The first Re multiplier is multiplied by the basic R0 to give the new Re after the first intervention. The second Re multiplier is multiplied by the basic R0 and the first Re multiplier to give the new Re after the second intervention.
Re after third intervention =
R0 * intervention1.multiplier * intervention2.multiplier * intervention3.multiplier
You can have any number of interventions. The template Excel input comes with 3. If you want fewer, delete one or more rows. If you want more, just add rows using the same format.
You can decrease
iter on the Internal sheet to 500 or less. The results will be less reliable and you will probably get some warnings from Stan, but if you just want to test something out, this may work. You should increase
iter back to at least 1000 for final projections.
If there are a small number of divergent transitions (less than ~10), there is usually little impact on the projections ((but this is not guaranteed). If there are a larger number, increase
adapt_delta to 0.9 or 0.95 or 0.99 on the Internal sheet. Run time will be slower. We don’t recommnd using
pairs(). If increasing
adapt_delta does not help, there may be a problem with your parameter distributions or data (e.g. hospital census data is not compatible with cumulative hospitals).
max_treedepth to 13 or 15 or more on the Internal sheet. Run time will be slower.