In WetRobot/popEpi: Functions for Epidemiological Analysis using Population Data

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "README-"
)
library(popEpi)

popEpi: Epidemiology with population data

The purpose of popEpi is to facilitate computing certain epidemiological statistics where population data is used. Current main attractions:

Splitting, merging population hazards, and aggregating

the lexpand function allows users to split their subject-level follow-up data into sub-intervals along age, follow-up time and calendar time, merge corresponding population hazard information to those intervals, and to aggregate the resulting data if needed.

data(sire)
sr <- sire[1,]
print(sr)

x <- lexpand(sr, birth = bi_date, entry = dg_date, exit = ex_date,
             status = status %in% 1:2, 
             fot = 0:5, per = 1994:2000)
print(x)

data(popmort)
x <- lexpand(sr, birth = bi_date, entry = dg_date, exit = ex_date,
             status = status %in% 1:2, 
             fot = 0:5, per = 1994:2000, pophaz = popmort)
print(x)

a <- lexpand(sr, birth = bi_date, entry = dg_date, exit = ex_date,
             status = status %in% 1:2,
             fot = 0:5, per = 1994:2000, aggre = list(fot, per))
print(a)

SIRs / SMRs

One can make use of the sir function to estimate indirectly standardised incidence or mortality ratios (SIRs/SMRs). The data can be aggregated by lexpand or by other means. While sir is simple and flexible in itself, one may also use sirspline to fit spline functions for the effect of e.g. age as a continuous variable on SIRs.

data(popmort)
data(sire)
c <- lexpand( sire, status = status %in% 1:2, birth = bi_date, exit = ex_date, entry = dg_date,
              breaks = list(per = 1950:2013, age = 1:100, fot = c(0,10,20,Inf)), 
              aggre = list(fot, agegroup = age, year = per, sex) )

se <- sir( coh.data = c, coh.obs = 'from0to1', coh.pyrs = 'pyrs', 
           ref.data = popmort, ref.rate = 'haz', 
           adjust = c('agegroup', 'year', 'sex'), print = 'fot')
se

(Relative) survival

The survtab function computes observed, net/relative and cause-specific survivals as well as cumulative incidence functions for Lexis data. Any of the supported survival time functions can be easily adjusted by any number of categorical variables if needed.

One can also use survtab_ag for aggregated data. This means the data does not have to be on the subject-level to compute survival time function estimates.

library(Epi)

data(sibr)
sire$cancer <- "rectal"
sibr$cancer <- "breast"
sr <- rbind(sire, sibr)

sr$cancer <- factor(sr$cancer)
sr <- sr[sr$dg_date < sr$ex_date, ]

sr$status <- factor(sr$status, levels = 0:2, 
                    labels = c("alive", "canD", "othD"))

x <- Lexis(entry = list(FUT = 0, AGE = dg_age, CAL = get.yrs(dg_date)), 
           exit = list(CAL = get.yrs(ex_date)), 
           data = sr,
           exit.status = status)

st <- survtab(FUT ~ cancer, data = x,
              breaks = list(FUT = seq(0, 5, 1/12)),
              surv.type = "cif.obs")
st