transph: Semiparametric regression models for infectious disease...
In ekenah/transtat: Statistical Methods for Infectious Disease Transmission

Description Usage Arguments Value Author(s)

Fits semiparametric pairwise regression models for infectious disease transmission using right-censored and/or left-truncated data on contact intervals in ordered pairs of individuals and infectious contact from external sources with individuals. Uses the Cox relative risk function. External sources of infection are handled by stratifying on an external row indicator.

transph(
  formula,
  sus,
  data,
  weights,
  subset,
  na.action,
  spline_df,
  itermax = 25,
  L1tol = 1e-04,
  ...
)

`formula`	A formula of the form "response ~ terms". The response must be an object returned by `Surv` of type "right" or "counting". The formula can include `offset` terms for fixed coefficients, and an external row indicator can be included in a `strata` term.
`sus`	A string giving the name of the variable in `data` that contains the susceptible member of each pair.
`data`	A data frame containing the variables named in `formula`.
`weights`	A vector of infector probabilities for the data rows. If who-infected-whom is not completely observed, these will be updated in an EM algorithm. If missing, all possible infectors of each susceptible are given equal initial weights. If who-infected-whom is observed, the weights are passed to `coxph`, where they will be used as case weights. Normally, there would be no weights in this case.
`subset`	An expression indicating which rows of `data` should be included in the model fit.
`na.action`	A missing-data filter applied to `data` after `subset`. Defaults to `options()$na.action`.
`itermax`	The maximum number of EM algorithm iterations if who- infected-whom is not completely observed. Setting `itermax = 1` will cause the model to to use only the initial weights (default or provided).
`L1tol`	The EM algorithm stops when the L1 distance between the old and new weights is below `L1tol` per transmission event.
`...`	Further arguments to be passed to `coxph`, such as `init`, `ties`, or `control`.
`method`	The method used to generate a monotonic cubic interpolating spline for the cumulative hazard. Options are `monoH.FC` and `hyman`. See documentation for `splinefun`.

A list of class transph with the following components:

call: The call to transph with complete formal arguments.
coefficients: The estimated coefficients.
coxph: The coxph object returned by the final Cox regression.
data: The data used to fit the model.
df: The number of estimated coefficients.
formula: The model formula.
iter: The number of iterations. It is one for a model in which who-infects-whom is completely observed. Otherwise, it is the number of EM algorithm iterations (including the initial model).
L1tol: The L1 tolerance per transmission event used to halt the EM algorithm.
loglik: The loglik element in the coxph.object returned by the final Cox regression, which contains the log partial likelihood at the initial coefficient values and the fitted coefficient values. For a null model, only the first element is present. This is NA when who-infected-whom is not completely observed.
smooth_method: The method element in the coxph.object returned by the final Cox regression, which is the name of approximation used to handle ties.
spline_df: The degrees of freedom in the smoothing spline used to calculate the contact interval hazard function(s).
sus: The vector identifying the susceptible member of each #' pair in the data used to fit the model.
var: The covariance matrix for the coefficient estimates. #' The rows and columns are named for the corresponding covariates.
weights: The final vector of estimated infector probabilities.
ymat: The outcome matrix returned by Surv