testing_bias_distribution

Models to understand dynamics of test positivity under epidemic dynamics and biased testing

• web version
• beta notes
• simple interaction model
• RTMBode/fitode example: not directly relevant, but hopefully a useful example of how to do trajectory-matching in fitode or RTMBode

Related projects. It's not super-clear we have any of sufficient interest, but JD is willing to share. We have a mathy publication with an earlier post-doc (Ali Gharouni, nice guy, now at PHO); Mike Li did some stuff during Covid which probably had some ideas but will be hard to figure out; JD advised some South African students who have a public repo, but it's not clear how far they got.

files

• Notes_Nov10.Rmd: @RichardSichengZhao 's summary notes for progress
• corrCheck.R: numerical calculations of correlation showing non-monotonicity
• Expected_Test_positivity_figure.R: beta model's test positivity as a function of testing focus/dispersion $\phi$, testing proportion t, prevalence i
• simple.md: basic notes on multiplex testing and cross-effects of diseases on each others' positivity/number of positive tests
• testing_funs.R: basic machinery for computing expected positivity based on the Beta model and Hazard Ratio Model
• testing_distrib.rmd: description and exploration of properties of testing_funs.R
• Logspace_comparing_methods.R: the file used to generate the pictures for log-space comparing 4 different methods of calculating positivity for Beta model.
• Qbeta_Issues.R: investigation of the Qbeta & qbeta issues.
• Single_variable_SIR.R: a numerical simulation of simple version SIR trajectory (fitODE later).
• inc_testing_positivity-ratio.R: As discussed, this file is used to generate figures for ratio of inc/test_positivity_proportion under different combination of inc, phi and testing_proportion.
• Average_i-phi_idea.R: RZhao 's Unsuccessful experiments. Will try something else but please ignore this for now.
• HR_Test_positivity_figure: Hazard Ratio model's test positivity curves as a function of testing Risk offsets \Phi, testing proportion T, prevalence V

In Pix

(TO DO)