population_setting: Simulate incidence under simple dissemination of screening in...

In roman-gulati/overdiag: A package to investigate empirical estimates of overdiagnosis

Description

Generate a data frame representing a population of pop.size individuals and record relevant disease diagnosed with and without screening and overdiagnoses under simple dissemination of screening.

Usage

population_setting(pop.size = 1e+05, onset.rate = 0.001, sojourn.min = 0,
  sojourn.max = 6, sensitivity = 0.5, overdiag.rate = 0.25,
  screen.start.year = 4, screen.stop.year = 30, followup.years = 30)

Arguments

pop.size	Number of individuals in the simulated population.
onset.rate	Annual incidence rate of relevant preclinical disease.
sojourn.min	Shortest relevant preclinical duration.
sojourn.max	Longest relevant preclinical duration.
sensitivity	Screen test episode sensitivity.
overdiag.rate	Proportion of screen detections that are overdiagnosed.
screen.start.year	Year of follow-up at which screening starts.
screen.stop.year	Year of follow-up at which screening stops.
followup.years	Number of years of follow-up.

Value

A data frame of simulated disease incidence organized by year of preclinical onset, sojourn time, and year of diagnosis.

See Also

trial_setting

Examples

library(plyr)
library(reshape)
population_incidence_3ways <- function(pop.size=1e5,
                                       screen.start.year=3){
    dset <- data.frame(sojourn.min=c(0, 0),
                       sojourn.max=c(12, 6),
                       sensitivity=0.5,
                       overdiag.rate=0.25)
    pset <- ddply(dset,
                  .(sojourn.min,
                    sojourn.max,
                    sensitivity,
                    overdiag.rate),
                  function(x)
                  with(x, population_setting(pop.size=pop.size,
                                             sojourn.min=sojourn.min,
                                             sojourn.max=sojourn.max,
                                             sensitivity=sensitivity,
                                             overdiag.rate=overdiag.rate,
                                             screen.start.year=screen.start.year))
                  )
    return(pset)
}

multipopulation_incidence_3ways <- function(pop.size=1e5,
                                            dissemination.pattern='default'){
    if(dissemination.pattern == 'default'){
        proportion <- c(0.05, 0.1, 0.15, 0.15, 0.05, 0.5)
        start.year <- c(2, 3, 4, 5, 6, 28)
        dissemination.name <- 'Cumulative uptake (years 2-6): 5%,15%,30%,45%,50%'
    } else {
        proportion <- c(0.1, 0.3, 0.1, 0.5)
        start.year <- c(2, 3, 4, 28)
        dissemination.name <- 'Cumulative uptake (years 2-4): 10%,40%,50%'
    }
    dissemination <- data.frame(pop.size=pop.size,
                                proportion=proportion,
                                start.year=start.year)
    stopifnot(with(dissemination, sum(proportion)) == 1)
    mpset <- ddply(dissemination,
                   .(proportion, start.year),
                   function(x){
                       with(x,
        population_incidence_3ways(pop.size=proportion*pop.size,
                                   screen.start.year=start.year))
                         })
    mpset <- ddply(mpset,
                   .(sojourn.min,
                     sojourn.max,
                     sensitivity,
                     overdiag.rate,
                     year),
                   summarize,
                   count_screen=sum(count_screen),
                   count_clinical=sum(count_clinical),
                   count_overdiag=sum(count_overdiag))
    mpset <- transform(mpset, dissemination=dissemination.name)
    return(mpset)
}
mpset_default <- multipopulation_incidence_3ways(dissemination.pattern='default')
mpset_variant <- multipopulation_incidence_3ways(dissemination.pattern='variant')
mpset_3ways <- rbind(mpset_default, mpset_variant)

roman-gulati/overdiag documentation built on May 27, 2019, 1:49 p.m.