Skip to contents

Dataset with 160 patients including cost information for false positive and false negative errors, enabling cost-benefit optimal cutpoint determination.

Usage

psychopdaROC_costbenefit

Format

A data frame with 160 rows and 5 variables:

patient_id

Character: Patient identifier (PT001-PT160)

outcome

Factor: "Event" or "No_Event" (28%/72% prevalence)

risk_score

Numeric: Risk prediction score (mean: 75 for event, 50 for no event)

false_positive_cost

Numeric: Cost of false positive ($100)

false_negative_cost

Numeric: Cost of false negative ($1000)

Source

Generated test data for ClinicoPath package

Details

False negative cost is 10× higher than false positive cost, representing clinical scenarios where missing a case is much more costly than a false alarm. Enables cost-ratio optimal cutpoint method.

Examples

data(psychopdaROC_costbenefit)
psychopdaROC(data = psychopdaROC_costbenefit, dependentVars = "risk_score",
             classVar = "outcome", positiveClass = "Event",
             method = "oc_cost_ratio")
#> Error in psychopdaROC(data = psychopdaROC_costbenefit, dependentVars = "risk_score",     classVar = "outcome", positiveClass = "Event", method = "oc_cost_ratio"): argument "refVar" is missing, with no default