Skip to contents

Reproduction number from modelled incidence

Source: R/estimator-rt-incidence.R
rt_from_incidence.Rd

Calculate a reproduction number estimate from modelled incidence using the methods described in the vignette "Estimating the reproduction number from modelled incidence" and using a set of empirical generation time distributions. This assumes that modelled incidence has the same time unit as the ip distribution, and that this is daily, if this is not the case then rescale_model() may be able to fix it.

Usage

rt_from_incidence(
  df = i_incidence_model,
  ip = i_discrete_ip,
  raw = i_incidence_data,
  approx = TRUE,
  .progress = interactive()
)

Arguments

df

modelled incidence estimate - a dataframe with columns:

  • time (ggoutbreak::time_period + group_unique) - A (usually complete) set of singular observations per unit time as a `time_period`

  • incidence.fit (double) - an estimate of the incidence rate on a log scale

  • incidence.se.fit (positive_double) - the standard error of the incidence rate estimate on a log scale

  • incidence.0.025 (positive_double) - lower confidence limit of the incidence rate (true scale)

  • incidence.0.5 (positive_double) - median estimate of the incidence rate (true scale)

  • incidence.0.975 (positive_double) - upper confidence limit of the incidence rate (true scale)

Any grouping allowed.

ip

an infectivity profile (aka generation time distribution) - a dataframe with columns:

  • boot (anything + default(1)) - a bootstrap identifier

  • probability (proportion) - the probability of new event during this period.

  • tau (integer + complete) - the days since the index event.

Minimally grouped by: boot (and other groupings allowed).

A default value is defined.

raw

the raw data that the modelled incidence is based on. This is optional. If not given the algorithm will assume independence, which will be faster to estimate, but with more uncertain Rt estimates. In some circumstances this assumption of independence can cause underestimation of Rt. If this is a risk there will be a warning given, and this parameter may need to be supplied. - a dataframe with columns:

  • count (positive_integer) - Positive case counts associated with the specified time frame

  • time (ggoutbreak::time_period + group_unique) - A (usually complete) set of singular observations per unit time as a `time_period`

Any grouping allowed.

approx

use a faster, but approximate, estimate of quantiles

.progress

show a CLI progress bar

Value

A dataframe containing the following columns:

  • time (ggoutbreak::time_period + group_unique) - A (usually complete) set of singular observations per unit time as a time_period

  • rt.fit (double) - an estimate of the reproduction number

  • rt.se.fit (positive_double) - the standard error of the reproduction number

  • rt.0.025 (double) - lower confidence limit of the reproduction number

  • rt.0.5 (double) - median estimate of the reproduction number

  • rt.0.975 (double) - upper confidence limit of the reproduction number

Any grouping allowed.

Details

N.B. for certain estimators (e.g. poisson_gam_model(), poisson_locfit_model()) a version of this function will be called automatically if the infectivity profile is supplied. The inbuilt version is preferred over this function for these estimators, as the full covariance matrix may be used and the initial part of the outbreak can be predicted more accurately. This function is for supporting the other count models in ggoutbreak. If you are rolling your own incidence estimates and want an Rt estimate then rt_incidence_timeseries_implementation() maybe better suited to your task.

Examples


tmp = ggoutbreak::test_poisson_rt_smooth %>%
  poisson_locfit_model() %>%
  rt_from_incidence(ip = test_ip, approx=FALSE)
#> Rt from incidence: assuming independence. 
#> (N.B. this message will only be displayed once.)
#> Warning: Estimates were assumed to be independent, but more that 1% of estimates
#> are at risk of Rt underestimation by more that 0.05 (absolute).
#> We advise re-running supplying a full variance-covariance matrix, or
#> a value to the `raw` parameter, or setting `quick=FALSE`. 
#> (N.B. this warning will only be displayed once.)

tmp2 = ggoutbreak::test_poisson_rt_smooth %>%
  poisson_locfit_model() %>%
  rt_from_incidence(ip = test_ip, approx=TRUE)

plot_data = dplyr::bind_rows(
  tmp %>% dplyr::mutate(class = "exact"),
  tmp2 %>% dplyr::mutate(class = "approx"),
) %>% dplyr::group_by(class)

if (interactive()) {
  plot_rt(plot_data, date_labels="%b %y")+
   sim_geom_function(ggoutbreak::test_poisson_rt_smooth)+
   ggplot2::coord_cartesian(ylim=c(0.5,3))+
   ggplot2::facet_wrap(~class)
}