trial_setting: Simulate incidence in a trial setting.

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

Description

Generate a data frame representing a screening trial of arm.size individuals in each arm and record relevant disease diagnosed with and without screening and overdiagnoses.

Usage

trial_setting(arm.size = 50000, onset.rate = 0.001, sojourn.min = 0,
  sojourn.max = 6, sensitivity = 0.5, attendance = 0.8,
  overdiag.rate = 0.25, screen.start.year = 1, screen.stop.year = 30,
  followup.years = 30)

Arguments

arm.size	Number of individuals in each trial arm.
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.
attendance	Proportion of participants who attend a screen test.
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

population_setting

Examples

library(plyr)
library(reshape)
trial_incidence_3types <- function(){
    tset <- data.frame(sojourn.min=0,
                       sojourn.max=6,
                       sensitivity=0.5,
                       attendance=0.8,
                       overdiag.rate=0.25)
    tset <- rbind(transform(tset, screen.start.year=1,
                                  screen.stop.year=5,
                                  type='screen-stop'),
                  transform(tset, screen.start.year=1,
                                  screen.stop.year=30,
                                  type='screen-continue'),
                  transform(tset, screen.start.year=5,
                                  screen.stop.year=30,
                                  type='screen-continue-delay'))
    tset <- ddply(tset,
                  .(sojourn.min,
                    sojourn.max,
                    sensitivity,
                    attendance,
                    overdiag.rate,
                    screen.start.year,
                    screen.stop.year,
                    type),
                  function(x)
                  with(x, trial_setting(sojourn.min=sojourn.min,
                                        sojourn.max=sojourn.max,
                                        sensitivity=sensitivity,
                                        attendance=attendance,
                                        overdiag.rate=overdiag.rate,
                                        screen.start.year=screen.start.year,
                                        screen.stop.year=screen.stop.year))
                  )
    newscreen <- subset(tset, arm == 'screen' & type == 'screen-continue')
    newcontrol <- subset(tset, arm == 'screen' & type == 'screen-continue-delay')
    newscreen <- transform(newscreen, type='screen-continue-control-start')
    newcontrol <- transform(newcontrol, arm='control', type='screen-continue-control-start')
    tset <- subset(tset, type != 'screen-continue-delay')
    tset <- rbind(tset, newscreen, newcontrol)
    return(tset)
}
tset_3types <- trial_incidence_3types()

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