adjlogistic <- function(formula,data,sens,spec,eps=1e-05, ...){
# Routine to do logistic regression corrected for known sens. or spec. of 
# the outcome. See Madger and Hughes (1997) Am. J. Epidem, 146:195-203.
#
# Creator: Jim Hughes
#
# Inputs:
#  formula = usual S-plus type formula for a logistic regression
#  data = data frame for evaluation of the formula
#  sens, spec = vectors giving the sensitivity and specificity of the outcome:
#               length should be the same as the row dimension of data
#  eps = stopping criteria for the EM procedure
#  ... = aditional arguments to model.frame
#
# Outputs:
#  result = a glm object with coefficients and variance adjusted for sens 
#           and spec
#
# Modified: 3/18/03
  tdata <- model.frame(formula,data, ...)
  keep <- match(row.names(data),row.names(tdata),nomatch=0)>0
  sens <- sens[keep] ; spec<- spec[keep]
  result <- glm(formula,tdata,family=binomial, ...)
  y <- result$y
  repeat{
    eb <- result$fitted.values
    beta <- result$coef
    EY <- ifelse(y==0,(1-sens)*eb/((1-sens)*eb+spec*(1-eb)),
                           sens*eb/(sens*eb+(1-spec)*(1-eb)))
    formula <- update(formula,EY ~ .)
    assign("EY",EY,frame=1)
    result <- glm(formula, tdata, family = binomial, ...)
# Put in any stopping criteria you want
    if (sqrt(sum((result$coef - beta)^2)) < eps) break
  }
  z <- model.matrix(formula,data=tdata)
  result$var <- solve(t(z)%*%(z*(eb*(1-eb) - EY*(1-EY))))
  result$sens <- sens
  result$spec <- spec
  result
}