Computes the fatal prevalence of resistance R_kd for each pathogen k and resistance profile delta, using per-profile mortality odds ratios (OR_death) from fit_mortality_rr_logistic() in place of LOS relative risks. The formula is identical to compute_fraction_associated() – only the RR source changes.

Usage

compute_R_kd_fatal(
  profiles_with_rr,
  probability_col = "probability",
  rr_profile_col = "RR_LOS_profile"
)

Arguments

profiles_with_rr: Named list from assign_rr_to_profiles() called with rr_col = "OR_death".
probability_col: Character. Profile probability column. Default "probability".
rr_profile_col: Character. Profile-level OR column as produced by assign_rr_to_profiles(). Default "RR_LOS_profile".

Value

Named list (one entry per pathogen) containing:

per_profile: profile data frame augmented with numerator_R_kd (= $R'_{K\delta} \cdot \text{OR}_{K\delta}$) and R_kd (= numerator / E[OR_k]) per profile.
R_k: overall fatal prevalence of resistance for pathogen k (sum of R_kd over resistant profiles only).
E_OR_k: expected OR = $\sum_\delta R'_{K\delta} \cdot \text{OR}_{K\delta}$.

Details

$$E[\text{OR}_k] = \sum_\delta R'_{K\delta} \cdot \text{OR}_{K\delta}$$

$$R_{kd} = \frac{R'_{K\delta} \cdot \text{OR}_{K\delta}}{E[\text{OR}_k]}$$

$$R_{k} = \sum_{\delta \neq \delta_0} R_{kd}$$

where $R'_{K\delta}$ is the profile probability from compute_resistance_profiles() and $\text{OR}_{K\delta}$ is assigned by assign_rr_to_profiles() with rr_col = "OR_death" (the all-susceptible profile carries OR = 1).

Usage pipeline:


  # 1. Fit mortality OR per class
  mort_or <- fit_mortality_rr_logistic(data, ...)

  # 2. Assign OR to profiles via max rule
  profiles_with_or <- assign_rr_to_profiles(
      profiles_output,
      rr_table = mort_or,
      rr_col   = "OR_death"
  )

  # 3. Compute R_kd (fatal prevalence of resistance)
  rkd <- compute_R_kd_fatal(profiles_with_or)

Compute Fatal Prevalence of Resistance (R_kd)

Usage

Arguments

Value

Details