Joint PIF: Combine impact fractions from independent and uncorrelated risk factors

A function to combine multiple impact fractions from independent uncorrelated risk factors. The function follows the formula from Ezzati et al. (2003) : \[ \text{PIF} = 1 - \prod\limits_{k} \big(1 - \text{PIF}_i) \]

Usage

pif_combine(pif_class_1, pif_class_2, ...)

Arguments

pif_class_1: A pif_class object (i.e. an estimation of pif or paf)
pif_class_2: Another pif_class object (i.e. an estimation of a different pif or paf)
...: Additional pif_class objects (i.e. estimations of pif or paf)

Value

A pif_aggregated object estimating the potential impact fraction for individual-level data from multiple distinct impact fractions.

References

Ezzati M, Vander Hoorn S, Rodgers A, Lopez AD, Mathers CD, Murray CJ (2003). “Estimates of global and regional potential health gains from reducing multiple major risk factors.” The Lancet, 362(9380), 271--280.

Examples

# Use the ensanut dataset
data(ensanut)

# EXAMPLE 1
# Setup the survey design
options(survey.lonely.psu = "adjust")
design <- survey::svydesign(data = ensanut, ids = ~1, weights = ~weight, strata = ~strata)
rr_1 <- function(X, theta) {
  exp(-2 +
    theta[1] * X[, "age"] + theta[2] * X[, "systolic_blood_pressure"] / 100)
}

rr_2 <- function(X, theta) {
  exp(-1 +
    theta[1] * X[, "age"] + X[, "systolic_blood_pressure"] )
}

rr_3 <- function(X, theta) {
  exp(-1 +
    theta[1] *X[, "systolic_blood_pressure"] )
}

cft <- function(X) {
  X[, "systolic_blood_pressure"] <- X[, "systolic_blood_pressure"] - 5
  return(X)
}

pif_1 <- pif(design,
  theta = log(c(1.25, 1.68)), rr_1, cft,
  additional_theta_arguments = c(0.01, 0.03), n_bootstrap_samples = 10,
)

pif_2 <- pif(design,
  theta = log(c(1.12, 1.45)), rr_2, cft,
  additional_theta_arguments = c(0.02, 0.025), n_bootstrap_samples = 10,
)

pif_3 <- pif(design,
  theta = 0.24, rr_3, cft,
  additional_theta_arguments = 0.01, n_bootstrap_samples = 10,
)

pif_combine(pif_1, pif_2, pif_3)
#> ── Potential Impact Fraction (PIF) ─────────────────────────────────────────────
#> # A tibble: 3 × 4
#>   counterfactual   relative_risk   potential_impact_fraction type          
#>   <chr>            <chr>                               <dbl> <chr>         
#> 1 Counterfactual_1 Relative_Risk_1                     0.998 point_estimate
#> 2 Counterfactual_1 Relative_Risk_1                     0.996 Lower 2.5%    
#> 3 Counterfactual_1 Relative_Risk_1                     1.00  Upper 97.5%   
#> ────────────────────────────────────────────────────────────────────────────────
#> • Number of bootstrap simulations: 10
#> ✖ A low number of bootstrap simulations will result in an unstable estimate.
#> • Use `as.data.frame` to access values.
#> • Use `summary` to save list of main results.