bootComb

bootComb computes confidence intervals for parameters formed by combining independent parameter estimates by propagating uncertainty via parametric bootstrap sampling for epidemiological analyses.

Key Features:

• Parametric bootstrap sampling: Samples from probability distributions associated with each parameter estimate to propagate uncertainty and generate empirical distributions of combined parameters.
• Interval estimation: Computes empirical confidence intervals from bootstrap distributions using highest density intervals or quantiles (e.g., 95% intervals).
• Distribution fitting from reported CIs: Derives best-fit parameter distributions from reported confidence intervals and provides sampling functions for bootstrap resampling.
• Supported distributions: Includes auxiliary functions for beta, normal, exponential, gamma, Poisson, and negative binomial distributions for fitting and sampling.
• Combination of independent estimates: Supports combining multiple independent parameter estimates into a single metric while propagating uncertainty through resampling.

Scientific Applications:

• Adjustment of incidence rates: Adjusts incidence rates for healthcare utilization by combining independent estimates and deriving confidence intervals.
• Derivation of disease prevalence: Derives disease prevalence from conditional prevalences by combining independent parameter estimates and their uncertainties.
• Seroprevalence adjustment: Adjusts seroprevalence estimates for test sensitivity and specificity by combining independent estimates and computing confidence intervals.
• General epidemiological inference: Produces confidence intervals for combined parameters required in epidemiological and public health studies.

Methodology:

Performs parametric bootstrap sampling by sampling from fitted distributions for each parameter; derives best-fit distributions from reported confidence intervals for beta, normal, exponential, gamma, Poisson, and negative binomial distributions; and computes empirical intervals using highest density intervals or quantiles.