Efficient odds ratio estimation under two-phase sampling using error-prone data from a multi-national HIV research cohort

Lotspeich, Shepherd, Amorim, Shaw, and Tao

The complete R package logreg2ph and code for the simulation settings included in the paper (https://doi.org/10.1111/biom.13512).

Install

To install the package, run the following in your R console:

devtools::install_github("sarahlotspeich/logreg2ph")

Note to Mac users:

This package employs Rcpp and RcppArmadillo, and it seems that for some Mac users there is conflict between these packages and Anaconda. If you experience errors installing the package, an R-only version is available on this repo.

Simulation settings

Inside the simulations subdirectory, you will find the following:

• Table1_SimSetup.R: simulations with outcome misclassification and a binary error-prone covariate, intended to inspect increasing Phase I and Phase II sample sizes
• Tables2&3_SimSetup.R: Simulations with outcome misclassification and a continuous covariate with additive errors, intended to inspect varied error variance
• Table4_SimSetup.R: Simulations with outcome misclassification and a continuous covariate with additive errors, intended to inspect varied outcome error rates
• TableS2_SimSetup.R: Simulations with outcome misclassification and binary error-prone covariate, intended to inspect robustness of the fully-parametric MLE and proposed SMLE under complex covariate error
• TableS3_SimSetup.R: Simulations with outcome misclassification and binary error-prone covariate, intended to inspect robustness of the fully-parametric MLE and proposed SMLE under complex outcome and covariate error
• TableS4_SimSetup.R: Simulations with continuous covariate with additive errors (classical measurement error), intended to compare performance of proposed SMLE to traditional regression calibration (RC) approach
• TableS5_SimSetup.R: Simulation results for the naive estimator under outcome misclassification and a continuous covariate under varied parameterizations
• TableS6_SimSetup.R: Simulations with outcome misclassification and continuous covariate with differentially biased additive errors, intended to demonstrate robustness of the proposed SMLE to different covariate error types
• TableS7_SimSetup.R: Simulations with outcome misclassification and continuous covariate with differential multiplicative errors, intended to demonstrate robustness of the proposed SMLE to different covariate error types

Generating Data

Inside each of the files above, you will find code to generate the appropriate data for that simulation setting, e.g.,

set.seed(918)

# Set sample sizes ----------------------------------------
N <- 1000 # Phase I size = N
n <- 250 # Phase II/audit size = n

# Generate true values Y, Xb, Xa ----------------------------
Xa <- rbinom(n = N, size = 1, prob = 0.25)
Xb <- rbinom(n = N, size = 1, prob = 0.5)
Y <- rbinom(n = N, size = 1, prob = (1 + exp(-(- 0.65 - 0.2 * Xb - 0.1 * Xa))) ^ (- 1))

# Generate error-prone Xb* from error model P(Xb*|Xb,Xa) ------
sensX <- specX <- 0.75
delta0 <- - log(specX / (1 - specX))
delta1 <- - delta0 - log((1 - sensX) / sensX)
Xbstar <- rbinom(n = N, size = 1, prob = (1 + exp(- (delta0 + delta1 * Xb + 0.5 * Xa))) ^ (- 1))

# Generate error-prone Y* from error model P(Y*|Xb*,Y,Xb,Xa) ---
sensY <- 0.95
specY <- 0.90
theta0 <- - log(specY / (1 - specY))
theta1 <- - theta0 - log((1 - sensY) / sensY)
Ystar <- rbinom(n = N, size = 1, prob = (1 + exp(- (theta0 - 0.2 * Xbstar + theta1 * Y - 0.2 * Xb - 0.1 * Xa))) ^ (- 1))

Then, the user has the option of two audit designs: simple random sampling (SRS) or 1:1 case-control sampling based on $Y^*$ (naive case-control). Based on these designs, the validation indicators V are generated as follows:

# Choose audit design: SRS or -----------------------------
## Naive case-control: case-control based on Y^* ----
audit <- "SRS" #or "Naive case-control"

# Draw audit of size n based on design --------------------
## V is a TRUE/FALSE vector where TRUE = validated --------
if(audit == "SRS")
{
    V <- seq(1, N) %in% sample(x = seq(1, N), size = n, replace = FALSE)
}
if(audit == "Naive case-control")
{
    V <- seq(1, N) %in% c(sample(x = which(Ystar == 0), size = 0.5 * n, replace = FALSE),
                          sample(x = which(Ystar == 1), size = 0.5 * n, replace = FALSE))
}

Finally, combine the generated data and validation indicators into an analytical dataset:

# Build dataset --------------------------------------------
sdat <- cbind(Y, Xb, Ystar, Xbstar, Xa, V)
# Make Phase-II variables Y, Xb NA for unaudited subjects ---
sdat[!V, c("Y", "Xb")] <- NA

Running Estimator Code

The R scripts each contain implementations for the estimators discussed in the paper. Examples of each are provided below:

1. Naive Analysis

naive <- glm(Ystar ~ Xbstar + Xa, family = "binomial", data = data.frame(sdat))
beta_naive <- naive$coefficients[2]
se_naive <- sqrt(diag(vcov(naive)))[2]

2. Complete-Case Analysis

cc <- glm(Y[V] ~ Xb[V] + Xa[V], family = "binomial")
beta_cc <- cc$coefficients[2]
se_cc <- sqrt(diag(vcov(cc)))[2]

3. Horvitz--Thompson Estimator (for Naive Case-Control Audit Only)

library(sandwich)
if (audit == "Naive case-control") {
  sample_wts <- ifelse(Ystar[V] == 0, 1 / ((0.5 * n) / (table(Ystar)[1])), 1 / ((0.5 * n) / (table(Ystar)[2])))
  ht <- glm(Y[V] ~ Xb[V] + Xa[V], family = "binomial",
            weights = sample_wts)
  beta_ht <- ht$coefficients[2]
  se_ht <- sqrt(diag(sandwich(ht)))[2]
}

4. Generalized Raking Estimator

### Influence function for logistic regression
### Taken from: https://github.com/T0ngChen/multiwave/blob/master/sim.r
inf.fun <- function(fit) {
  dm <- model.matrix(fit)
  Ihat <- (t(dm) %*% (dm * fit$fitted.values * (1 - fit$fitted.values))) / nrow(dm)
  ## influence function
  infl <- (dm * resid(fit, type = "response")) %*% solve(Ihat)
  infl
}

naive_infl <- inf.fun(naive) # error-prone influence functions based on naive model
colnames(naive_infl) <- paste0("if", 1:3)

# Add naive influence functions to sdat -----------------------------------------------
sdat <- cbind(id = 1:N, sdat, naive_infl)

# Calibrate raking weights to the sum of the naive influence functions ----------------
library(survey)
if (audit == "SRS") {
  sstudy <- twophase(id = list(~id, ~id),
                     data = data.frame(sdat),
                     subset = ~V)
} else if (audit == "Naive case-control") {
  sstudy <- twophase(id = list(~id, ~id),
                     data = data.frame(sdat),
                     strat = list(NULL, ~Ystar),
                     subset = ~V)
}
scal <- calibrate(sstudy, ~ if1 + if2 + if3, phase = 2, calfun = "raking")

# Fit analysis model using calibrated weights -----------------------------------------
rake <- svyglm(Y ~ Xb + Xa, family = "binomial", design = scal)
beta_rake <- rake$coefficients[2]
se_rake <- sqrt(diag(vcov(rake)))[2]

5. Maximum Likelihood Estimator (MLE) (for Binary Xb* Only)

# Script: two-phase log-likelihood specification adapted from Tang et al. (2015) named ~/code/Tang_twophase_loglik_binaryX.R
source("Tang_twophase_loglik_binaryX.R")
fit_Tang <- nlm(f = Tang_twophase_loglik,
                p = rep(0, 12),
                hessian = TRUE,
                Val = "V",
                Y_unval = "Ystar",
                Y_val="Y",
                X_unval = "Xbstar",
                X_val = "Xb",
                C = "Xa",
                data = sdat)
beta_mle <- fit_Tang$estimate[10]
se_mle <- sqrt(diag(solve(fit_Tang$hessian)))[10]

6. Sieve Maximum Likelihood Estimator (SMLE)

# Construct B-spline basis -------------------------------
nsieve <- 4
B <- matrix(0, nrow = N, ncol = nsieve)
B[which(Xa == 0 & Xbstar == 0), 1] <- 1
B[which(Xa == 0 & Xbstar == 1), 2] <- 1
B[which(Xa == 1 & Xbstar == 0), 3] <- 1
B[which(Xa == 1 & Xbstar == 1), 4] <- 1
colnames(B) <- paste0("bs", seq(1, nsieve))
sdat <- cbind(sdat, B)

library("logreg2ph")
smle <- logreg2ph(Y_unval = "Ystar",
                  Y_val = "Y",
                  X_unval = "Xbstar",
                  X_val = "Xb",
                  C = "Xa",
                  Validated = "V",
                  Bspline = colnames(B),
                  data = sdat,
                  noSE = FALSE,
                  MAX_ITER = 1000,
                  TOL = 1E-4)
beta_smle <- smle$Coefficients$Coefficient[2]
se_smle <- smle$Coefficients$SE[2]