Skip to contents

Plot Infection Type (HAI/CAI) by Location (ICU/Ward), Faceted by Hospital

Source: R/prep_eda_plots.R
plot_infection_type_by_location.Rd

100 types (e.g. HAI vs CAI) among patients in each location (e.g. ICU vs Ward). Each bar is normalised within its location, so the two bars in a panel directly compare \(P(\text{infection type} \mid \text{location})\) – i.e. whether infection type is associated with location, per hospital.

Usage

plot_infection_type_by_location(
  data,
  patient_col = "PatientInformation_id",
  center_col = "center_name",
  location_col = "location",
  infection_col = "infection_type",
  location_levels = c("ICU", "Ward"),
  infection_levels = c("HAI", "CAI"),
  style = c("bar", "heatmap"),
  colours = NULL,
  ncol = 5,
  base_size = 12,
  show_counts = TRUE,
  title = NULL
)

Arguments

data

Data frame. Long-format AMR / cleaned stewardship data.

patient_col

Character. Patient ID column. Default "PatientInformation_id".

center_col

Character. Centre/facility column. Default "center_name".

location_col

Character. Location column. Default "location".

infection_col

Character. Infection-type column. Default "infection_type".

location_levels

Character vector. Location values to keep (become the x-axis bars). Default c("ICU", "Ward").

infection_levels

Character vector. Infection-type values to keep (become the stacked fills). Default c("HAI", "CAI").

style

Character. "bar" (default) draws a 100 bar per hospital (x = location, fill = infection type, normalised within location). "heatmap" draws a tile grid per hospital (x = location, y = infection type, fill = within-hospital joint proportion).

colours

Named character vector of fill colours keyed by infection_levels. Auto-generated if NULL.

ncol

Integer. Number of facet columns. Default 5.

base_size

Numeric. Base font size. Default 12.

show_counts

Logical. Print unique-patient counts inside each segment. Default TRUE.

title

Character. Custom title. Auto-generated if NULL.

Value

A ggplot object (one panel per centre).

Details

Deduplication: unique patients per centre x location x infection type. Only the requested location_levels and infection_levels are kept (other/blank/compound values are dropped).