Once you have downloaded the data, you can use fcds to pre-process and import the FCDS data into R.

fcds <- fcds_import("STAT_dataset_2018.dat")

The imported, pre-processed data is automatically cached in the .fcds directory in your home directory for later use. Calling fcds_load() will re-load the cached data.

fcds <- fcds_load()

You can customize the folder where the cached data files are stored; see ?fcds_default_data_path for more information.

You can also list the files currently stored in the data cache with fcds_cache_ls() and you can clean outdated files with fcds_cache_clean(), which by default retains the most recently imported FCDS data file.

Example Data

If you don't have access to the FCDS data, we have provided a small example data set with similar properties to the FCDS data, but without any overlap in the attributes of cases.

fcds <- fcds::fcds_example

fcds[1:5, ]

Looking up FCDS Values

FCDS Constants

Lookup the values of variables in the imported data (referred to as a labels of the original data value) with fcds_const()

fcds_const(NULL)

fcds_const("cancer_status")

If you supply full = TRUE, the original column name (name_original) and value from the FCDS data will be reported alongsize the cleaned column name (name_clean) and the value label used by the fcds package.

fcds_const("cancer_site_group", full = TRUE)

FCDS Variables

All of the variables in the imported FCDS data are documented in ?fcds_import. You can also list common groups of fcds variables using fcds_vars().

fcds_vars("demo")

fcds_vars("cancer")

These column names can be used in conjunction with dplyr::select(), or you can supply FCDS data to fcds_vars() using the .data argument to select matching columns.

fcds %>% select(fcds_vars("id", "demo"))

fcds_vars(.data = fcds, "id", "icdo3")

A Worked Example: Prostate Cancer

Note: This data uses the fcds_example data and therefore bears no resemblance to reality. The examples herein are only relevant to the demonstration of the functions of the fcds package.

fcds_prostate <- 
  fcds %>% 
  # Filter to Prostate Cancer in two Florida counties
  filter(
    cancer_site_group == "Prostate Gland",
    county_name %in% c("Hillsborough", "Pinellas"),
    sex == "Male"
  ) %>% 
  # Count incidences of Prostate Cancer
  count_fcds(county_name, cancer_site_group) %>% 
  # Fill in missing age groups w.r.t. specified columns
  complete_age_groups(
    county_name, cancer_site_group, tidyr::nesting(year_group, year)
  ) %>% 
  # Calculate age-adjusted rate from incidence
  age_adjust() %>% 
  # Compute average yearly age-adjusted rate over 5 year range
  mutate(rate = rate / 5)

fcds_prostate

This data can readily be passed to ggplot2 for visualization.

library(ggplot2)

ggplot(fcds_prostate) +
  aes(year_group, rate, color = county_name, group = county_name) +
  geom_line() +
  geom_point()

GerkeLab/fcds documentation built on July 30, 2020, 7:04 p.m.