Package 'sircovid' reference manual

Title:	SIR Model for COVID-19
Description:	Mechanistic models of transmission of the SARS-Cov-2 (COVID-19) virus implemented as stochastic compartmental models in odin and dust. Uses mcstate to perfom Bayesian evidence synthesis from several surveillance data streams through the estimation of transmission parameters.
Authors:	Marc Baguelin [aut, cre], Sangeeta Bhatia [ctb], Edward Knock [aut], Lilith Whittles [aut], Rich FitzJohn [aut], John Lees [aut], OJ Watson [ctb], Anne Cori [aut], Pablo Perez-Guzman [aut], Raphael Sonabend [aut], Yasin Elmaci [aut]
Maintainer:	Marc Baguelin <[email protected]>
License:	MIT + file LICENSE
Version:	0.15.1
Built:	2026-06-17 09:10:48 UTC
Source:	https://github.com/mrc-ide/sircovid

Add incidence to trajectories

Description

Augment (or remove) trajectories from a mcstate_trajectories or mcstate_pmcmc object.

Usage

add_trajectory_incidence(obj, states, suffix = "_inc")

drop_trajectory_incidence(obj)
add_trajectory_incidence(obj, states, suffix = "_inc")

drop_trajectory_incidence(obj)

Arguments

obj

A mcstate_trajectories or mcstate_pmcmc object to update

states

A vector of cumulative states to compute incidence for

suffix

A string to append to the input states to create the incidence variables

The basic sircovid model

Description

The "basic" sircovid model. This is a dust model.

Examples

# Set up the basic model for England with default parameters and
# an initial seeding of early February:
p <- basic_parameters(sircovid_date("2020-02-07"), "england")
mod <- basic$new(p, 0, 10)

# Set the initial state and index as we would us for a run
# (without setting an initial state there is no seeding)
initial <- basic_initial(mod$info(), 10, p)
mod$update_state(state = initial)
mod$set_index(basic_index(mod$info())$run)

# Run the model up to the end of march
step_end <- sircovid_date("2020-03-31") / p$dt

# The filtered state is returned at the end of the run
mod$run(step_end)

# More state can be retrieved using the "state" method
mod$state(1:6)
# Set up the basic model for England with default parameters and
# an initial seeding of early February:
p <- basic_parameters(sircovid_date("2020-02-07"), "england")
mod <- basic$new(p, 0, 10)

# Set the initial state and index as we would us for a run
# (without setting an initial state there is no seeding)
initial <- basic_initial(mod$info(), 10, p)
mod$update_state(state = initial)
mod$set_index(basic_index(mod$info())$run)

# Run the model up to the end of march
step_end <- sircovid_date("2020-03-31") / p$dt

# The filtered state is returned at the end of the run
mod$run(step_end)

# More state can be retrieved using the "state" method
mod$state(1:6)

Compare observed and modelled data for the basic model

Description

Compare observed and modelled data from the basic model. This conforms to the mcstate interface.

Usage

basic_compare(state, observed, pars)
basic_compare(state, observed, pars)

Arguments

state

State vector for the end of the current day. This is assumed to be filtered following basic_index() so contains rows corresponding to ICU and deaths.

observed

Observed data. This will be a list with elements icu (number of ICU beds occupied) and deaths (number of deaths over this day).

pars

A list of parameters, as created by basic_parameters()

Value

A vector of log likelihoods, the same length as the number of particles (the number of columns in the modelled state)

Examples

state <- rbind(icu = 10:15, deaths_inc = 1:6)
observed <- list(icu = 13, deaths = 3)
pars <- basic_parameters(sircovid_date("2020-02-07"), "england")
basic_compare(state, observed, pars)
basic_compare(state * 5, observed, pars)
state <- rbind(icu = 10:15, deaths_inc = 1:6)
observed <- list(icu = 13, deaths = 3)
pars <- basic_parameters(sircovid_date("2020-02-07"), "england")
basic_compare(state, observed, pars)
basic_compare(state * 5, observed, pars)

Index of basic model

Description

Index of "interesting" elements for the basic model. This function conforms to the mcstate interface.

Usage

basic_index(info)
basic_index(info)

Arguments

info

The result of running the ⁠$info()⁠ method on an initialised basic model

Value

A list with element run, indicating the locations of the ICU and D compartments.

Examples

p <- basic_parameters(sircovid_date("2020-02-07"), "england")
mod <- basic$new(p, 0, 10)
basic_index(mod$info())
p <- basic_parameters(sircovid_date("2020-02-07"), "england")
mod <- basic$new(p, 0, 10)
basic_index(mod$info())

Initial conditions for the basic model

Description

Create initial conditions for the basic model. This matches the interface required for mcstate

Usage

basic_initial(info, n_particles, pars)
basic_initial(info, n_particles, pars)

Arguments

info

The result of running the ⁠$info()⁠ method on an initialised basic model

n_particles

The number of particles required. Currently only uniform initial seeding is implemented so this has no effect

pars

A parameter list created by basic_parameters(); from this list we will use the population element.

Value

A numeric vector of initial conditions

Examples

p <- basic_parameters(sircovid_date("2020-02-07"), "england")
mod <- basic$new(p, 0, 10)
basic_initial(mod$info(), 10, p)
p <- basic_parameters(sircovid_date("2020-02-07"), "england")
mod <- basic$new(p, 0, 10)
basic_initial(mod$info(), 10, p)

Parameters for the basic model

Description

Parameters for the "basic" model.

Usage

basic_parameters(
  start_date,
  region,
  beta_date = NULL,
  beta_value = NULL,
  beta_type = "piecewise-linear",
  severity = NULL,
  exp_noise = 1e+06,
  initial_seed_size = 30,
  initial_seed_pattern = 1
)
basic_parameters(
  start_date,
  region,
  beta_date = NULL,
  beta_value = NULL,
  beta_type = "piecewise-linear",
  severity = NULL,
  exp_noise = 1e+06,
  initial_seed_size = 30,
  initial_seed_pattern = 1
)

Arguments

start_date

The start date, as a sircovid_date() (i.e., the number of days into 2020)

region

The region to run the model for. This will bee used to get population data, which is currently fixed within the package and is limited to "uk", the four constituent nations ("england", "wales", "scotland", "northern_ireland") and the 7 NHS regions (e.g., "midlands"). These names are case insensitive.

beta_date

A vector of dates (each as a sircovid_date()) for changes in beta (the contact rate parameter), or NULL if a single value is used for all times (see sircovid_parameters_piecewise_linear() or sircovid_parameters_piecewise_constant(), where this is passed as date). If beta_type = "piecewise-constant", then the first date must be 0.

beta_value

A vector of values for beta (the contact rate parameter). If not given, and if beta_date is NULL then a value of 0.1 will be used through the whole simulation, otherwise if beta_date is NULL this must be a scalar. If beta_date is given then beta_date and beta_value must have the same length (see sircovid_parameters_piecewise_linear() or sircovid_parameters_piecewise_constant(), where this is passed as value).

beta_type

The type of form used for beta (the contact rate parameter), which currently can be "piecewise-linear" or "piecewise-constant"

severity

Severity data, via Bob Verity's markovid package. This needs to be NULL (use the default bundled data version in the package), a data.frame object (for raw severity data) or a list (for data that has already been processed by sircovid for use). New severity data comes from Bob Verity via the markovid package, and needs to be carefully calibrated with the progression parameters.

exp_noise

Rate of exponential noise used in the compare function - typically set to a large value so that noise is small but non-zero. If set to Inf then there is no noise in the observation process (not realistic but useful for testing).

initial_seed_size

Initial size of seeding from the S to E compartment; all seeding is in the 15-19 year old group from start_date according to initial_seed_pattern. The default is 30.

initial_seed_pattern

A vector of seeding weights for the initial seeding. The length represents the number of steps to seed over from the start_date, and the initial_seed_size is split over these steps according to those weights. If start_date is not a multiple of the step size (and thus falls between two steps) then we weight over an additional step and adjust the weights according to how far the start_date is from the previous full step.

Value

A list of inputs to the model, many of which are fixed and represent data. These correspond largely to user() calls within the odin code, though some are also used in processing just before the model is run.

Examples

basic_parameters(sircovid_date("2020-02-01"), "uk")
basic_parameters(sircovid_date("2020-02-01"), "uk")

Check a model type

Description

Check that a model type is valid (i.e. in sircovid_models)

Usage

check_sircovid_model(sircovid_model)
check_sircovid_model(sircovid_model)

Arguments

sircovid_model

String

Combine Rt estimates

Description

Combine Rt across multiple runs.

Usage

combine_rt(rt, samples, rank = TRUE, weight = "infections_inc")
combine_rt(rt, samples, rank = TRUE, weight = "infections_inc")

Arguments

rt

A list of Rt calculations from lancelot_Rt_trajectories() (though any Rt calculation that confirms to this will work)

samples

A list of samples from mcstate::pmcmc

rank

A boolean deciding whether to rank trajectories by increasing incidence or not before combining Rt estimates

weight

A string specifying what trajectory to use for weighting, defaults to "infections_inc" (used for Rt calculation)

Value

A list of Rt output in the same structure as the first element of rt. All Rt estimates will be aggregated across regions (or whatever else you are aggregating on) based on the parameters in samples.

Combine Rt estimates from EpiEstim

Description

Combine Rt estimates from EpiEstim across multiple runs.

Usage

combine_rt_epiestim(rt, samples, q = NULL, rank = TRUE)
combine_rt_epiestim(rt, samples, q = NULL, rank = TRUE)

Arguments

rt

A list of Rt calculations from lancelot_rt_trajectories_epiestim() (though any Rt calculation that confirms to this will work)

samples

A list of samples from mcstate::pmcmc

q

A vector of quantiles to return values for

rank

A boolean deciding whether to rank trajectories by increasing incidence or not before combining Rt estimates

Value

A list of Rt output in the same structure as the first element of rt. Rt estimates will be aggregated across regions (or whatever else you are aggregating on) based on the parameters in samples.

Combine trajectories

Description

Combine trajectories across multiple runs

Usage

combine_trajectories(samples, rank = TRUE)
combine_trajectories(samples, rank = TRUE)

Arguments

samples

A list of samples from mcstate::pmcmc

rank

Logical, indicating if trajectories should be ranked before combination.

Value

A set of combined trajectories (not a combined samples object).

Create GPU model

Description

Create a model for the GPU. This requires a working nvcc toolchain and GPU device to work properly. Note that this will cache the compilation within a session, so if you want to change GPU options you will need to restart R. However, if you change the model (e.g., by changing the result of options like rewrite_dims) the model will recompile. Compilation speed is slow enough that the cachine will be obvious.

Usage

compile_gpu(model = "lancelot", ..., real_type = "float", gpu = TRUE)
compile_gpu(model = "lancelot", ..., real_type = "float", gpu = TRUE)

Arguments

model

The name of a sircovid model, either lancelot or basic

...

Additional arguments passed to odin.dust::odin_dust_ and from there into either odin's options or the gpu options.

real_type

The type to use for floating point numbers. The default is "float" which differs from the CPU model but is typically what you want on the GPU.

gpu

The argument passed to odin.dust::odin_dust_ as gpu. The default here is TRUE, but you may want to pass the results of dust::dust_cuda_options in order to control compilation.

Value

A model generator, like lancelot that can run on the GPU available to this computer.

Get the ranking of sample by a given variable, e.g. infections

Description

Get the ranking of sample

Usage

get_sample_rank(sample, by = "infections")
get_sample_rank(sample, by = "infections")

Arguments

sample

An mcstate_pmcmc object

by

The name of the variable used for ranking. The ranking is computed based on the value of this variable at the last time step. Default is set to "infections", which will give the rank based on the cumulative number of infections.

Value

A vector of integers giving the ranks

Inflate strains in model state

Description

Inflate model state, so that additional strain compartments are created but zerod. Use this to run the model for a while, then set it up to work with new strains.

Usage

inflate_state_strains(state1, info1, info2)
inflate_state_strains(state1, info1, info2)

Arguments

state1

The state matrix with one strain

info1

The model info with one strain (or a dust model object)

info2

The model info with two strains (or a dust model object)

Value

An expanded model state with two strains

Author(s)

Richard Fitzjohn

Inflate vacc classes in model state

Description

Inflate model state, so that additional vacc class compartments are created but zerod. Use this to run the model for a while, then set it up to work with new vacc classes (e.g. booster doses).

Usage

inflate_state_vacc_classes(state1, info1, info2)
inflate_state_vacc_classes(state1, info1, info2)

Arguments

state1

The state matrix with X vacc classes

info1

The model info with X vacc classes

info2

The model info with Y > X vacc classes

Value

An expanded model state with Y > X vacc classes

The lancelot sircovid model

Description

##' Our "current" sircovid model. This is a dust model.

Check data for particle filter

Description

Check that data for the particle filter has required columns and constraints among columns. This function does not alter the data at present (though it does return it invisibly).

Usage

lancelot_check_data(data)
lancelot_check_data(data)

Arguments

data

A data.frame of data

Value

Invisibly, a data frame, identical to data

Check severity parameters

Description

Checks severity probabilities are valid

Usage

lancelot_check_severity(pars)
lancelot_check_severity(pars)

Arguments

pars

A list of probabilities from lancelot_parameters.

Value

Returns pars invisibly if all severity probabilities lie in

Compare observed and modelled data for the lancelot model

Description

Compare observed and modelled data from the lancelot model. This conforms to the mcstate interface.

Usage

lancelot_compare(state, observed, pars)
lancelot_compare(state, observed, pars)

Arguments

state

State vector for the end of the current day. This is assumed to be filtered following lancelot_index() so contains 10 rows corresponding to ICU, general beds, admissions, deaths and seroconversion compartments.

observed

Observed data. At the moment please see the tests for a full list as this changes frequently (and this function may be removed in future).

pars

A list of parameters, as created by lancelot_parameters()

Value

A vector of log likelihoods, the same length as the number of particles (the number of columns in the modelled state)

Compute "IFR" excluding immunity

Description

Compute "IFR" excluding immunity.

Usage

lancelot_ifr_excl_immunity(time, pars)
lancelot_ifr_excl_immunity(time, pars)

Arguments

time

A vector of time steps to calculate IFR at

pars

An unnamed list of lancelot_parameters() objects

Index of lancelot model

Description

Index of "interesting" elements for the lancelot model. This function conforms to the mcstate interface.

Usage

lancelot_index(
  info,
  rt = TRUE,
  cum_admit = TRUE,
  diagnoses_admitted = TRUE,
  cum_infections_disag = TRUE,
  cum_n_vaccinated = TRUE,
  D_all = TRUE,
  D_hosp = TRUE,
  infections_inc_per_strain = TRUE,
  severity = FALSE,
  protected = FALSE
)
lancelot_index(
  info,
  rt = TRUE,
  cum_admit = TRUE,
  diagnoses_admitted = TRUE,
  cum_infections_disag = TRUE,
  cum_n_vaccinated = TRUE,
  D_all = TRUE,
  D_hosp = TRUE,
  infections_inc_per_strain = TRUE,
  severity = FALSE,
  protected = FALSE
)

Arguments

info

The result of running the ⁠$info()⁠ method on an initialised basic model

rt

Logical, whether to output trajectories required for calculating Rt (default = TRUE)

cum_admit

Logical, whether to output cumulative admissions by age (default = TRUE)

diagnoses_admitted

Logical, whether to output cumulative combined confirmed admissions and inpatient diagnoses by age and vaccine class (default = TRUE)

cum_infections_disag

Logical, whether to output cumulative infections by age and vaccine class (default = TRUE)

cum_n_vaccinated

Logical, whether to output cumulative number vaccinated by age and vaccine class (default = TRUE)

D_all

Logical, whether to output all deaths by age and vaccine class (default = TRUE)

D_hosp

Logical, whether to output hospital deaths by age (default = TRUE)

infections_inc_per_strain

Logical, whether to output infections incidence per strain (default = TRUE)

severity

Logical, whether to output severity calculations (default = FALSE)

protected

Logical, whether to output protected levels

Value

A list with element run, indicating the locations of (in order) (1) ICU, (2) general, (3) deaths in community, (4) deaths in hospital, (5) total deaths, (6) cumulative confirmed admissions,(7) cumulative confirmed new admissions, (8) probability of a positive sero test, (9) probability of a positive pillar 2 test, and with element state containing the same values followed by 17 S compartments (one per age group, then one for carehome workers and carehome residents respectively) and 17 "cumulative admission" compartments.

Examples

p <- lancelot_parameters(sircovid_date("2020-02-07"), "england")
mod <- lancelot$new(p, 0, 10)
lancelot_index(mod$info())
p <- lancelot_parameters(sircovid_date("2020-02-07"), "england")
mod <- lancelot$new(p, 0, 10)
lancelot_index(mod$info())

Initial conditions for the lancelot model

Description

Create initial conditions for the lancelot model. This matches the interface required for mcstate

Usage

lancelot_initial(info, n_particles, pars)
lancelot_initial(info, n_particles, pars)

Arguments

info

The result of running the ⁠$info()⁠ method on an initialised basic model

n_particles

The number of particles required. Currently only uniform initial seeding is implemented so this has no effect

pars

A parameter list created by basic_parameters(); from this list we will use the population element.

Value

A numeric vector of initial conditions

Examples

p <- lancelot_parameters(sircovid_date("2020-02-07"), "england")
mod <- lancelot$new(p, 0, 10)
lancelot_initial(mod$info(), 10, p)
p <- lancelot_parameters(sircovid_date("2020-02-07"), "england")
mod <- lancelot$new(p, 0, 10)
lancelot_initial(mod$info(), 10, p)

Parameters for the lancelot model

Description

Parameters for the "lancelot" model.

Usage

lancelot_parameters(
  start_date,
  region,
  beta_date = NULL,
  beta_value = NULL,
  beta_type = "piecewise-linear",
  population = NULL,
  has_carehomes = TRUE,
  carehome_beds = NULL,
  severity = NULL,
  progression = NULL,
  observation = NULL,
  sens_and_spec = NULL,
  initial_seed_size = 30,
  initial_seed_pattern = 1,
  eps = 0.1,
  m_CHW = 4e-06,
  m_CHR = 5e-05,
  strain_transmission = 1,
  strain_seed_date = NULL,
  strain_seed_size = NULL,
  strain_seed_pattern = NULL,
  strain_rel_gamma_E = 1,
  strain_rel_gamma_A = 1,
  strain_rel_gamma_P = 1,
  strain_rel_gamma_C_1 = 1,
  strain_rel_gamma_C_2 = 1,
  strain_rel_gamma_PCR_pre = 1,
  strain_rel_gamma_PCR_pos = 1,
  strain_rel_p_sympt = 1,
  strain_rel_p_hosp_if_sympt = 1,
  strain_rel_p_icu = 1,
  strain_rel_p_death = 1,
  strain_rel_p_G_D = 1,
  rel_susceptibility = 1,
  rel_p_sympt = 1,
  rel_p_hosp_if_sympt = 1,
  rel_p_death = 1,
  rel_infectivity = 1,
  vaccine_progression_rate = NULL,
  vaccine_schedule = NULL,
  vaccine_index_dose2 = NULL,
  vaccine_index_booster = NULL,
  vaccine_catchup_fraction = 1,
  n_doses = 2L,
  vacc_skip_progression_rate = NULL,
  vacc_skip_to = NULL,
  vacc_skip_weight = NULL,
  waning_rate = 0,
  exp_noise = 1e+06,
  cross_immunity = 1
)
lancelot_parameters(
  start_date,
  region,
  beta_date = NULL,
  beta_value = NULL,
  beta_type = "piecewise-linear",
  population = NULL,
  has_carehomes = TRUE,
  carehome_beds = NULL,
  severity = NULL,
  progression = NULL,
  observation = NULL,
  sens_and_spec = NULL,
  initial_seed_size = 30,
  initial_seed_pattern = 1,
  eps = 0.1,
  m_CHW = 4e-06,
  m_CHR = 5e-05,
  strain_transmission = 1,
  strain_seed_date = NULL,
  strain_seed_size = NULL,
  strain_seed_pattern = NULL,
  strain_rel_gamma_E = 1,
  strain_rel_gamma_A = 1,
  strain_rel_gamma_P = 1,
  strain_rel_gamma_C_1 = 1,
  strain_rel_gamma_C_2 = 1,
  strain_rel_gamma_PCR_pre = 1,
  strain_rel_gamma_PCR_pos = 1,
  strain_rel_p_sympt = 1,
  strain_rel_p_hosp_if_sympt = 1,
  strain_rel_p_icu = 1,
  strain_rel_p_death = 1,
  strain_rel_p_G_D = 1,
  rel_susceptibility = 1,
  rel_p_sympt = 1,
  rel_p_hosp_if_sympt = 1,
  rel_p_death = 1,
  rel_infectivity = 1,
  vaccine_progression_rate = NULL,
  vaccine_schedule = NULL,
  vaccine_index_dose2 = NULL,
  vaccine_index_booster = NULL,
  vaccine_catchup_fraction = 1,
  n_doses = 2L,
  vacc_skip_progression_rate = NULL,
  vacc_skip_to = NULL,
  vacc_skip_weight = NULL,
  waning_rate = 0,
  exp_noise = 1e+06,
  cross_immunity = 1
)

Arguments

start_date

The start date, as a sircovid_date() (i.e., the number of days into 2020)

region

beta_date

beta_value

beta_type

The type of form used for beta (the contact rate parameter), which currently can be "piecewise-linear" or "piecewise-constant"

population

Population data. A vector of length 17 for the population size for age groups 0-4, 5-9, ..., 75-79, 80+. If NULL, the population data will be sourced within the package for the specified region (only if available).

has_carehomes

Logical, whethere or not the model has carehomes.

carehome_beds

The number of care home beds in the region. If NULL, this will be sourced within the package for the specified region (only if available).

severity

progression

Progression data

observation

Either NULL or a list of observation parameters. If NULL, then a list of observation parameters will be generated using lancelot_parameters_observation(exp_noise)

sens_and_spec

Either NULL or a list of diagnostic test sensitivity and specificity parameters. If NULL, then a list of sensitivity and specificity parameters will be generated using lancelot_parameters_sens_and_spec()

initial_seed_size

Initial size of seeding from the S to E compartment; all seeding is in the 15-19 year old group from start_date according to initial_seed_pattern. The default is 30.

initial_seed_pattern

eps

Change in contact rate for carehome residents

m_CHW

Contact rate between carehome workers and either residents or workers

m_CHR

Contact rate between carehome residents

strain_transmission

Vector of length two for relative transmissibility of each strain modelled. Length will define the number of strains used in the model, either 1 or 2.

strain_seed_date

Either NULL (no seeding) or a sircovid_date corresponding to the date the seeding of strain 2 begins.

strain_seed_size

Either NULL (no seeding) or the size of strain 2 seeding from the S to E compartment; all seeding is in the 15-19 year old group from strain_seed_date according to strain_seed_pattern.

strain_seed_pattern

Either NULL (no seeding) or a vector of seeding weights for the initial seeding. The length represents the number of steps to seed over from the start_date, and the strain_seed_size is split over these steps according to those weights. If start_seed_date is not a multiple of the step size (and thus falls between two steps) then we weight over an additional step and adjust the weights according to how far the strain_seed_date is from the previous full step.

strain_rel_gamma_E

Vector of relative rates of progression out of E (gamma_E) for each strain modelled. If 1 all strains have same rates. Otherwise vector of same length as strain_transmission, with entries that determines the relative scaling of the defaults for each strain.

strain_rel_gamma_A

Vector of relative rates of progression out of I_A (gamma_A) for each strain modelled. If 1 all strains have same rates. Otherwise vector of same length as strain_transmission, with entries that determines the relative scaling of the defaults for each strain.

strain_rel_gamma_P

Vector of relative rates of progression out of I_P (gamma_P) for each strain modelled. If 1 all strains have same rates. Otherwise vector of same length as strain_transmission, with entries that determines the relative scaling of the defaults for each strain.

strain_rel_gamma_C_1

Vector of relative rates of progression out of I_C_1 (gamma_C_1) for each strain modelled. If 1 all strains have same rates. Otherwise vector of same length as strain_transmission, with entries that determines the relative scaling of the defaults for each strain.

strain_rel_gamma_C_2

Vector of relative rates of progression out of I_C_2 (gamma_C_2) for each strain modelled. If 1 all strains have same rates. Otherwise vector of same length as strain_transmission, with entries that determines the relative scaling of the defaults for each strain.

strain_rel_gamma_PCR_pre

Vector of relative rates of PCR pre-positive duration for each strain modelled. if '1' all strains have same rates. Otherwise vector of same length as strain_transmission, with entries that determines the relative scaling of the defaults for each strain.

strain_rel_gamma_PCR_pos

Vector of relative rates of PCR positivity duration for each strain modelled. if '1' all strains have same rates. Otherwise vector of same length as strain_transmission, with entries that determines the relative scaling of the defaults for each strain.

strain_rel_p_sympt

Vector of relative probabilities of symptoms for each strain modelled. If 1 all strains have same probabilities of symptoms. Otherwise vector of same length as strain_transmission, where the first value should be 1 (for the first strain) and subsequent values between 0 and 1. In this case parameters will be "mirrored" for pseudostrains i.e. the relative probability of symptoms will be assume the same irrespective of previous infection with another strain. Alternatively, a vector of twice the length of strain_transmission can be provided to allow specifying directly relative probability of symptoms for each pseudostrain (with strain 3 - 1.2 and strain 4 = 2.1). To ensure valid probabilities, p_sympt is upper-truncated at 1 after scaling.

strain_rel_p_hosp_if_sympt

Vector of relative probabilities of hospitalisation given symptoms for each strain modelled. If 1 all strains have same probabilities of hospitalisation. Otherwise vector of same length as strain_transmission, where the first value should be 1 (for the first strain) and subsequent values between 0 and 1. In this case parameters will be "mirrored" for pseudostrains i.e. the relative probability of hospitalisation will be assume the same irrespective of previous infection with another strain. Alternatively, a vector of twice the length of strain_transmission can be provided to allow specifying directly relative probability of hospitalisation for each pseudostrain (with strain 3 - 1.2 and strain 4 = 2.1). To ensure valid probabilities, p_hosp_if_sympt is upper-truncated at 1 after scaling.

strain_rel_p_icu

Vector of relative probabilities of icu given hospitalised for each strain modelled. If 1 all strains have same probabilities of icu admission. Otherwise vector of same length as strain_transmission, where the first value should be 1 (for the first strain) and subsequent values between 0 and 1. In this case parameters will be "mirrored" for pseudostrains i.e. the relative probability of icu admission will be assume the same irrespective of previous infection with another strain. Alternatively, a vector of twice the length of strain_transmission can be provided to allow specifying directly relative probability of icu admission for each pseudostrain (with strain 3 - 1.2 and strain 4 = 2.1). To ensure valid probabilities, p_icu is upper-truncated at 1 after scaling.

strain_rel_p_death

Vector of relative probabilities of death for each strain modelled. If 1 all strains have same probabilities of death. Otherwise vector of same length as strain_transmission, where the first value should be 1 (for the first strain) and subsequent values between 0 and 1. In this case parameters will be "mirrored" for pseudostrains i.e. the relative probability of death will be assume the same irrespective of previous infection with another strain. Alternatively, a vector of twice the length of strain_transmission can be provided to allow specifying directly relative probability of death for each pseudostrain (with strain 3 - 1.2 and strain 4 = 2.1). To ensure valid probabilities, p_death is upper-truncated at 1 after scaling.

strain_rel_p_G_D

Vector of relative probabilities of death in the coummunity for each strain modelled. If 1 all strains have same probabilities of death. Otherwise vector of same length as strain_transmission, where the first value should be 1 (for the first strain) and subsequent values between 0 and 1. In this case parameters will be "mirrored" for pseudostrains i.e. the relative probability of death will be assume the same irrespective of previous infection with another strain. Alternatively, a vector of twice the length of strain_transmission can be provided to allow specifying directly relative probability of death for each pseudostrain (with strain 3 - 1.2 and strain 4 = 2.1). To ensure valid probabilities, p_G_D is upper-truncated at 1 after scaling.

rel_susceptibility

A vector or array of values representing the relative susceptibility of individuals in different vaccination groups. If a vector, the first value should be 1 (for the non-vaccinated group) and subsequent values be between 0 and 1. In that case relative susceptibility will be the same across all age groups within one vaccination category, and will be the same for all pathogen strains. Specifying an array instead of a vector allows different relative susceptibilities by age (first dimension of the array), pathogen strain (second dimension) and vaccination group (third dimension); in that case, the first layer (3rd dimension) of rel_susceptibility should be 1 (for the non-vaccinated group) for the first column (first infection with first strain) and other values between 0 and 1

rel_p_sympt

A vector or matrix of values of same dimension as rel_susceptibility representing the relative probability of symptomatic infection in different vaccination groups. If a vector, the first value should be 1 (for the non-vaccinated group) and subsequent values be between 0 and 1. In that case the relative reduction in probability of symptomatic infection will be the same across all age groups and all strains within one vaccination category. Specifying an array instead of a vector allows different relative reductions in probability of symptomatic infection by age (first dimension of the array), pathogen strain (second dimension) and vaccination group (third dimension); in that case, the first layer of rel_p_sympt should be 1 (for the non-vaccinated group) for the first column (first infection with first strain) and other values between 0 and 1

rel_p_hosp_if_sympt

A vector or array of values of same dimension as rel_susceptibility representing the relative probability of hospitalisation for symptomatic cases in different vaccination groups. If a vector, the first value should be 1 (for the non-vaccinated group) and subsequent values be between 0 and 1. In that case the relative reduction in probability of hospitalisation for symptomatic cases will be the same across all age groups and all strains within one vaccination category. Specifying an array instead of a vector allows different relative reductions in probability of hospitalisation for symptomatic cases by age (first dimension of the array), pathogen strain (second dimension) and vaccination group (third dimension); in that case, the first layer of rel_p_hosp_if_sympt should be 1 (for the non-vaccinated group) for the first column (first infection with first strain) and other values between 0 and 1

rel_p_death

A vector or array of values of same dimension as rel_susceptibility representing the relative probability of death for severe cases (either in hospital or the community) in different vaccination groups. If a vector, the first value should be 1 (for the non-vaccinated group) and subsequent values be between 0 and 1. In that case the relative reduction in probability of death for severe cases will be the same across all age groups and all strains within one vaccination category. Specifying an array instead of a vector allows different relative reductions in probability of death for severe cases by age (first dimension of the array), pathogen strain (second dimension) and vaccination group (third dimension); in that case, the first layer of rel_p_death should be 1 (for the non-vaccinated group) for the first column (first infection with first strain) and other values between 0 and 1

rel_infectivity

A vector or array of values representing the relative infectivity of individuals in different vaccination groups, if they are infected. If a vector, the first value should be 1 (for the non-vaccinated group) and subsequent values be between 0 and 1. In that case relative infectivity will be the same across all age groups and all strains within one vaccination category. Specifying an array instead of a vector allows different relative infectivities by age (first dimension of the array), pathogen strain (second dimension) and vaccination group (third dimension); in that case, the first layer of rel_infectivity should be 1 (for the non-vaccinated group) for the first column (first infection with first strain) and other values between 0 and 1

vaccine_progression_rate

A vector or matrix of values of same dimension as rel_susceptibility representing the rate of movement between different vaccination classes. If a vector, it should have as many values as vaccination classes, and the same rates of progression will be used for all age groups (the first rate is the vaccination rate and the last of the rates is the rate of returning to the initial vacination class); if a matrix, the element on row i and column j is the rate of progression from the jth vaccination class to the (j+1)th for age group i.

vaccine_schedule

A vaccine_schedule object indicating the people to be vaccinated by group over time

vaccine_index_dose2

The index to use for the second dose

vaccine_index_booster

The indices to use for the booster doses

vaccine_catchup_fraction

A value between 0 and 1 indicating the proportion of doses not distributed according to schedule (e.g. because too many people were in the I or H compartments and could not be vaccinated at the scheduled time) that we postpone to a later date. A value of 0 means we do not catch up at all on any missed doses; a value of 1 means we try to catch up for all missed doses. This is set to 1 by default

n_doses

Number of doses given out, including boosters. Default is 2.

vacc_skip_progression_rate

A vector of length equal to the number of vaccine strata. Represents the base progression rate for "vaccine skip moves" from each strata. The rate is assumed to be the same across all age groups. Must be 0 for all j for which vacc_skip[j] = 0. Default is a vector of 0s

vacc_skip_to

A vector of integers of length equal to the number of vaccine strata. The value of vacc_skip_to[j] represents the vaccine stratum an individual in vaccine stratum j can move to in a "vaccine skip move". Values can be 0 (which mean there is no vaccine skip move from that stratum), or greater than or equal to j + 2 (we already account for moves between successive vaccine strata). More than one vaccine skip move to the same vaccine stratum is not allowed, so all non-zero values in vacc_skip_to must be unique. Overlapping vaccine skip moves are also not allowed - if a there is a vaccine skip move from i to j there can be no vaccine skip moves from stratum k where ⁠i < k < j⁠. For example, having vaccine skip moves from 2 to 4 and 3 to 5 would not be allowed, but moves from 2 to 4 and 4 to 6 would be allowed. Default is a vector of 0s

vacc_skip_weight

A vector of weights of length equal to the number of vaccine strata. If movement into vacc_skip_to[j] from vacc_skip_to[j] - 1 is controlled by doses, then the "vaccine skip move" is too, and the weight represents how much those in strata j are weighted in dose distribution relative to those in vacc_skip_to - 1. Must be between 0 and 1. Must be 0 for any j for which vacc_skip_to = 0. Default is a vector of 0s

waning_rate

A single value or a vector of values representing the rates of waning of immunity after infection; if a single value the same rate is used for all age groups; if a vector of values if used it should have one value per age group.

exp_noise

cross_immunity

A value or vector of same length as strain_transmission that controls the amount of immunity conferred by previous infection with one strain. If a scalar is given then same level of cross immunity is assumed between both strains. Otherwise a vector of length two should be provided where the first value is the relative protection against infection with strain 2 following infection with strain 1 (i.e. while in the R1 compartment), and vice versa for the second value. Values between 0 and 1 are allowed with values of 1 (default) indicating complete cross-immunity, and values of 0 mean no cross-immunity. Modelling 'superinfections' (being exposed to one strain after recovering from another) can be turned off by setting cross_immunity = 1.

Value

Examples


region <- "london"
lancelot_parameters(sircovid_date("2020-02-01"), region)

# example set up of vaccination parameters independent of age
# 3 groups: 1) unvaccinated, 2) vaccinated with partial immunity
# 3) fully vaccinated (but with an imperfect vaccine). People return
# to group 1 after a period of time in group 3 to mirror waning immunity

# Assumption: immediately after vaccination susceptibility is reduced by
# 20%, and then by 50% when you reach full effect of the vaccination,
# then susceptibility returns to 100% upon waning of vaccine=induced
# immunity
# effect of vaccination similar across all age groups
rel_susceptibility <- c(1, 0.8, 0.5)

# The vaccine also reduces the risk of symptoms
rel_p_sympt <- c(1, 0.6, 0.3)
# and the risk of hospitalisation for those with symptoms
rel_p_hosp_if_sympt <- c(1, 0.95, 0.95)
# and the risk of death for those with severe disease
rel_p_death <- c(1, 0.9, 0.9)

# The vaccine also reduces infectivity of infected individuals by half
rel_infectivity <- c(1, 0.5, 0.5)

# On average 10000 first doses of vaccine are given every day
# Second doses are given on average 12 weeks after the first dose
# vaccine-induced immunity wanes after a period which is exponentially
# distributed and lasts on average 26 weeks (half a year);
# they are similar across all age groups
vaccine_progression_rate <- c(0, 0, 1/(26*7))

daily_doses <- rep(10000, 365)
mean_days_between_doses <- 12 * 7
n <- vaccine_priority_population(region, uptake = 1)
schedule <- vaccine_schedule_future(
  0, daily_doses, mean_days_between_doses, n)

# generate model parameters
p <- lancelot_parameters(
       sircovid_date("2020-02-01"), region,
       rel_susceptibility = rel_susceptibility,
       rel_p_sympt = rel_p_sympt,
       rel_p_hosp_if_sympt = rel_p_hosp_if_sympt,
       rel_p_death = rel_p_death,
       rel_infectivity = rel_infectivity,
       vaccine_progression_rate = vaccine_progression_rate,
       vaccine_schedule = schedule,
       vaccine_index_dose2 = 2)

# vaccination parameters are automatically copied across all age groups#
# (and across strains but here we only have 1 strain which is the 2nd
# dimension here)
p$rel_susceptibility
p$rel_p_sympt
p$rel_p_hosp_if_sympt
p$rel_p_death
p$rel_infectivity
# Note that this is only the "base" rate as we fill in the first
# column dynamically based on vaccine_daily_doses
p$vaccine_progression_rate

### same example as above BUT assume a different effect of vaccine in the
### first age group
n_groups <- 19
n_strains <- 1

# Assumption: vaccine is twice more effective at reducing susceptibility
# in the first age group
rel_susceptibility_agegp1 <- c(1, 0.4, 0.25)
rel_susceptibility_other_agegp <- c(1, 0.8, 0.5)
rel_susceptibility <- array(NA, dim = c(n_groups, n_strains, 3))
rel_susceptibility[1, , ] <- rel_susceptibility_agegp1
for (i in seq(2, n_groups)) {
  rel_susceptibility[i, , ] <- rel_susceptibility_other_agegp
}
rel_susceptibility

# But vaccine has the same impact on probability of symptoms and
# hospitalisation for the symptomatic across all age groups
rel_p_sympt <- array(rep(rel_p_sympt, each = n_groups),
  dim = c(n_groups, n_strains, 3))
rel_p_hosp_if_sympt <-
  array(rep(rel_p_hosp_if_sympt, each = n_groups),
  dim = c(n_groups, n_strains, 3))
rel_p_death <-
  array(rep(rel_p_hosp_if_sympt, each = n_groups),
  dim = c(n_groups, n_strains, 3))

# And vaccine has the same impact on onwards infectivity across age groups
rel_infectivity <- array(rep(rel_infectivity, each = n_groups),
  dim = c(n_groups, n_strains, 3))

# the period of build-up of immunity is the same for all age groups,
# lasting on average 2 weeks,
# but the first age group loses immunity more quickly
# (on average after 3 months) than the other age groups
# (on average after 6 months)
vaccine_progression_rate <- cbind(0, 0,
                                  c(1 / (13 * 7),
                                  rep( 1 / (26 * 7), n_groups - 1)))

# generate model parameters
p <- lancelot_parameters(
       sircovid_date("2020-02-01"), region,
       rel_susceptibility = rel_susceptibility,
       rel_p_sympt = rel_p_sympt,
       rel_p_hosp_if_sympt = rel_p_hosp_if_sympt,
       rel_p_death = rel_p_death,
       rel_infectivity = rel_infectivity,
       vaccine_progression_rate = vaccine_progression_rate,
       vaccine_schedule = schedule,
       vaccine_index_dose2 = 2)

# TODO: add an example of manually set up vaccine schedule

region <- "london"
lancelot_parameters(sircovid_date("2020-02-01"), region)

# example set up of vaccination parameters independent of age
# 3 groups: 1) unvaccinated, 2) vaccinated with partial immunity
# 3) fully vaccinated (but with an imperfect vaccine). People return
# to group 1 after a period of time in group 3 to mirror waning immunity

# Assumption: immediately after vaccination susceptibility is reduced by
# 20%, and then by 50% when you reach full effect of the vaccination,
# then susceptibility returns to 100% upon waning of vaccine=induced
# immunity
# effect of vaccination similar across all age groups
rel_susceptibility <- c(1, 0.8, 0.5)

# The vaccine also reduces the risk of symptoms
rel_p_sympt <- c(1, 0.6, 0.3)
# and the risk of hospitalisation for those with symptoms
rel_p_hosp_if_sympt <- c(1, 0.95, 0.95)
# and the risk of death for those with severe disease
rel_p_death <- c(1, 0.9, 0.9)

# The vaccine also reduces infectivity of infected individuals by half
rel_infectivity <- c(1, 0.5, 0.5)

# On average 10000 first doses of vaccine are given every day
# Second doses are given on average 12 weeks after the first dose
# vaccine-induced immunity wanes after a period which is exponentially
# distributed and lasts on average 26 weeks (half a year);
# they are similar across all age groups
vaccine_progression_rate <- c(0, 0, 1/(26*7))

daily_doses <- rep(10000, 365)
mean_days_between_doses <- 12 * 7
n <- vaccine_priority_population(region, uptake = 1)
schedule <- vaccine_schedule_future(
  0, daily_doses, mean_days_between_doses, n)

# generate model parameters
p <- lancelot_parameters(
       sircovid_date("2020-02-01"), region,
       rel_susceptibility = rel_susceptibility,
       rel_p_sympt = rel_p_sympt,
       rel_p_hosp_if_sympt = rel_p_hosp_if_sympt,
       rel_p_death = rel_p_death,
       rel_infectivity = rel_infectivity,
       vaccine_progression_rate = vaccine_progression_rate,
       vaccine_schedule = schedule,
       vaccine_index_dose2 = 2)

# vaccination parameters are automatically copied across all age groups#
# (and across strains but here we only have 1 strain which is the 2nd
# dimension here)
p$rel_susceptibility
p$rel_p_sympt
p$rel_p_hosp_if_sympt
p$rel_p_death
p$rel_infectivity
# Note that this is only the "base" rate as we fill in the first
# column dynamically based on vaccine_daily_doses
p$vaccine_progression_rate

### same example as above BUT assume a different effect of vaccine in the
### first age group
n_groups <- 19
n_strains <- 1

# Assumption: vaccine is twice more effective at reducing susceptibility
# in the first age group
rel_susceptibility_agegp1 <- c(1, 0.4, 0.25)
rel_susceptibility_other_agegp <- c(1, 0.8, 0.5)
rel_susceptibility <- array(NA, dim = c(n_groups, n_strains, 3))
rel_susceptibility[1, , ] <- rel_susceptibility_agegp1
for (i in seq(2, n_groups)) {
  rel_susceptibility[i, , ] <- rel_susceptibility_other_agegp
}
rel_susceptibility

# But vaccine has the same impact on probability of symptoms and
# hospitalisation for the symptomatic across all age groups
rel_p_sympt <- array(rep(rel_p_sympt, each = n_groups),
  dim = c(n_groups, n_strains, 3))
rel_p_hosp_if_sympt <-
  array(rep(rel_p_hosp_if_sympt, each = n_groups),
  dim = c(n_groups, n_strains, 3))
rel_p_death <-
  array(rep(rel_p_hosp_if_sympt, each = n_groups),
  dim = c(n_groups, n_strains, 3))

# And vaccine has the same impact on onwards infectivity across age groups
rel_infectivity <- array(rep(rel_infectivity, each = n_groups),
  dim = c(n_groups, n_strains, 3))

# the period of build-up of immunity is the same for all age groups,
# lasting on average 2 weeks,
# but the first age group loses immunity more quickly
# (on average after 3 months) than the other age groups
# (on average after 6 months)
vaccine_progression_rate <- cbind(0, 0,
                                  c(1 / (13 * 7),
                                  rep( 1 / (26 * 7), n_groups - 1)))

# generate model parameters
p <- lancelot_parameters(
       sircovid_date("2020-02-01"), region,
       rel_susceptibility = rel_susceptibility,
       rel_p_sympt = rel_p_sympt,
       rel_p_hosp_if_sympt = rel_p_hosp_if_sympt,
       rel_p_death = rel_p_death,
       rel_infectivity = rel_infectivity,
       vaccine_progression_rate = vaccine_progression_rate,
       vaccine_schedule = schedule,
       vaccine_index_dose2 = 2)

# TODO: add an example of manually set up vaccine schedule

Lancelot observation parameters

Description

Lancelot observation parameters

Usage

lancelot_parameters_observation(exp_noise = 1e+06)
lancelot_parameters_observation(exp_noise = 1e+06)

Arguments

exp_noise

Rate parameter for the exponentially-distributed noise used in the likelihood calculation. Typically a large value so noise is small.

Value

A list of parameter values

Lancelot progression parameters

Description

Lancelot progression parameters. The s_ parameters are the scaling parameters for the Erlang distibution (a.k.a 'k'), while the gamma_ parameters are the gamma parameters of that distribution. These need to be aligned with Bob's severity outputs, and we will come up with a better way of coordinating the two.

Usage

lancelot_parameters_progression(
  dt,
  gamma_E = NULL,
  gamma_A = NULL,
  gamma_P = NULL,
  gamma_C_1 = NULL,
  gamma_C_2 = NULL,
  gamma_ICU_pre = NULL,
  gamma_ICU_D = NULL,
  gamma_ICU_W_D = NULL,
  gamma_ICU_W_R = NULL,
  gamma_H_D = NULL,
  gamma_H_R = NULL,
  gamma_W_D = NULL,
  gamma_W_R = NULL,
  gamma_G_D = NULL,
  gamma_U = NULL,
  gamma_PCR_pre = NULL,
  gamma_PCR_pos = NULL
)
lancelot_parameters_progression(
  dt,
  gamma_E = NULL,
  gamma_A = NULL,
  gamma_P = NULL,
  gamma_C_1 = NULL,
  gamma_C_2 = NULL,
  gamma_ICU_pre = NULL,
  gamma_ICU_D = NULL,
  gamma_ICU_W_D = NULL,
  gamma_ICU_W_R = NULL,
  gamma_H_D = NULL,
  gamma_H_R = NULL,
  gamma_W_D = NULL,
  gamma_W_R = NULL,
  gamma_G_D = NULL,
  gamma_U = NULL,
  gamma_PCR_pre = NULL,
  gamma_PCR_pos = NULL
)

Arguments

dt

The step size

gamma_E

Time-varying parameters for the Erlang rate parameter of the duration in the E (exposed) compartment. See Details.

gamma_A

Time-varying parameters for the Erlang rate parameter of the duration in the I_A (asymptomatic) compartment. See Details.

gamma_P

Time-varying parameters for the Erlang rate parameter of the duration in the I_P (presymptomatic) compartment. See Details.

gamma_C_1

Time-varying parameters for the Erlang rate parameter of the duration in the I_C_1 (first stage symptomatic) compartment. See Details.

gamma_C_2

Time-varying parameters for the Erlang rate parameter of the duration in the I_C_2 (second stage symptomatic) compartment. See Details.

gamma_ICU_pre

Time-varying parameters for the Erlang rate parameter of the duration in the ICU_pre (general beds stay before ICU) compartment. See Details.

gamma_ICU_D

Time-varying parameters for the Erlang rate parameter of the duration in the ICU_D (ICU stay for individuals who die in ICU) compartment. See Details.

gamma_ICU_W_D

Time-varying parameters for the Erlang rate parameter of the duration in the ICU_W_D (ICU stay for individuals who go on to die in stepdown) compartment. See Details.

gamma_ICU_W_R

Time-varying parameters for the Erlang rate parameter of the duration in the ICU_W_R (ICU stay for individuals who go on to recover in stepdown) compartment. See Details.

gamma_H_D

Time-varying parameters for the Erlang rate parameter of the duration in the H_D (general beds stay for individuals who die in general beds) compartment. See Details.

gamma_H_R

Time-varying parameters for the Erlang rate parameter of the duration in the H_R (general beds stay for individuals who recover in general beds) compartment. See Details.

gamma_W_D

Time-varying parameters for the Erlang rate parameter of the duration in the W_D (stepdown for individuals who die in stepdown) compartment. See Details.

gamma_W_R

Time-varying parameters for the Erlang rate parameter of the duration in the W_R (stepdown for individuals who recover in stepdown) compartment. See Details.

gamma_G_D

Time-varying parameters for the Erlang rate parameter of the duration in the G_D (delay to death in the community/care homes) compartment. See Details.

gamma_U

Time-varying parameters for the Erlang rate parameter of the duration of the delay from hospital admission to diagnosis for those not confirmed on admission. Note this duration is a single-stage Erlang. See Details.

gamma_PCR_pre

Time-varying parameters for the Erlang rate parameter the duration of PCR pre-positivity affecting REACT and ONS infection prevalence. See Details.

gamma_PCR_pos

Time-varying parameters for the Erlang rate parameter the duration of PCR positivity affecting REACT and ONS infection prevalence. See Details.

Value

A list of parameter values

Time-varying parameters

Every time varying parameter has the same format, which can be NULL (in which case a default single value is used) or a list with date and value for changes in the parameter. If value is scalar then date can be NULL or missing. If value is a vector then date must be a vector of sircovid dates of the same length as value.

Lancelot sensitivity and specificity parameters

Description

Lancelot observation parameters

Usage

lancelot_parameters_sens_and_spec(
  sero_specificity_1 = 0.9,
  sero_sensitivity_1 = 0.99,
  sero_specificity_2 = 0.9,
  sero_sensitivity_2 = 0.99,
  pillar2_specificity = 0.99,
  pillar2_sensitivity = 0.99,
  ons_specificity = 0.99,
  ons_sensitivity = 0.99,
  react_specificity = 0.99,
  react_sensitivity = 0.99
)
lancelot_parameters_sens_and_spec(
  sero_specificity_1 = 0.9,
  sero_sensitivity_1 = 0.99,
  sero_specificity_2 = 0.9,
  sero_sensitivity_2 = 0.99,
  pillar2_specificity = 0.99,
  pillar2_sensitivity = 0.99,
  ons_specificity = 0.99,
  ons_sensitivity = 0.99,
  react_specificity = 0.99,
  react_sensitivity = 0.99
)

Arguments

sero_specificity_1

Specificity of the first serology test assay

sero_sensitivity_1

Sensitivity of the first serology test assay

sero_specificity_2

Specificity of the second serology test assay

sero_sensitivity_2

Sensitivity of the second serology test assay

pillar2_specificity

Specificity of the Pillar 2 test

pillar2_sensitivity

Sensitivity of the Pillar 2 test

ons_specificity

Specificity of the ONS test

ons_sensitivity

Sensitivity of the ONS test

react_specificity

Specificity of the REACT test

react_sensitivity

Sensitivity of the REACT test

Value

A list of parameter values

Lancelot severity parameters

Description

Lancelot severity parameters

Usage

lancelot_parameters_severity(
  dt,
  severity = NULL,
  has_carehomes = TRUE,
  p_C = NULL,
  p_H = NULL,
  p_H_CHR = NULL,
  p_ICU = NULL,
  p_H_D = NULL,
  p_ICU_D = NULL,
  p_W_D = NULL,
  p_G_D = NULL,
  p_G_D_CHR = NULL,
  p_R = NULL,
  p_star = NULL
)
lancelot_parameters_severity(
  dt,
  severity = NULL,
  has_carehomes = TRUE,
  p_C = NULL,
  p_H = NULL,
  p_H_CHR = NULL,
  p_ICU = NULL,
  p_H_D = NULL,
  p_ICU_D = NULL,
  p_W_D = NULL,
  p_G_D = NULL,
  p_G_D_CHR = NULL,
  p_R = NULL,
  p_star = NULL
)

Arguments

dt

The step size

severity

Severity data, used to determine default severity parameter age-scalings and to provide default severity parameter values. Can be NULL (use the default bundled data version in the package), or a data.frame object (for raw severity data).

has_carehomes

Logical, whether or not the model has carehomes

p_C

Time-varying parameters for p_C (the probability of an infected individual becoming symptomatic). See Details.

p_H

Time-varying parameters for p_H (the probability of a symptomatic individual requiring hospitalisation). See Details.

p_H_CHR

Time-varying parameters for p_H (the probability of a symptomatic individual requiring hospitalisation) for care home residents. If NULL then the value for the oldest age group is used. See Details.

p_ICU

Time-varying parameters for p_ICU (the probability of a hospitalised individual going to ICU). See Details.

p_H_D

Time-varying parameters for p_H_D (the probability of death in general beds).See Details.

p_ICU_D

Time-varying parameters for p_ICU_D (the probability of death in ICU). See Details.

p_W_D

Time-varying parameters for p_W_D (the probability of death in stepdown). See Details.

p_G_D

Time-varying parameters for p_G_D (the probability of individuals requiring hospitalisation dying in the community or a care home). See Details.

p_G_D_CHR

Time-varying parameters for p_G_D (the probability of individuals requiring hospitalisation dying in a care home) for care home residents. If NULL then the value for the oldest age group is used. See Details.

p_R

Time-varying parameters for p_R (the probability of an non- fatally infected individual having immunity post-infection). See Details.

p_star

Time-varying parameters for p_star (the probability of patients being confirmed as covid on admission to hospital). See Details.

Value

A list of severity parameters

Time-varying parameters

Every time varying parameter has the same format, which can be NULL (in which case the value from severity is used) or a list with date and value for changes in the parameter. If value is scalar then date can be NULL or missing. If value is a vector then date must be a vector of sircovid dates of the same length as value.

Compute "Rt"

Description

Compute "Rt" for a single simulated trajectory and parameter set.

Usage

lancelot_Rt(
  time,
  S,
  p,
  prob_strain = NULL,
  type = NULL,
  interpolate_every = NULL,
  interpolate_critical_dates = NULL,
  interpolate_min = NULL,
  eigen_method = "power_iteration",
  R = NULL,
  weight_Rt = FALSE,
  keep_strains_Rt = FALSE
)
lancelot_Rt(
  time,
  S,
  p,
  prob_strain = NULL,
  type = NULL,
  interpolate_every = NULL,
  interpolate_critical_dates = NULL,
  interpolate_min = NULL,
  eigen_method = "power_iteration",
  R = NULL,
  weight_Rt = FALSE,
  keep_strains_Rt = FALSE
)

Arguments

time

A vector of time steps that the model was run over

S

A (n groups x n vaccine classes) x steps matrix of "S" compartment counts

p

A lancelot_parameters() object

prob_strain

A (n groups x n strains) x n time steps matrix of "prob_strain" outputs from the model. For a 2 strain model for example, prob_strain[1, j] and prob_strain[n_groups + 1, j] should give, for the j^th time step, the probabilities that new infections in group 1 are of strains 1 and 2 respectively. The default is NULL, but it must be specified if there is more than one strain

type

A character vector of possible Rt types to compute. Can be any or all of eff_Rt_all, eff_Rt_general, Rt_all and Rt_general

interpolate_every

Spacing (in days) to use between interpolated points

interpolate_critical_dates

Optional vector of critical sircovid dates to use when interpolating. Interpolation will be done in blocks between the first time step of these dates, with each block starting on the first time step of the date given. So if you give a interpolate_critical_dates of c(20, 50) then blocks start on first time step of days 20 and 50, i.e.: ⁠[1, 20)⁠, ⁠[20, 50)⁠, ⁠[50, end]⁠.

interpolate_min

The minimum number of steps to include within a block. If there are fewer points than this then all points are used (i.e., no interpolation is done) or interpolate_every is reduced until at least this many points were used. This can be used to specify a lower bound on the error of small regions. If Rt is small it won't matter that much. You do need to specify something though or interpolation will not happen, and do not use less than 3 as we use spline interpolation and that will not work with fewer than 3 points.

eigen_method

The eigenvalue method to use (passed to eigen1::eigen1 as method)

R

A (n groups x n strains x n vaccine classes) x time steps matrix of "R" compartment counts, required for multi-strain models.

weight_Rt

If TRUE then computes the weighted average of the Rt for all strains, otherwise all calculations are returned with an additional dimension to index each strain.

keep_strains_Rt

Additional argument for when weight_Rt is TRUE (has no impact otherwise). If TRUE, then the Rt for each strain is returned along with the weighted average, otherwise just the weighted average is returned. When TRUE, the dimension indexing these lists strains first, and then the weighted average.

Value

A list with elements time, beta, and any of the type values specified above.

Compute Rt for a set of trajectories

Description

Compute "Rt" for a set of simulated trajectories (e.g., the result of the ⁠$iterate()⁠ method of lancelot, mcstate::pmcmc() or mcstate::pmcmc_predict(). The trajectories may or may not share parameters.

Usage

lancelot_Rt_trajectories(
  time,
  S,
  pars,
  prob_strain = NULL,
  initial_time_from_parameters = TRUE,
  shared_parameters = NULL,
  type = NULL,
  interpolate_every = NULL,
  interpolate_critical_dates = NULL,
  interpolate_min = NULL,
  eigen_method = "power_iteration",
  R = NULL,
  weight_Rt = FALSE,
  keep_strains_Rt = FALSE
)
lancelot_Rt_trajectories(
  time,
  S,
  pars,
  prob_strain = NULL,
  initial_time_from_parameters = TRUE,
  shared_parameters = NULL,
  type = NULL,
  interpolate_every = NULL,
  interpolate_critical_dates = NULL,
  interpolate_min = NULL,
  eigen_method = "power_iteration",
  R = NULL,
  weight_Rt = FALSE,
  keep_strains_Rt = FALSE
)

Arguments

time

A vector of time steps

S

A 3d ((n groups x n vaccine classes) x n trajectories x n time steps) array of "S" compartment counts

pars

Either a single lancelot_parameters() object (shared parameters) or an unnamed list of lancelot_parameters() objects, the same length as ncol(S).

prob_strain

A 3d ((n groups x n strains) x n trajectories x n time steps) array of "prob_strain" model outputs. Default is NULL, but it must be specified if there is more than one strain.

initial_time_from_parameters

If TRUE, then time[[1]] is replaced by the value of initial_time from the parameters. This is usually what you want. (From sircovid 0.12.13 this parameter means "initial time is zero" and will probably be updated in a future version).

shared_parameters

Should pars be treated as a single shared list? Leave as NULL to detect automatically, set to TRUE or FALSE to force it to be interpreted one way or the other which may give more easily interpretable error messages.

type

A character vector of possible Rt types to compute. Can be any or all of eff_Rt_all, eff_Rt_general, Rt_all and Rt_general

interpolate_every

Spacing (in days) to use between interpolated points

interpolate_critical_dates

interpolate_min

eigen_method

The eigenvalue method to use (passed to eigen1::eigen1 as method)

R

A 3d ((n groups x n strains x n vaccine classes) x n trajectories x n time steps) array of "R" compartment counts, required for multi-strain models.

weight_Rt

If TRUE then computes the weighted average of the Rt for all strains, otherwise all calculations are returned with an additional dimension to index each strain.

keep_strains_Rt

Value

As for lancelot_Rt(), except that every element is a matrix, not a vector.

Compute Rt using EpiEstim for a set of trajectories

Description

Compute "Rt" using EpiEtim for a set of simulated trajectories (e.g., the result of the ⁠$iterate()⁠ method of lancelot, mcstate::pmcmc() or mcstate::pmcmc_predict(). The trajectories should share parameters.

Usage

lancelot_rt_trajectories_epiestim(
  step,
  incidence,
  p,
  gt_distr = NULL,
  sliding_window_ndays = 7,
  mean_prior = 1,
  sd_prior = 1,
  n_GT = 10000,
  n_R = 1000,
  save_all_Rt_sample = TRUE,
  q = NULL
)
lancelot_rt_trajectories_epiestim(
  step,
  incidence,
  p,
  gt_distr = NULL,
  sliding_window_ndays = 7,
  mean_prior = 1,
  sd_prior = 1,
  n_GT = 10000,
  n_R = 1000,
  save_all_Rt_sample = TRUE,
  q = NULL
)

Arguments

step

A vector of steps

incidence

A matrix (n trajectories x n steps) of incidence counts

p

A single lancelot_parameters() object.

gt_distr

A vector giving the discrete (daily) distribution of the generation time. The first value must be zero. If NULL, the generation time distribution will be determined automatically from p.

sliding_window_ndays

An integer giving the length of the sliding window on which Rt will be estimated

mean_prior

The mean prior for Rt

sd_prior

The standard deviation of the prior for Rt

n_GT

An integer giving the number of generation times to be drawn to construct the discrete distribution of the generation time

n_R

An integer giving the number of Rt values to sample from for each incidence trajectory. These will then be aggregated across all incidence trajectories.

save_all_Rt_sample

A boolean determining whether to save all samples of Rt estimated or only a summary

q

A vector of quantiles to return values for

Value

A list with elements t_start (vector of first days of the sliding windows over which Rt is estimated), t_end (vector of last days of the sliding windows over which Rt is estimated), Rt a matrix (only present if save_all = TRUE) containing for each sliding window (each col in the matrix) a sample of n_R * nrow(inc) values of Rt for that sliding window (rows of the matrix) Rt_summary a matrix containing for each sliding window (each col in the matrix) the quantiles (q) and the mean of Rt for that sliding window (rows of the matrix) gt_distr a vector giving the discrete (daily) distribution of the generation time

Modify severity and transmission of variants

Description

Applies a modifier to severity and transmission parameters

Usage

modify_severity(efficacy, efficacy_strain_2, strain_severity_modifier)
modify_severity(efficacy, efficacy_strain_2, strain_severity_modifier)

Arguments

efficacy, efficacy_strain_2

Vaccine efficacy parameters for strains 1 and 2 respectively. Expects a list with names rel_susceptibility, rel_p_sympt, rel_p_hosp_if_sympt, rel_infectivity, rel_p_death. Element columns correspond to vaccine strata and rows to age groups.

strain_severity_modifier

List of modifiers to be applied to efficacy variables; length should correspond to number of strains and each element should be a list with same names as efficacy

Value

Returns a list with same length and names as efficacy and where each element has dimensions n_groups x n_strains x n_vacc_strata

Regions

Description

Generate lists of regions that we use for various tasks.

Usage

regions(type)
regions(type)

Arguments

type

The name of a region type; must be one of "all", "england", or "nations"

Value

A character vector

Examples

sircovid::regions("england")
sircovid::regions("england")

Reorder Rt or IFR trajectories

Description

Reorder Rt or IFR trajectories

Usage

reorder_rt_ifr(x, rank)
reorder_rt_ifr(x, rank)

Arguments

x

An Rt_trajectories object, as returned by lancelot_Rt_trajectories()

rank

A vector of ranks to reorder by

Value

An Rt_trajectories object with appropriately reordered elements

Reorder samples

Description

Reorder samples according to a predefined ranking

Usage

reorder_sample(sample, rank)
reorder_sample(sample, rank)

Arguments

sample

An mcstate_pmcmc object

rank

A vector of ranks to reorder by

Value

An mcstate_pmcmc object with appropriately reordered elements

Rotate strains

Description

Rotate strains, so that strain 1 becomes the sum of strains 1 and 2 and strain 2 is empty. Use this to allow sequential replacement of strains.

Usage

rotate_strains(state, info, ...)
rotate_strains(state, info, ...)

Arguments

state

Model state

info

Model info

...

Additional arguments, ignored. This exists so that this function can be used as an argument to mcstate::multistage_epoch. Practically your two model informations in this case would be equivalent.

Date handling for sircovid

Description

We need to map "dates" onto dust::dust's concept of model "step" and we do this by mapping a date such as 2020-03-02 into the number of days into 2020 (62 here, with the 1st of January being day 1). We call this integer number a "sircovid date".

Usage

sircovid_date(date)

sircovid_date_as_date(date)

as_sircovid_date(date)

as_date(date)
sircovid_date(date)

sircovid_date_as_date(date)

as_sircovid_date(date)

as_date(date)

Arguments

date

A Date object, or something that can be converted to one, or a "sircovid date"; see Details

Details

There are several related functions here

sircovid_date converts its argument into an R Date object, then applies this tranformation. If the argument is not a Date object or a string representing one, an error will be thrown.
sircovid_date_to_date does the reverse conversion to sircovid_date, converting an integer sircovid date into an R Date
as_sircovid_date does the same conversion as sircovid_date but will assume that an integer already represents a sircovid date and will return it unmodified rather than erroring.
as_date does a string to date conversion, using as.Date() but requiring the dates are in ISO 8601 (YYYY-MM-DD) format (it is a helper that avoids conversion to NA, instead throwing an error)

Value

An integer, being the number of days into 2020

Examples

# Convert dates into sircovid dates:
sircovid::sircovid_date("2020-01-01")
sircovid::sircovid_date(c("2020-03-01", "2020-10-01"))

# Reverse the conversion:
sircovid::sircovid_date_as_date(1)
sircovid::sircovid_date_as_date(c(61, 275))

# Double conversion not possible with sircovid_date...
try(sircovid::sircovid_date(61))
# ...but allowed with as_sircovid_date
sircovid::as_sircovid_date(61)

# Strict date conversion with as_date
sircovid::as_date("2020-03-01")
try(sircovid::as_date("03-01-2020"))
# Convert dates into sircovid dates:
sircovid::sircovid_date("2020-01-01")
sircovid::sircovid_date(c("2020-03-01", "2020-10-01"))

# Reverse the conversion:
sircovid::sircovid_date_as_date(1)
sircovid::sircovid_date_as_date(c(61, 275))

# Double conversion not possible with sircovid_date...
try(sircovid::sircovid_date(61))
# ...but allowed with as_sircovid_date
sircovid::as_sircovid_date(61)

# Strict date conversion with as_date
sircovid::as_date("2020-03-01")
try(sircovid::as_date("03-01-2020"))

Available sircovid Models

Description

List the models available in this package

Usage

sircovid_models()
sircovid_models()

Expand beta steps

Description

Expand value_step based on a series of steps. Use this to convert between the values passed to sircovid_parameters_piecewise_linear() and the actual values for a given set of steps.

Usage

sircovid_parameters_expand_step(step, value_step)
sircovid_parameters_expand_step(step, value_step)

Arguments

step

A vector of steps

value_step

A vector of values

Value

A numeric vector the same length as step

Construct piecewise constant array

Description

Construct a piecewise constant quantity over time array for use within sircovid models.

Usage

sircovid_parameters_piecewise_constant(date, value, dt)
sircovid_parameters_piecewise_constant(date, value, dt)

Arguments

date

Either NULL, if one value of the quantity will be used for all time steps, or a vector of times that will be used as change points. Must be provided as a sircovid_date(), i.e., days into 2020. The first date must be 0.

value

A vector of values to use for the quantity - either a scalar (if date is NULL) or a vector the same length as date.

dt

The timestep that will be used in the simulation. This must be of the form 1 / n where n is an integer representing the number of steps per day. Ordinarily this is set by sircovid internally to be 0.25 but this will become tuneable in a future version.

Value

Returns a vector of piecewise constant values, one per timestep, until the values stabilise. After this point the quantity is assumed to be constant.

Examples

# If "date" is NULL, then the quantity is constant and this function is
# trivial:
sircovid::sircovid_parameters_piecewise_constant(NULL, 0.1, 0.25)

date <- sircovid::sircovid_date(
   c("2019-12-31", "2020-02-01", "2020-02-14", "2020-03-15"))
value <- c(0, 3, 1, 2)
y <- sircovid::sircovid_parameters_piecewise_constant(date, value, 1)

# The implied time series looks like this:
t <- seq(0, date[[4]])
plot(t, y, type = "o")
points(date, value, pch = 19, col = "red")

# After 2020-03-15, the quantity value will be fixed at 2, the value
# that it reached at that date.

# You can see this using sircovid_parameters_expand_step
# If a vector of dates is provided then, it's more complex. We'll
# use dt of 1 here as it's easier to visualise
t <- seq(0, 100, by = 1)
sircovid::sircovid_parameters_expand_step(t, y)
plot(t, sircovid::sircovid_parameters_expand_step(t, y), type = "o")
points(date, value, pch = 19, col = "red")

# If dt is less than 1, this is scaled, but the pattern of
# change is the same
y <- sircovid::sircovid_parameters_piecewise_constant(date, value, 0.5)
t <- seq(0, date[[4]], by = 0.5)
plot(t, y, type = "o", cex = 0.25)
points(date, value, pch = 19, col = "red")
# If "date" is NULL, then the quantity is constant and this function is
# trivial:
sircovid::sircovid_parameters_piecewise_constant(NULL, 0.1, 0.25)

date <- sircovid::sircovid_date(
   c("2019-12-31", "2020-02-01", "2020-02-14", "2020-03-15"))
value <- c(0, 3, 1, 2)
y <- sircovid::sircovid_parameters_piecewise_constant(date, value, 1)

# The implied time series looks like this:
t <- seq(0, date[[4]])
plot(t, y, type = "o")
points(date, value, pch = 19, col = "red")

# After 2020-03-15, the quantity value will be fixed at 2, the value
# that it reached at that date.

# You can see this using sircovid_parameters_expand_step
# If a vector of dates is provided then, it's more complex. We'll
# use dt of 1 here as it's easier to visualise
t <- seq(0, 100, by = 1)
sircovid::sircovid_parameters_expand_step(t, y)
plot(t, sircovid::sircovid_parameters_expand_step(t, y), type = "o")
points(date, value, pch = 19, col = "red")

# If dt is less than 1, this is scaled, but the pattern of
# change is the same
y <- sircovid::sircovid_parameters_piecewise_constant(date, value, 0.5)
t <- seq(0, date[[4]], by = 0.5)
plot(t, y, type = "o", cex = 0.25)
points(date, value, pch = 19, col = "red")

Construct piecewise linear array

Description

Construct a piecewise linear quantity over time array for use within sircovid models.

Usage

sircovid_parameters_piecewise_linear(date, value, dt)
sircovid_parameters_piecewise_linear(date, value, dt)

Arguments

date

value

A vector of values to use for the quantity - either a scalar (if date is NULL) or a vector the same length as date.

dt

Value

Returns a vector of piecewise linear values, one per timestep, until the values stabilise. After this point the quantity is assumed to be constant.

Examples

# If "date" is NULL, then the quantity is constant and this function is
# trivial:
sircovid::sircovid_parameters_piecewise_linear(NULL, 0.1, 0.25)

date <- sircovid::sircovid_date(
   c("2020-02-01", "2020-02-14", "2020-03-15"))
value <- c(3, 1, 2)
y <- sircovid::sircovid_parameters_piecewise_linear(date, value, 1)

# The implied time series looks like this:
t <- seq(0, date[[3]])
plot(t, y, type = "o")
points(date, value, pch = 19, col = "red")

# After 2020-03-15, the quantity value will be fixed at 2, the value
# that it reached at that date.

# You can see this using sircovid_parameters_expand_step
# If a vector of dates is provided then, it's more complex. We'll
# use dt of 1 here as it's easier to visualise
t <- seq(0, 100, by = 1)
sircovid::sircovid_parameters_expand_step(t, y)
plot(t, sircovid::sircovid_parameters_expand_step(t, y), type = "o")
points(date, value, pch = 19, col = "red")

# If dt is less than 1, this is scaled, but the pattern of
# change is the same
y <- sircovid::sircovid_parameters_piecewise_linear(date, value, 0.5)
t <- seq(0, date[[3]], by = 0.5)
plot(t, y, type = "o", cex = 0.25)
points(date, value, pch = 19, col = "red")
# If "date" is NULL, then the quantity is constant and this function is
# trivial:
sircovid::sircovid_parameters_piecewise_linear(NULL, 0.1, 0.25)

date <- sircovid::sircovid_date(
   c("2020-02-01", "2020-02-14", "2020-03-15"))
value <- c(3, 1, 2)
y <- sircovid::sircovid_parameters_piecewise_linear(date, value, 1)

# The implied time series looks like this:
t <- seq(0, date[[3]])
plot(t, y, type = "o")
points(date, value, pch = 19, col = "red")

# After 2020-03-15, the quantity value will be fixed at 2, the value
# that it reached at that date.

# You can see this using sircovid_parameters_expand_step
# If a vector of dates is provided then, it's more complex. We'll
# use dt of 1 here as it's easier to visualise
t <- seq(0, 100, by = 1)
sircovid::sircovid_parameters_expand_step(t, y)
plot(t, sircovid::sircovid_parameters_expand_step(t, y), type = "o")
points(date, value, pch = 19, col = "red")

# If dt is less than 1, this is scaled, but the pattern of
# change is the same
y <- sircovid::sircovid_parameters_piecewise_linear(date, value, 0.5)
t <- seq(0, date[[3]], by = 0.5)
plot(t, y, type = "o", cex = 0.25)
points(date, value, pch = 19, col = "red")

Process severity data

Description

Process severity data

Usage

sircovid_parameters_severity(params)
sircovid_parameters_severity(params)

Arguments

params

Value

A list of age-structured probabilities

Upgrade model state after change

Description

Upgrade model state

Usage

upgrade_state(state_orig, info_orig, info_new, allowed = NULL)
upgrade_state(state_orig, info_orig, info_new, allowed = NULL)

Arguments

state_orig

Old model info

info_orig

Old model info

info_new

New model info

allowed

Character vector of allowed states to add; new values will be initialised as zeros.

Value

A new copy of the state

Create vaccine eligibility vector

Description

Create a vector of vaccine eligibility, based on a minimum age. The assumption is that everyone below that age is ineligible and everyone including and above is eligibility. This will practically be modified further by assumptions of uptake. Note that carehome residents and workers are always eligible for vaccination, regardless of the minimum age.

Usage

vaccine_eligibility(min_age)
vaccine_eligibility(min_age)

Arguments

min_age

Minimum age to be eligible for vaccination (includes this age).

Value

A vector of length 19

Compute vaccination order

Description

Compute vaccination priority following JCVI ordering.

Usage

vaccine_priority_proportion(
  uptake,
  prop_hcw = NULL,
  prop_very_vulnerable = NULL,
  prop_underlying_condition = NULL
)

vaccine_priority_population(
  region,
  uptake,
  prop_hcw = NULL,
  prop_very_vulnerable = NULL,
  prop_underlying_condition = NULL,
  carehomes = TRUE
)
vaccine_priority_proportion(
  uptake,
  prop_hcw = NULL,
  prop_very_vulnerable = NULL,
  prop_underlying_condition = NULL
)

vaccine_priority_population(
  region,
  uptake,
  prop_hcw = NULL,
  prop_very_vulnerable = NULL,
  prop_underlying_condition = NULL,
  carehomes = TRUE
)

Arguments

uptake

A vector of length 19 with fractional uptake per group. If a single number is given it is shared across all groups (note that this includes under-18s)

prop_hcw

Assumed fraction of healthcare workers in each group (length 19) - if NULL we use a default that is a guess with hopefully the right general shape.

prop_very_vulnerable

Assumed fraction "very vulnerable" in each group (length 19) - if NULL we use a default that is a guess with hopefully the right general shape.

prop_underlying_condition

Assumed fraction "underlying condition" in each group (length 19) - if NULL we use a default that is a guess with hopefully the right general shape.

region

Region to use to get total population numbers

carehomes

Logical parameter, whether or not we have carehomes in the model. Default is TRUE

Details

https://tinyurl.com/8uwtatvm

Assmuing independance between job (e.g. HCW) and clinical condition

But assuming one is either counted as "clinically extremely vulnerable" or "with underlying health conditions" but not both

The JCVI priority groups, in decending order are:

residents in a care home for older adults and their carers
all those 80 years of age and over and frontline health and social care workers
all those 75 years of age and over
all those 70 years of age and over and clinically extremely vulnerable individuals
all those 65 years of age and over
all individuals aged 16 years to 64 years with underlying health conditions which put them at higher risk of serious disease and mortality
all those 60 years of age and over
all those 55 years of age and over
all those 50 years of age and over
all those 40-49 years of age and over
all those 30-39 years of age and over
all those 18-29 years of age and over

Value

A matrix with n_groups rows (19) and columns representing priority groups, and element (i, j) is the proportion in group i who should be vaccinated as part of priority group j, accounting for uptake so that the sum over the rows corresponds to the total fractional uptake in that group. For vaccine_priority_population, the total number of individuals replaces the proportion (based on the demography used by sircovid).

Remap state to include vaccination

Description

Remap state that was run with one stratum to support vaccination. This is useful where fitting was done pre-vaccination and the model state needs to be expanded to support vaccination for simulations.

Usage

vaccine_remap_state(state_orig, info_orig, info_vacc)
vaccine_remap_state(state_orig, info_orig, info_vacc)

Arguments

state_orig

An original state matrix (rows representing state, columns representing particles or samples). This must be the complete model state.

info_orig

The results of info() from the model object without vaccination

info_vacc

The results of info() from the model object with vaccination

Value

A 3d array with more rows than state.

Create vaccine schedule

Description

Create a vaccine schedule for use with lancelot_parameters

Usage

vaccine_schedule(date, doses, n_doses = 2L)
vaccine_schedule(date, doses, n_doses = 2L)

Arguments

date

A single date, representing the first day that vaccines will be given

doses

A 3d array of doses representing (1) the model group (19 rows for the lancelot model), (2) the dose (must be length 2 at present) and (3) time (can be anything nonzero). The values represent the number of vaccine doses in that group for that dose for that day. So for doses[i, j, k] then it is for the ith group, the number of jth doses on day (k - 1) + date

n_doses

The number of doses in the schedule. Typically (and by default) this will be 2, but if using booster doses 3 (and in future we may extend further).

Value

A vaccine_schedule object

Create historical vaccine schedule

Description

Create a historical vaccine schedule from data

Usage

vaccine_schedule_from_data(data, region, uptake, carehomes = TRUE)
vaccine_schedule_from_data(data, region, uptake, carehomes = TRUE)

Arguments

data

A data.frame with columns date, age_band_min, and numbered doses columns, e.g. if there are three doses these should be dose1, dose2 and dose3. Values of age_band_min should be either multiples of 5 or NA - the latter means those doses are not age-specific and will be distributed across all ages according to priority after all age-specific doses have already been dealt with

region

Region to use to get total population numbers

uptake

A matrix of 19 rows, and number of columns equal to number of doses. The (i,j)th entry gives the fractional uptake of dose j for group i. Should be non-increasing across rows

carehomes

Logical parameter, whether or not we have carehomes in the model. Default is TRUE

Value

A vaccine_schedule object

Create vaccination schedule

Description

Create future vaccination schedule from a projected number of daily doses.

Usage

vaccine_schedule_future(
  start,
  daily_doses_value,
  mean_days_between_doses,
  priority_population,
  lag_groups = NULL,
  lag_days = NULL,
  booster_daily_doses_value = NULL,
  booster_proportion = rep(1L, 19)
)
vaccine_schedule_future(
  start,
  daily_doses_value,
  mean_days_between_doses,
  priority_population,
  lag_groups = NULL,
  lag_days = NULL,
  booster_daily_doses_value = NULL,
  booster_proportion = rep(1L, 19)
)

Arguments

start

Either a sircovid_date object corresponding to the first date in daily_doses_value, or a vaccine_schedule object corresponding to previously carried out vaccination.

daily_doses_value

A vector of doses per day.

mean_days_between_doses

Assumed mean days between doses one and two

priority_population

Output from vaccine_priority_population, giving the number of people to vaccinate in each age (row) and priority group (column)

lag_groups

Row indices, corresponding to age groups in which a lag should be added to the start time of the dose schedule returned by vaccine_schedule, if NULL then no lag is added. Ignored if lag_groups is NULL.

lag_days

If lag_groups is not NULL then specifies the number of days to add the start of the dose schedule for the given groups. Ignored if lag_groups is NULL.

booster_daily_doses_value

A vector of booster doses per day.

booster_proportion

Proportion of the groups in priority_population to boost, default is all groups; ignored if booster_daily_doses_value is NULL.

High-level vaccine scenario creation

Description

Create a vaccination scenario

Usage

vaccine_schedule_scenario(
  schedule_past,
  doses_future,
  end_date,
  mean_days_between_doses,
  priority_population,
  lag_groups = NULL,
  lag_days = NULL,
  boosters_future = NULL,
  boosters_prepend_zero = TRUE,
  booster_proportion = rep(1L, 19)
)
vaccine_schedule_scenario(
  schedule_past,
  doses_future,
  end_date,
  mean_days_between_doses,
  priority_population,
  lag_groups = NULL,
  lag_days = NULL,
  boosters_future = NULL,
  boosters_prepend_zero = TRUE,
  booster_proportion = rep(1L, 19)
)

Arguments

schedule_past

A vaccine_schedule object corresponding to previously carried out vaccination.

doses_future

A named vector of vaccine doses to give in the future. Names must be in ISO date format.

end_date

The final day in the future to create a schedule for. After this date the model will assume 0 vaccine doses given so an overestimate is probably better than an underestimate.

mean_days_between_doses

Assumed mean days between doses one and two

priority_population

Output from vaccine_priority_population, giving the number of people to vaccinate in each age (row) and priority group (column)

lag_groups

lag_days

If lag_groups is not NULL then specifies the number of days to add the start of the dose schedule for the given groups. Ignored if lag_groups is NULL.

boosters_future

Optional named vector of booster doses to give in the future. Names must be in ISO date format.

boosters_prepend_zero

If TRUE (default) and boosters_future is not NULL then sets booster doses to zero before the first date in boosters_future. This is in contrast to when it is FALSE and the previous value in schedule_past is replicated until the first date in boosters_future. Note that this should rarely be FALSE as this will likely lead to duplicating daily doses that are already replicated in doses_future.

booster_proportion

Proportion of the groups in priority_population to boost, default is all groups; ignored if booster_daily_doses_value is NULL.

Value

A vaccine_schedule object

Package 'sircovid'

Help Index

Add incidence to trajectories

Description

Usage

Arguments

The basic sircovid model

Description

Examples

Compare observed and modelled data for the basic model

Description

Usage

Arguments

Value

Examples

Index of basic model

Description

Usage

Arguments

Value

Examples

Initial conditions for the basic model

Description

Usage

Arguments

Value

Examples

Parameters for the basic model

Description

Usage

Arguments

Value

Examples

Check a model type

Description

Usage

Arguments

Combine Rt estimates

Description

Usage

Arguments

Value

Combine Rt estimates from EpiEstim

Description

Usage

Arguments

Value

Combine trajectories

Description

Usage

Arguments

Value

Create GPU model

Description

Usage

Arguments

Value

Get the ranking of sample by a given variable, e.g. infections

Description

Usage

Arguments

Value

Inflate strains in model state

Description

Usage

Arguments

Value

Author(s)

Inflate vacc classes in model state

Description

Usage

Arguments

Value

The lancelot sircovid model

Description

Check data for particle filter

Description

Usage

Arguments

Value