Skip to contents

Validation of patient-level prediction models with the HADES PatientLevelPrediction Package

Source: R/Module-PatientLevelPredictionValidation.R
PatientLevelPredictionValidationModule.Rd

Module for performing patient-level prediction model validation for models built using the PatientLevelPrediction package.

Super class

Strategus::StrategusModule -> PatientLevelPredictionValidationModule

Public fields

tablePrefix

The table prefix to append to the results tables

Methods

Public methods

Inherited methods

Method new()

Initialize the module

Usage

PatientLevelPredictionValidationModule$new()

Method execute()

Executes the PatientLevelPrediction package to validate a PLP model

Usage

PatientLevelPredictionValidationModule$execute(
  connectionDetails,
  analysisSpecifications,
  executionSettings
)

Arguments

connectionDetails

An object of class connectionDetails as created by the DatabaseConnector::createConnectionDetails() function.

analysisSpecifications

An object of type AnalysisSpecifications as created by createEmptyAnalysisSpecificiations().

analysisSpecifications

An object of type AnalysisSpecifications as created by createEmptyAnalysisSpecificiations().

executionSettings

An object of type ExecutionSettings as created by createCdmExecutionSettings() or createResultsExecutionSettings().

Method createResultsDataModel()

Create the results data model for the module

Usage

PatientLevelPredictionValidationModule$createResultsDataModel(
  resultsConnectionDetails,
  resultsDatabaseSchema,
  tablePrefix = self$tablePrefix
)

Arguments

resultsConnectionDetails

The connection details to the results database which is an object of class connectionDetails as created by the DatabaseConnector::createConnectionDetails() function.

resultsConnectionDetails

The connection details to the results database which is an object of class connectionDetails as created by the DatabaseConnector::createConnectionDetails() function.

resultsDatabaseSchema

The schema in the results database that holds the results data model.

tablePrefix

A prefix to apply to the database table names (optional).

Method uploadResults()

Upload the results for the module

Usage

PatientLevelPredictionValidationModule$uploadResults(
  resultsConnectionDetails,
  analysisSpecifications,
  resultsDataModelSettings
)

Arguments

resultsConnectionDetails

The connection details to the results database which is an object of class connectionDetails as created by the DatabaseConnector::createConnectionDetails() function.

resultsConnectionDetails

The connection details to the results database which is an object of class connectionDetails as created by the DatabaseConnector::createConnectionDetails() function.

analysisSpecifications

An object of type AnalysisSpecifications as created by createEmptyAnalysisSpecificiations().

analysisSpecifications

An object of type AnalysisSpecifications as created by createEmptyAnalysisSpecificiations().

resultsDataModelSettings

The results data model settings as created using [@seealso createResultsDataModelSettings()]

Method createModuleSpecifications()

Creates the PatientLevelPredictionValidation Module Specifications

Usage

PatientLevelPredictionValidationModule$createModuleSpecifications(
  validationList = list(PatientLevelPrediction::createValidationDesign(plpModelList =
    list(file.path("location_to_model")), targetId = 1, outcomeId = 3,
    restrictPlpDataSettings = PatientLevelPrediction::createRestrictPlpDataSettings(),
    populationSettings = NULL, recalibrate = "weakRecalibration", runCovariateSummary =
    TRUE), PatientLevelPrediction::createValidationDesign(plpModelList =
    list(file.path("location_to_model")), targetId = 4, outcomeId = 3,
    restrictPlpDataSettings = PatientLevelPrediction::createRestrictPlpDataSettings(),

       populationSettings = NULL, recalibrate = "weakRecalibration", runCovariateSummary
    = TRUE)),
  logLevel = "INFO"
)

Arguments

validationList

A list of validation designs from PatientLevelPrediction::createValidationDesign

logLevel

The logging level while executing the model validation.

Method validateModuleSpecifications()

Validate the module specifications

Usage

PatientLevelPredictionValidationModule$validateModuleSpecifications(
  moduleSpecifications
)

Arguments

moduleSpecifications

The PatientLevelPredictionValidation module specifications

Method clone()

The objects of this class are cloneable with this method.

Usage

PatientLevelPredictionValidationModule$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.