NestedCohort-package: Survival Analysis of Cohort Studies With Missing Covariate...
In NestedCohort: Survival Analysis for Cohorts with Missing Covariate Information

Description Details Author(s) References See Also Examples

NestedCohort fits Kaplan-Meier and Cox Models when you have missing covariate or strata data on a sample of a cohort. Missingness can be either by happenstance or by design (for example, the case-cohort and case-control within cohort designs). NestedCohort estimates standardized survival, survival differences, and attributable risks.

Package:	NestedCohort
Type:	Package
Version:	1.1-3
Date:	2012-12-12
License:	GPL (>= 2)

To fit Kaplan-Meier, use nested.km(). If you only want hazard ratios from a Cox model, used nested.coxph(). If you want standardized survival and attributable risk estimates, used nested.stdsurv().

Author: Hormuzd A. Katki

Maintainer: Hormuzd A. Katki <katkih@mail.nih.gov>

Katki, H.A. and Mark, S.D. Survival Analysis for Cohorts with Missing Covariate Information. The R Journal, 2008, 8(1), 14-9.

Mark, S.D. and Katki, H.A. Specifying and Implementing Nonparametric and Semiparametric Survival Estimators in Two-Stage (sampled) Cohort Studies with Missing Case Data. Journal of the American Statistical Association, 2006, 101, 460-471.

The survival package and the survey package

# Get zinc dataset
data(zinc)

# Fit and plot Kaplan-Meier
mod <- nested.km(survfitformula="Surv(futime01,ec01==1)~znquartiles",
                 samplingmod="ec01*basehist",data=zinc)
plot(mod,ymin=.6,xlab="Time (Days)",ylab="Survival",main="Survival by Quartile of Zinc",lty=1:4,)
legend(2000,0.7,c("Q1","Q2","Q3","Q4"),lty=1:4)

# Fit Cox model, get hazard ratios
coxmod <- nested.coxph(coxformula="Surv(futime01,ec01==1)~
          sex+agepill+smoke+drink+mildysp+moddysp+sevdysp+anyhist+zncent",
          samplingmod="ec01*basehist",data=zinc)
summary(coxmod)

# Fit Cox model, get and plot standardized survivals, survival differences, and attributable risks
mod <- nested.stdsurv(outcome="Surv(futime01,ec01==1)",
                      exposures="znquartiles",
                      confounders="sex+agestr+smoke+drink+mildysp+moddysp+sevdysp+anyhist",
                      samplingmod="ec01*basehist",exposureofinterest="Q4",plot=TRUE,
                      main="Time to Esophageal Cancer by Quartiles of Zinc",data=zinc)