Skip to contents

Calculate the probability imprinting occurs n years after birth

Source: R/calculation_funs.R
get_p_infection_year.Rd

Given an individual's birth year, the year of observation, and pre-calculated influenza circulation intensities, calculate the probability that the first influenza infection occurs exactly 0, 1, 2, ... 12 years after birth.

Usage

get_p_infection_year(
  birth_year,
  observation_year,
  intensity_df,
  max_year,
  baseline_annual_p_infection = 0.28
)

Arguments

birth_year

year of birth (numeric). Must be between 1918 and the current calendar year.

observation_year

year of observation, which affects the birth cohort's age.

intensity_df

data frame of annual intensities, output by get_country_intensity_data().

max_year

maximum year for which to output probabilities. Must be greater than or equal to observation_year. (If in doubt, set equal to observation year.)

baseline_annual_p_infection

average annual probability of primary infection. The default, 0.28, was estimated using age-seroprevalence data in doi:10.1126/science.aag1322 Gostic et al. Science, (2016).

Value

a vector whose entries show the probability that a person born in year 0 was first infected by influenza in year 0, 1, 2, 3, ...12 We only consider the first 13 probabilities (i.e. we assume everyone imprints before age 13. These outputs are not normalized, so the vector sum asymptotically approaches one, but is not exactly equal to one. For cohorts born <13 years prior to the year of observation, the output vector will have less than 13 entries.

Details

The probability of primary influenza infection n years after birth is calculated based on a modified geometric distribution: let p be the average annual probability of a primary influenza infection. Then the probability that primary infection occurs n=0,1,2,... years after birth is \(p*(1-p)^{n}\).

This function modifies the geometric model above to account for changes in annual circulation intensity, so that annual probabilities of primary infection \(p_i\) are scaled by the intensity in calendar year i. Details are given in doi:10.1126/science.aag1322 Gostic et al. Science, (2016).

Examples

# For a cohort under 12 years old and born in 2000, return the
# probabilities of primary infection in 2000, 2001, ... 2012:
get_p_infection_year(
  birth_year = 2000,
  observation_year = 2022,
  intensity_df = get_country_intensity_data("Canada", 2022),
  max_year = 2022
)
#>  [1] 0.0024984455 0.0000000000 0.0000000000 0.0180147686 0.0000000000
#>  [6] 0.2742563000 0.0000000000 0.0008075884 0.0078780356 0.4875814033
#> [11] 0.0270627329 0.0665796926 0.0393163301

# If the cohort is still under age 12 at the time of observation, return
# a truncated vector of probabilities:
get_p_infection_year(
  birth_year = 2020,
  observation_year = 2022,
  intensity_df = get_country_intensity_data("Mexico", 2022),
  max_year = 2022
)
#> [1] 0.10778376 0.04336504 0.30279601