Table 1.
A comparison of ern with other packages built to estimate
from epidemiological data.
Checkmarks (✓) indicate the presence of a feature and crosses (×) indicate the absence. A cross with an asterisk (×*) denotes a feature not built-in the package but technically possible though not straightforward for the average user (e.g., they may require additional modelling knowledge and/or the use of advanced/less documented features).
Fig 1.
Overview of the ern data pipeline to estimate .
Fig 2.
Fecal shedding distribution example.
A possible choice for the mean fecal shedding distribution used for SARS-CoV-2 wastewater data.
Fig 3.
Output of the function plot_diagnostic_ww.
The top panel shows the wastewater concentration data used as input (step line) along with the smoothed version of this time series (curve). The middle panel represents the daily incidence inferred from the smoothed wastewater concentration data (using the Richardson-Lucy deconvolution algorithm). The grey band gives a confidence band reflecting the uncertainty associated with the fecal shedding distribution. (The confidence width is set with prm.R$CI.) The estimated incidence is proportional to the parameter scaling.factor, here assumed equal to 1. The bottom panel shows the mean estimates (solid line), along with a 95% confidence interval (grey band) reflecting various sources of uncertainty. The horizontal dashed line represents the
threshold value of 1, which is epidemiologically important.
Fig 4.
Output of the function plot_diagnostic_cl.
The top panel shows the observed case report data used as input. The second panel from the top shows daily reports, smoothed and, in this case, inferred from the input aggregate (weekly) reports. When this inference is made, this panel also summarises the ensemble of daily report time series with a grey band, whose limits are given by the 2.5% and 97.5% percentiles by day. The second panel from the bottom appears only in the case where the input data is coarser than daily and compares the observed (aggregate) reports (black points) to aggregates from inferred daily reports (red points with 95% confidence bars), so that the user can check whether inferred daily reports are plausible against the input data. The bottom panel shows the mean estimates (solid line), along with a 95% confidence interval (grey band) reflecting various sources of uncertainty. The horizontal dashed line represents the
threshold value of 1, which is epidemiologically important.
Table 2.
Sample computing times (in seconds) for Rt estimates using different R packages.
The wastewater data is taken from the data set shipped with the package, which consists of four months of daily SARS-CoV-2 concentration measurements for the city of Zurich. The clinical data are simulated weekly reports. See S6 File for more details.