paf.confidence.one2one: Approximate Confidence Intervals for the Population...

In INSP-RH/pif: Potential Impact Fraction and Population Attributable Fraction for Cross-Sectional Data

Description

Function that calculates approximate confidence intervals of the Population Attributable Fraction paf considering a one to one Relative Risk rr with unidimensional theta parameter values

Usage

paf.confidence.one2one(X, thetahat, rr, thetalow, thetaup,
  weights = rep(1/nrow(as.matrix(X)), nrow(as.matrix(X))), confidence = 95,
  confidence_theta = 99, check_thetas = TRUE, deriv.method.args = list(),
  deriv.method = c("Richardson", "complex"), method = c("empirical",
  "approximate"), Xvar = var(X), check_exposure = TRUE, check_rr = TRUE,
  check_integrals = TRUE)

Arguments

X	Random sample (data.frame) which includes exposure and covariates.
thetahat	Estimative of theta for the Relative Risk function
rr	Function for Relative Risk which uses parameter theta. The order of the parameters shound be rr(X, theta). **Optional**
thetalow	Lower bound of the confidence interval.
thetaup	Upper bound of the confidence interval.
weights	Survey weights for the random sample X.
confidence	Confidence level % (default: 95)
confidence_theta	Confidence level % of theta corresponding to the interval [thetalow, thetaup] (default: 99%).
check_thetas	Check that thetas are correctly specified
deriv.method.args	method.args for hessian.
deriv.method	method for hessian. Don't change this unless you know what you are doing.
method	Either "empirical" (default) or "approximate".
Xvar	Variance of exposure levels (for "approximate" method).
check_exposure	Check that exposure X is positive and numeric
check_rr	Check that Relative Risk function rr equals 1 when evaluated at 0
check_integrals	Check that counterfactual and relative risk's expected values are well defined for this scenario

Author(s)

Rodrigo Zepeda-Tello [e-mail]

Dalia Camacho-García-Formentí [e-mail]

Examples

#Example 1: Exponential Relative Risk
#--------------------------------------------
set.seed(18427)
X <- data.frame(rnorm(1000, 3,.7))
thetahat <- 0.4
thetalow <- 0.1
thetaup  <- 0.7
paf.confidence.one2one(X, thetahat, function(X, theta){exp(theta*X)}, 
thetalow, thetaup)

#Approximate method:
Xmean <- data.frame(mean(X[,1]))
Xvar  <- var(X[,1])
paf.confidence.one2one(Xmean, thetahat, function(X, theta){exp(theta*X)}, 
thetalow, thetaup, Xvar = Xvar, method = "approximate")

#Example 2: Multivariate example
#--------------------------------------------
set.seed(18427)
X1 <- rnorm(1000,3,.7)
X2 <- rnorm(1000,3,.7)
X  <- data.frame(X1,X2)
thetahat <- c(0.12, 0.03)
thetalow <- c(0.05, 0.01)
thetaup  <- c(0.25, 0.06)
rr <- function(X, theta){exp(theta[1]*X[,1] + theta[2]*X[,2])}
paf.confidence.one2one(X, thetahat, rr, thetalow, thetaup) 

#Approximate method:
Xmean <- data.frame(t(colMeans(X)))
Xvar  <- var(X)
paf.confidence.one2one(Xmean, thetahat, 
function(X, theta){exp(theta[1]*X[,1] + theta[2]*X[,2])}, 
thetalow, thetaup, 
Xvar = Xvar, method = "approximate")

INSP-RH/pif documentation built on May 7, 2019, 6:01 a.m.