R/diagnostics.R

In rprev: Estimating Disease Prevalence from Registry Data

# This function compares the actual variance of the yearly incidence rates with
# rates simulated from a Poisson process with overall rate equal to the overall
# mean rate.
#
# param inc Vector of absolute incidence values for each included year.
# param N_sim Number of simulations to perform.
# return Vector of p-values for over- and under-dispersion based on the
#   position of the observed sequence variance in the distribution.
test_dispersion <- function(inc, N_sim = 1e5) {
    var_sim <- vapply(seq(N_sim), function(i) var(rpois(length(inc), mean(inc))), numeric(1))
    c(sum(var_sim > var(inc))/N_sim, sum(var_sim <= var(inc))/N_sim)
}

#' Test simulated prevalence fit.
#'
#' Calculates a Chi squared test between predicted yearly contributions to
#' prevalence, and the observed values obtained from the registry, indicating
#' whether the simulated prevalence values are accurate.
#'
#' @param object A \code{prevalence} object.
#' @return P-value from a chi-squared test of difference between prevalence
#'   prediction and counted prevalence at the index date.
#' @examples
#' data(prevsim)
#'
#' \dontrun{
#'
#' obj <- prevalence(Surv(time, status) ~ age(age) + sex(sex) + entry(entrydate) + event(eventdate),
#'                   data=prevsim, num_years_to_estimate = c(5, 10), population_size=1e6,
#'                   start = "2005-09-01",
#'                   num_reg_years = 8, cure = 5)
#'
#' test_prevalence_fit(obj)
#' }
#' @import data.table
#'
#' @export
#' @family prevalence functions
test_prevalence_fit <- function(object) {
    # Needed for CRAN
    prev_registry <- NULL
    alive_at_index <- NULL
    incident_date <- NULL
    sim <- NULL

    object$simulated[, prev_registry := as.numeric(incident_date >= object$registry_start & alive_at_index)]
    predicted <- round(mean(object$simulated[, sum(prev_registry), by=sim][[2]]))
    poisson.test(c(object$counted, predicted))$p.value
}