In cancerpolicy/bcimodel: Model the impact of early detection and treatment interventions for breast cancer

---

title: "Compare old and new code for US incidence survival curves" output: html_document: number_sections: yes toc: yes toc_depth: 2 html_notebook: number_sections: yes toc: yes toc_depth: 2

---

Setup

#--------------------------------------------------------------------------------
# Packages
#--------------------------------------------------------------------------------
library(bcimodel)
library(knitr)

#--------------------------------------------------------------------------------
# Knitr
#--------------------------------------------------------------------------------
opts_chunk$set(echo=FALSE, message=FALSE, warning=FALSE)

New incidence

data(incratesf)

# Compute survival from incidence/mortality databases
newInc <- interpolate_cumsurv(incratesf, 
                          ratevar='Female.Rate.Per.100K',
                          country='United States')
newInc100 <- interpolate_cumsurv(incratesf, 
                          ratevar='Female.Rate.Per.100K',
                          country='United States',
                          maxage=100)

Old incidence

incfile <- '~/Documents/jbirnbau/archive/screentreat/data/bc_1975-1979_incidence.csv'
format_clinical_incidence <- function(
    incfile
        ### Path to SEER clinical incidence table
) {

    # Read in data
    seer <- read.csv(incfile, stringsAsFactors=FALSE)
    seer <- within(seer, {
                   Age <- gsub('[0-9]+=', '', Ages)
                   Age <- as.character(gsub(' years', '', Age))
                   Age[Age=='85+'] <- 87
                   Age <- suppressWarnings(sapply(strsplit(Age, '-'),
                                 function(x) mean(as.numeric(x))))
            })
    seer <- seer[!is.nan(seer$Age),]
    seer <- seer[!is.na(seer$Age),]

    age_max <- max(seer$Age)
    incidence <- 
        data.frame(age = 0:age_max, 
                   incidence_rate = with(seer, 
                                          approx(Age, 
                                                 Crude.Rate, 
                                                 xout=0:age_max))$y/100000)

    incidence$incidencefree_rate <- 1-incidence$incidence_rate
    incidence_free_survival <- c()
    for(i in 1:(age_max+1)){
      incidence_free_survival[i] <- 
          prod(incidence$incidencefree_rate[1:i])
    }
    incidence$incidencefree_survival <- incidence_free_survival
    rm(incidence_free_survival)

    return(incidence)
}
oldInc <- format_clinical_incidence(incfile)

Compare rates

Compare survivals

plot(newInc100$age, newInc100$cumsurv, col='black', type='l')
lines(newInc$age, newInc$cumsurv, col='red')
lines(oldInc$age, oldInc$incidencefree_survival, col='green')

cancerpolicy/bcimodel documentation built on June 30, 2019, 12:39 a.m.