This has brought across Charlie's param changes along with the triggering changes:
Update to the parameter values squire is using to incorporate:
Up-to-date estimates of the probability of being hospitalised (prob_hosp) and the probability of requiring an ICU bed once hospitalised (prob_severe), derived from Salje et al. (https://science.sciencemag.org/content/369/6500/208).
Up-to-date estimates of the infection fatality ratio (which along with prob_hosp and prob_severe is used to define the parameters specifying probability of dying within hospital prob_non_severe_death_treatment and prob_severe_death_treatment). These estimates are derived from the recent Imperial College report (https://www.imperial.ac.uk/mrc-global-infectious-disease-analysis/covid-19/report-34-ifr/, submitted for peer review).
In terms of code changes, this PR does the following:
Create a new data object elderly_pop.rda which contains information on the 80+ age-group breakdown (80-84, 85-89 and 90+), required as our IFR estimates are now in 5 year age bands up to 90+; and an associated function get_elderly_population that extracts that information for a single specified country.
A new function parse_country_IFR which adjusts the value of prob_non_severe_death_treatment and prob_severe_death_treatment for the oldest age-group (80+) based on the age-composition of this class (i.e. proportion of individuals in 80-84, 85-89 and 90+, as specified in elderly_pop.rda.
A small number of new tests to ensure get_elderly_population and parse_country_IFR behave as we'd expect them to.
TODO:
[x] Update projections and t0_variables to get all tt_<variables>, such as tt_dur_get_mv_survive etc so these are propogated correctly.
[x] Finish NEWS changes for this.
[x] Update the documentation surrounding parameter values
[X] Add in somewhere the old probabilities back in (for ease of comparison)
- [] Add in time varying CFR probabilities for the dexamethasone etc (next version)
This has brought across Charlie's param changes along with the triggering changes:
Update to the parameter values squire is using to incorporate:
prob_hosp) and the probability of requiring an ICU bed once hospitalised (prob_severe), derived from Salje et al. (https://science.sciencemag.org/content/369/6500/208).prob_hospandprob_severeis used to define the parameters specifying probability of dying within hospitalprob_non_severe_death_treatmentandprob_severe_death_treatment). These estimates are derived from the recent Imperial College report (https://www.imperial.ac.uk/mrc-global-infectious-disease-analysis/covid-19/report-34-ifr/, submitted for peer review).In terms of code changes, this PR does the following:
elderly_pop.rdawhich contains information on the 80+ age-group breakdown (80-84, 85-89 and 90+), required as our IFR estimates are now in 5 year age bands up to 90+; and an associated functionget_elderly_populationthat extracts that information for a single specified country.parse_country_IFRwhich adjusts the value ofprob_non_severe_death_treatmentandprob_severe_death_treatmentfor the oldest age-group (80+) based on the age-composition of this class (i.e. proportion of individuals in 80-84, 85-89 and 90+, as specified inelderly_pop.rda.get_elderly_populationandparse_country_IFRbehave as we'd expect them to.TODO:
[x] Update
projectionsandt0_variablesto get alltt_<variables>, such astt_dur_get_mv_surviveetc so these are propogated correctly.[x] Finish NEWS changes for this.
[x] Update the documentation surrounding parameter values
[X] Add in somewhere the old probabilities back in (for ease of comparison)
- [] Add in time varying CFR probabilities for the dexamethasone etc(next version)