GLMPolyModels: Modeling the Force of Infection Using Polynomial Models

In TeaKov/serostat: Modeling Infectious Disease Parameters Based on Serological and Social Contact Data R Package

Description

Models the (age-dependent) force of infection using polynomial generalized linear models, given input parameters and the specification of a model (see arguments). All models for the force of infection are cast into the well-known generalized linear model framework for binomial response data and are relying on different link-functions depending on the specified model.

Usage

GLMPolyModels(parameters, fun)

Arguments

parameters

Contains all input parameters: Age describes the population's age categories, Pos represent the number of seropositive individuals by age category, Tot are the total number of individuals in each specified age category, and Bin represents the response variable (binary, 1=past infection, 0=otherwise). One either specifies both the Pos and Tot parameters or the Bin parameter. For the Griffith's model it is required to specify a tau parameter, as the interval (0,τ) represents the period of inherited immunity and since more than 95% of the measles cases occur before the age of 10 years. The order of the polynomial of Grenfell's and Anderson's model can be specified by the order parameter (default=2).

fun

This specifies which polynomial model you want to use: Muench is the catalytic model of Muench, MuenchAlt is the alternative model of Muench, Griffiths is Griffiths' model in which the force of infection increases linearly by age, and GrenfellAnderson is Grenfell's and Anderson's model which extends the models of Muench and Griffiths to allow for higher-order polynomial functions representing force of infection. The Weibull model implies that the force of infection is a monotone function of exposure time, and consequently the infection times are Weibull distributed.

Examples

# Load Bulgarian Hepatitis A data
data("HAV_BUL_64")

# Prepare the parameters
params <- list(Tot=HAV_BUL_64$Tot, Pos=HAV_BUL_64$Pos, Age=HAV_BUL_64$Age)

# Muench's model
result <- GLMPolyModels(params, Muench)
summary(result$res)

# Muench's alternative model
result <- GLMPolyModels(params, MuenchAlt)
summary(result$res)

# Griffiths' model
params_griffiths <- c(params, tau=1)
result <- GLMPolyModels(params_griffiths, Griffiths)
summary(result$res)

# Grenfell's and Anderson's model
params_ga <- c(params, order=2)
result <- GLMPolyModels(params_ga, GrenfellAnderson)
summary(result$res)

# Load Belgian Hepatitis C data
data("HCV_BE_06_EXTRA")
data("HCV_BE_06")

# Prepare the parameters
params <- list(duration=HCV_BE_06_EXTRA$duration,
  infected=HCV_BE_06_EXTRA$infected, Tot=HCV_BE_06$V2,
  Pos=(HCV_BE_06$V2*HCV_BE_06$V4), Age=HCV_BE_06$V1)
result <- GLMPolyModels(params, Weibull)
summary(result$res)

TeaKov/serostat documentation built on May 21, 2019, 1:21 p.m.