How can we estimate the % of observations from a given surveillance system or in a certain outbreak response context?

[avallecam]

In Gostic et al. 2020, in the accounting for incomplete observations, authors share the factors that change the assumption of a constant fraction of observed infections.

In White et al 2019, in the Adjustment for changes in reporting fraction section, authors mention that changes in case ascertainment informed changes in Rt and R0 estimates.

Questions:

• From a workflow perspective, can we use {cfr} to estimate ascertainment and inform on the % of observations to add in the obs argument in epinow()?
• Do you have examples of study evaluations on designed surveillance systems that aim to estimate this % of expected observations?
• Or, do you usually expect to input data from other data sources?

[jamesmbaazam]

In Gostic et al. 2020, in the accounting for incomplete observations, authors share the factors that change the assumption of a constant fraction of observed infections.

In White et al 2019, in the Adjustment for changes in reporting fraction section, authors mention that changes in case ascertainment informed changes in Rt and R0 estimates.

Questions:

• From a workflow perspective, can we use {cfr} to estimate ascertainment and inform on the % of observations to add in the obs argument in epinow()?

I believe in theory, you can pass the result of {cfr} to scale in obs_opts() (with uncertainty around it) since any value you supply here is based on your assumptions/prior knowledge anyway. You must also be aware of the assumptions/limitations/interpretation of the results and underlying methods used by {cfr} to estimate ascertainment. For example, {cfr} uses "Normal approximation to binomial likelihood", from a quick glance at the linked vignette. @sbfnk Thoughts?

• Do you have examples of study evaluations on designed surveillance systems that aim to estimate this % of expected observations?

Are you looking for something like the following:

• https://bmcpublichealth.biomedcentral.com/articles/10.1186/1471-2458-14-147#Sec12
• https://www.imperial.ac.uk/mrc-global-infectious-disease-analysis/disease-areas/covid-19/report-31-syria/
• https://www.sciencedirect.com/science/article/pii/S0022519322003241#sec2

• Or, do you usually expect to input data from other data sources?

I'm not sure I understand this.