Comparison with other methods

[epiverse-trace]

Comparison with other methods

[adamkucharski]

Script comparing timeseries estimation with fixed CFR and estimation based on known outcomes in linelist, using global.health 2022 data as example: https://github.com/adamkucharski/ebola-cfr/blob/main/scripts/script_2022.R

[adamkucharski]

In terms of adjusting for truncation, estimating CFR based only on cases in linelist with known outcomes should produce an unbiased estimate if the delay from onset-to-outcome is the same for fatal and non-fatal outcomes (because we are calculating the ratio of two variables sampled from the early part of the two same distributions).

[adamkucharski]

For coarseDataTools, EMforCFR() might be a useful starting point for comparison with Ebola data.

Update: built in CFR functions won't be relevant for our specific problem, because EMforCFR estimates relative CFR between groups using survival analysis (rather than an absolute population CFR like our package approach).

[pratikunterwegs]

Closing as this is now covered conceptually in the Readme.