Computes the fatal pathogen distribution for a given infectious syndrome (L) using facility-level microbiology data. This quantity represents the fractional contribution of each pathogen (K) to fatal infection cases, and is used in Step 4 of YLL estimation (Eq. 11).

Usage

calculate_P_Lk(
  data,
  syndrome_col,
  syndrome_name,
  polymicrobial_col,
  patient_col,
  pathogen_col,
  outcome_col,
  death_value = "Death",
  specimen_col = NULL,
  specimen_name = NULL,
  glass_ref = NULL,
  facility_col = NULL,
  facility_name = NULL,
  pathogen_name = NULL
)

Arguments

data: Data frame of facility-level microbiology records.
syndrome_col: Character. Column containing infectious syndrome labels (L).
syndrome_name: Character. Syndrome to analyse (e.g. "Bloodstream infection").
polymicrobial_col: Character. Column flagging polymicrobial patients (1 = polymicrobial, 0 = monomicrobial).
patient_col: Character. Unique patient identifier column.
pathogen_col: Character. Pathogen (organism) column (K).
outcome_col: Character. Final patient outcome column.
death_value: Character. Value indicating a fatal outcome. Default "Death".
specimen_col: Character or NULL. Specimen type column. When supplied, specimen_name must also be provided.
specimen_name: Character or NULL. Specimen value to restrict to (e.g. "Blood").
glass_ref: Character vector of valid pathogen names, or a data frame with columns specimen and pathogen. When supplied, polymicrobial patients are restricted to pathogens on the GLASS list for the given specimen. Monomicrobial patients are never filtered. NULL skips GLASS filtering.
facility_col: Character or NULL. Facility identifier column. When supplied without facility_name, returns both facility-level and pooled P_LK.
facility_name: Character or NULL. Restricts computation to a single named facility.
pathogen_name: Character vector or NULL. Restricts output to specific pathogen(s).

Value

A named list:

P_Lk: Data frame of pooled P_LK across facilities. Columns: pathogen_col, N_F_LK (weighted fatal patient count for K), N_F_L (total fatal patients for L), P_Lk (fraction).
facility_level: Per-facility P_LK data frame (only when facility_col is supplied and facility_name is NULL).

Details

Unit of analysis: the patient. Each fatal patient contributes a total weight of 1, split equally across their valid pathogens: $$w_{r,k} = \frac{1}{m_r}$$ where $m_r$ is the number of distinct pathogens for patient $r$. Monomicrobial patients ($m_r = 1$) receive weight 1; polymicrobial patients have their death split equally (Approach 1).

Pooled formula across facilities: $$P_{LK}^{\text{pooled}} = \frac{\sum_f N^{F}_{f,L,K}}{\sum_f N^{F}_{f,L}}$$

References

Bhaswati Ganguli. DALY Methodology for AMR (YLD notes). March 2026.

Examples

if (FALSE) { # \dontrun{
pkL <- calculate_P_Lk(
  data              = bsi_data,
  syndrome_col      = "infectious_syndrome",
  syndrome_name     = "Bloodstream infection",
  polymicrobial_col = "is_polymicrobial",
  patient_col       = "PatientInformation_id",
  pathogen_col      = "organism_name",
  outcome_col       = "final_outcome",
  death_value       = "Death",
  facility_col      = "center_name"
)
pkL$P_Lk
pkL$facility_level
} # }