Lisa's comments from Como-model presentations - Githubissues

Como-DTC-Collaboration / como-models

Pedagogical materials for learning COVID-19 transmission dynamics modelling

https://Como-DTC-Collaboration.github.io/como-models/

MIT License

3 stars 0 forks source link

Lisa's comments from Como-model presentations #30

Closed ben18785 closed 2 years ago

ben18785 commented 3 years ago

SEIRD:

we want to include more text about the conditions that this might be appropriate for: specifically about natural birth and death
need to remove S0 from the R0 equation
rename R(0) as something else
proportion infected and recovered opposed to number of infected and recovered
express incidence as cases per 1000 of the population

SEIRD with age structure:

again add a caveat about natural birth and death; also mention that it would be easy to include the UN data on demographics which would be important if we wanted to run for longer
mention also that we would need to include aging in the model if we were running it for longer
put units on the axes
shouldn't the SEIRD with age be named SEIRDAge opposed to SEIRAge
we should probably go into infection fatality ratios for this model

Uncertainty in contact matrix:

add in a comment about contact matrices being representative of the whole country; mention about the granularity of data

Hui Jia: optimisation:

It's worth indicating that we are not including reporting fraction
Make the axis limits the same for the profile likelihood plots

Numerical solvers:

Might be worth talking about how variable step size solvers work

Ioana: SEIRDV

Make axis fonts larger
Scenario analysis: try starting vaccination at different points in the epidemic
Mention in the vignette that we may be in fact vaccinating E, I and Rs- Produce some graphs with interventions on / off

Idil: SEIRD with birth and natural death and the non-pharmaceutical interventions

Replenishments for Covid happen at a faster rate than how births and deaths influence things: important to caveat this in the vignette
Again would be good to do an intervention analysis, I think: turn on vs off

Liangti: GUI

Add a disclaimer about "don't blame me if you use this to make a decision and it's bad". Have a look at the Como app for this.
Change "State" to "Model Compartments"
Lisa to share a "building up complexity" pres for malaria

ben18785 commented 2 years ago

Addressed by #38 which was merged in #46