Day-weighted exposure over windows

Exposure rows cover intervals (a month, a residence).

Usage

dayWeightedExposure(exposure, windows = NULL, valueColumn = "value_as_number")

Arguments

exposure

A data frame with person_id, exposure_start_date, exposure_end_date and a value column.

windows

A data frame with person_id, window_start and window_end, one row per window, optionally with window_id. If NULL, each person's whole exposure period is the window.

valueColumn

Name of the value column of exposure.

Value

A data frame with, per window: weighted_mean (day-weighted mean), naive_mean (mean of the overlapping rows), n_rows, days_covered, window_days and coverage (share of the window days with exposure data; below 1 when the window is only partly observed or exposure rows overlap).

Examples

exposure <- data.frame(
  person_id = 1,
  exposure_start_date = as.Date(c("2016-01-01", "2016-02-01")),
  exposure_end_date = as.Date(c("2016-01-31", "2016-02-29")),
  value_as_number = c(10, 20)
)
windows <- data.frame(
  person_id = 1,
  window_start = as.Date("2016-01-16"),
  window_end = as.Date("2016-02-10")
)
dayWeightedExposure(exposure, windows)
#>   window_id person_id window_start window_end window_days weighted_mean
#> 1         1         1   2016-01-16 2016-02-10          26      13.84615
#>   naive_mean n_rows days_covered coverage
#> 1         15      2           26        1