ddsynth

Bayesian hierarchical synthesis of incubation period distributions from individual-level data and published summary statistics.

Overview

The incubation period -- the interval between pathogen exposure and symptom onset -- is a critical parameter for quarantine policy and outbreak response, yet individual-level exposure data remain scarce in the published literature. For most pathogens, only summary statistics are available, and restricting inference to individual-level data alone would leave too few datasets for reliable between-study heterogeneity estimation.

ddsynth introduces a Bayesian hierarchical framework that jointly models individual-level observations and published summary statistics under a unified federated analysis approach. Key features:

• Accurately recovers incubation period distributions across a range of data availability scenarios
• Outperforms approaches that use summary data alone
• Quantifies between-study heterogeneity, including by outbreak country, pathogen variant, and exposure pathway
• Covers 18 pathogens with outbreak potential, spanning six pathogen groups

Dataset coverage

The package includes a curated dataset of incubation period studies spanning outbreaks worldwide:

Datasets cover six pathogen groups:

Prerequisites

R

R (≥ 4.0.0) is required. Download it from CRAN.

Stan

Several functions in this package use Stan via the rstan package. Stan requires a working C++ toolchain:

• Windows: Install Rtools and follow the RStan Getting Started guide.
• macOS: Install the Xcode Command Line Tools (xcode-select --install) and follow the RStan Getting Started guide.
• Linux: Ensure g++ and make are available (e.g. sudo apt install build-essential on Debian/Ubuntu), then install rstan normally.

Install rstan from CRAN once the toolchain is ready:

install.packages("rstan")

Verify the installation works by running the built-in example:

example(stan_model, package = "rstan", run.dontrun = TRUE)

Installation

Install the development version of ddsynth directly from GitHub using either remotes or pak:

# Using remotes (install remotes first if needed)
install.packages("remotes")
remotes::install_github("cm401/ddsynth")

# Alternatively, using pak
install.packages("pak")
pak::pkg_install("cm401/ddsynth")

Quick start

library(ddsynth)

See the package vignettes for detailed usage examples:

browseVignettes("ddsynth")