validateMultiplePlp: externally validate the multiple plp models across new...

In OHDSI/PatientLevelPrediction: Develop Clinical Prediction Models Using the Common Data Model

externally validate the multiple plp models across new datasets

Description

This function loads all the models in a multiple plp analysis folder and validates the models on new data

Usage

validateMultiplePlp(
  analysesLocation,
  validationDatabaseDetails,
  validationRestrictPlpDataSettings = createRestrictPlpDataSettings(),
  recalibrate = NULL,
  cohortDefinitions = NULL,
  saveDirectory = NULL
)

Arguments

analysesLocation	The location where the multiple plp analyses are
validationDatabaseDetails	A single or list of validation database settings created using createDatabaseDetails()
validationRestrictPlpDataSettings	The settings specifying the extra restriction settings when extracting the data created using createRestrictPlpDataSettings().
recalibrate	A vector of recalibration methods (currently supports 'RecalibrationintheLarge' and/or 'weakRecalibration')
cohortDefinitions	A list of cohortDefinitions
saveDirectory	The location to save to validation results

Details

Users need to input a location where the results of the multiple plp analyses are found and the connection and database settings for the new data

Value

Nothing. The results are saved to the saveDirectory

Examples

 
# first develop a model using runMultiplePlp
connectionDetails <- Eunomia::getEunomiaConnectionDetails()
Eunomia::createCohorts(connectionDetails = connectionDetails)
databaseDetails <- createDatabaseDetails(connectionDetails = connectionDetails,
                                         cdmDatabaseId = "1",
                                         cdmDatabaseName = "Eunomia",
                                         cdmDatabaseSchema = "main",
                                         targetId = 1,
                                         outcomeIds = 3)
covariateSettings <- 
 FeatureExtraction::createCovariateSettings(useDemographicsGender = TRUE,
   useDemographicsAge = TRUE, useConditionOccurrenceLongTerm = TRUE)
modelDesign <- createModelDesign(targetId = 1, 
                                 outcomeId = 3,
                                 modelSettings = setLassoLogisticRegression(seed = 42),
                                 covariateSettings = covariateSettings)
saveLoc <- file.path(tempdir(), "valiateMultiplePlp", "development")
results <- runMultiplePlp(databaseDetails = databaseDetails,
               modelDesignList = list(modelDesign),
               saveDirectory = saveLoc)
# now validate the model on a Eunomia but with a different target
analysesLocation <- saveLoc
validationDatabaseDetails <- createDatabaseDetails(connectionDetails = connectionDetails,
                                                   cdmDatabaseId = "2",
                                                   cdmDatabaseName = "EunomiaNew",
                                                   cdmDatabaseSchema = "main",
                                                   targetId = 4,
                                                   outcomeIds = 3)
newSaveLoc <- file.path(tempdir(), "valiateMultiplePlp", "validation")
validateMultiplePlp(analysesLocation = analysesLocation,
                    validationDatabaseDetails = validationDatabaseDetails,
                    saveDirectory = newSaveLoc)
# the results could now be viewed in the shiny app with viewMultiplePlp(newSaveLoc)

OHDSI/PatientLevelPrediction documentation built on Feb. 14, 2025, 9:44 a.m.