Priors on probability of diagnosis

This function returns a keyed list of priors related to probability of diagnosis. Called with no arguments, the default values are returned. Custom hyperpriors can be specified by passing values for the parameters specified below. The return value of this function must be added to a covidestim object using the addition operator (see examples).

priors_diagnosis(
  rr_diag_asy_vs_sym = c(2, 18),
  rr_diag_sym_vs_sev = c(2, 2),
  p_diag_if_sev = c(20, 5)
)

Arguments

rr_diag_asy_vs_sym

A two-element numeric vector containing c(alpha, beta) parameters/hyperpriors of a Beta distribution modeling the rate ratio of diagnosis among asymptomatic versus symptomatic, but non-"severe" infections.

Lowering the mean of this prior would reflect a belief that even fewer asymptomatic indivivduals are diagnosed than currently assumed.

rr_diag_sym_vs_sev

A two-element numeric vector containing c(alpha, beta) parameters/hyperpriors of a Beta distribution modeling the rate ratio of diagnosis at the symptomatic vs severe stage of infection.

p_diag_if_sev

A two-element numeric vector containing c(alpha, beta) parameters/hyperpriors of a Beta distribution modeling the probability of being diagnosed if severely ill.

Value

An S3 object of class 'priors'

Details

Boundary avoiding priors are used by default; a weakly informative prior is used for the probability of diagnosis if severely ill.

Examples

cfg <- covidestim(ndays = 50, region = 'New York') + priors_diagnosis(p_diag_if_sev = c(2, 2))
#> Error in covidestim(ndays = 50, region = "New York"): unused argument (ndays = 50)