Comprehensive ROC Analysis with Advanced Features

Value

A psychopdarocResults object containing:

resultsTable: Detailed results for each threshold
simpleResultsTable: Summary AUC results with confidence intervals
sensSpecTable: Confusion matrix at optimal cutpoint
plotROC: ROC curve visualization
delongTest: DeLong test results (if requested)
idiTable: IDI results with confidence intervals (if requested)
nriTable: NRI results with confidence intervals (if requested)
Additional plots and tables based on options selected

Details

This function provides an extensive ROC analysis toolkit that goes beyond basic ROC curve generation. Key features include:

Core ROC Analysis:

AUC calculation with confidence intervals
Multiple cutpoint optimization methods (12 different approaches)
16 different optimization metrics (Youden, accuracy, F1, etc.)
Bootstrap confidence intervals
Manual cutpoint specification

Advanced Statistical Methods:

DeLong test for comparing multiple AUCs
IDI (Integrated Discrimination Index) with bootstrap CI
NRI (Net Reclassification Index) with bootstrap CI
Partial AUC calculations
ROC curve smoothing (multiple methods)
Classifier performance comparison

Visualization Options:

ROC curves (individual and combined)
Sensitivity/specificity vs threshold plots
Predictive value vs prevalence plots
Precision-recall curves
Dot plots showing class distributions
Interactive ROC plots
Confidence bands and quantile confidence intervals

Subgroup Analysis:

Stratified analysis by grouping variables
Cost-benefit optimization with custom cost ratios
Hospital/site comparisons

Note

This function originally developed by Lucas Friesen in the psychoPDA module. Enhanced version with additional features added to the ClinicoPath module.

References

DeLong, E. R., DeLong, D. M., & Clarke-Pearson, D. L. (1988). Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics, 44(3), 837-845.

Pencina, M. J., D'Agostino, R. B., D'Agostino, R. B., & Vasan, R. S. (2008). Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond. Statistics in Medicine, 27(2), 157-172.

Youden, W. J. (1950). Index for rating diagnostic tests. Cancer, 3(1), 32-35.

Super classes

jmvcore::Analysis -> ClinicoPath::psychopdarocBase -> psychopdarocClass

Methods

Public methods

psychopdarocClass$clone()

Inherited methods

Method `clone()`

The objects of this class are cloneable with this method.

Usage

psychopdarocClass$clone(deep = FALSE)

Arguments

deep: Whether to make a deep clone.

Examples

if (FALSE) { # \dontrun{
# Load example medical data
data(medical_roc_data)

# Basic ROC analysis
result1 <- psychopdaroc(
  data = medical_roc_data,
  dependentVars = "biomarker1",
  classVar = "disease_status", 
  positiveClass = "Disease"
)

# Compare multiple biomarkers with DeLong test
result2 <- psychopdaroc(
  data = medical_roc_data,
  dependentVars = c("biomarker1", "biomarker2", "biomarker3"),
  classVar = "disease_status",
  positiveClass = "Disease",
  delongTest = TRUE,
  combinePlots = TRUE
)

# Advanced analysis with IDI/NRI
result3 <- psychopdaroc(
  data = medical_roc_data,
  dependentVars = c("biomarker1", "biomarker2"),
  classVar = "disease_status",
  positiveClass = "Disease", 
  calculateIDI = TRUE,
  calculateNRI = TRUE,
  refVar = "biomarker1",
  nriThresholds = "0.3,0.7"
)

# Cost-benefit optimization
result4 <- psychopdaroc(
  data = medical_roc_data,
  dependentVars = "biomarker1",
  classVar = "disease_status",
  positiveClass = "Disease",
  method = "oc_cost_ratio",
  costratioFP = 2.5  # False positives cost 2.5x false negatives
)

# Subgroup analysis by hospital
result5 <- psychopdaroc(
  data = medical_roc_data,
  dependentVars = "biomarker1",
  classVar = "disease_status",
  positiveClass = "Disease",
  subGroup = "hospital"
)

# Comprehensive analysis with all features
result6 <- psychopdaroc(
  data = medical_roc_data,
  dependentVars = c("biomarker1", "biomarker2"),
  classVar = "disease_status",
  positiveClass = "Disease",
  method = "maximize_metric",
  metric = "youden",
  plotROC = TRUE,
  sensSpecTable = TRUE,
  showThresholdTable = TRUE,
  delongTest = TRUE,
  calculateIDI = TRUE,
  partialAUC = TRUE,
  bootstrapCI = TRUE,
  precisionRecallCurve = TRUE,
  compareClassifiers = TRUE
)

# Financial risk assessment example
data(financial_roc_data)

financial_result <- psychopdaroc(
  data = financial_roc_data,
  dependentVars = c("credit_score", "income_debt_ratio", "employment_score"),
  classVar = "default_status",
  positiveClass = "Default",
  direction = "<=",  # Lower credit scores indicate higher risk
  method = "oc_cost_ratio",
  costratioFP = 0.1,  # False positives (rejected good clients) cost less
  delongTest = TRUE,
  subGroup = "client_type"
)

# Educational assessment example
data(education_roc_data)

education_result <- psychopdaroc(
  data = education_roc_data,
  dependentVars = c("exam_score", "project_score", "peer_score"),
  classVar = "pass_status",
  positiveClass = "Pass",
  method = "maximize_metric",
  metric = "accuracy",
  calculateIDI = TRUE,
  refVar = "exam_score",
  subGroup = "class_section"
)

# Manufacturing quality control example
data(manufacturing_roc_data)

quality_result <- psychopdaroc(
  data = manufacturing_roc_data, 
  dependentVars = c("dimension_score", "surface_score", "strength_score"),
  classVar = "quality_status",
  positiveClass = "Defect",
  method = "oc_equal_sens_spec",  # Balanced sensitivity/specificity
  plotROC = TRUE,
  showCriterionPlot = TRUE,
  showDotPlot = TRUE,
  subGroup = "production_line"
)
} # }