Skip to contents

Observed-versus-Model Marginal Resistance Validation

Source: R/daly_resistance_validation.R
validate_marginal_calibration.Rd

For every hospital x pathogen x antibiotic-class cell with adequate observed testing support, compares the observed marginal resistance rate against the model's fitted probability \(\Phi(\mu_{ed})\), averaged over the same tested-event cohort and over posterior draws. This is a model-validation calculation: it uses the fitted probability for TESTED cells (unlike compute_event_profile_probabilities(), which never overwrites an observed value). A calibration residual near 0 and interval coverage near the nominal CI level indicate the model reproduces the observed marginal resistance rates; it does not by itself validate joint (pairwise/profile) structure.

Usage

validate_marginal_calibration(
  fitted_model,
  n_posterior_draws_for_validation = 2000L,
  seed = 123L,
  ci_level = 0.95,
  min_tested = NULL,
  min_resistant = NULL,
  min_susceptible = NULL
)

Arguments

fitted_model

List returned by fit_bayesian_multivariate_probit().

n_posterior_draws_for_validation

Integer. Posterior draws used. Default 2000L.

seed

Integer. Random seed (draw subsampling only). Default 123L.

ci_level

Numeric. Credible interval coverage. Default 0.95.

min_tested, min_resistant, min_susceptible

Integer or NULL. Eligibility thresholds. When all NULL (default), eligibility is taken from fitted_model$eligibility_report$marginal (the same approved rules used for profile panels) rather than re-deriving new thresholds. Supplying any of these overrides that report and applies the supplied thresholds instead.

Value

Tibble, one row per eligible hospital x pathogen x class cell: n_events, n_tested, n_resistant, n_susceptible, observed_resistance, model_resistance_mean/lower/upper, absolute_error, calibration_residual, interval_contains_observed.