In OhdsiReportGenerator: Observational Health Data Sciences and Informatics Report Generator

Predictors

Candidate predictors The predictors are determined using data recorded relative to the target cohort index date. The settings endDays specifies the number of days relative to the target index that is the latest point in time for determining the covariates. For example, an endDays of -1 means that no data recorded on index or after are used by the covariates (the covariates only use data recorded up to index-1 day).

if(class(covariateSettings) == 'covariateSettings'){
  covariateSettings <- list(covariateSettings)
}

cat('\n There are ', length(covariateSettings), ' covariate settings. \n')

for(i in 1:length(covariateSettings)){

   cat('\n Covariate Setting', i, '\n')

  dataC <- data.frame(
    name = c(
      'function',
      names(lapply(covariateSettings[[i]], function(x) paste(x, collapse = '-', sep='-')))), 
    value = c(
      attr(covariateSettings[[i]],"fun"), 
      unlist(lapply(covariateSettings[[i]], function(x) paste(x, collapse = '-', sep='-'))
      ))
  )
  row.names(dataC) <- NULL

print(knitr::kable(x = dataC, caption = paste('covariate setting ', i)))
  cat('\n \n')
}

Feature Engineering

if(class(featureEngineeringSettings) == 'featureEngineeringSettings'){
  featureEngineeringSettings <- list(featureEngineeringSettings)
}

if(!is.null(attr(featureEngineeringSettings[[1]],"fun"))){
  if(attr(featureEngineeringSettings[[1]],"fun") != 'sameData'){

    cat('\n There are ', length(featureEngineeringSettings), ' feature engineering settings. \n')

    for(i in 1:length(featureEngineeringSettings)){

      cat('\n The function ', attr(featureEngineeringSettings[[i]],"fun"), ' with inputs: \n')

      feData <- data.frame(
        name = names(featureEngineeringSettings), 
        value = unlist(lapply(featureEngineeringSettings, function(x) paste(x, sep = '', collapse='-')))
        )

      print(knitr::kable(x = feData , caption = paste('feature engineering setting ', i)))
      cat('\n \n')

    }

  } else{
  cat('\n None \n')  
  }
} else{
  cat('\n None \n')  
}

Pre-processing

The following pre-processing were applied:

if(preprocessSettings$normalize){
  cat('\n - The data were normalized using the formula: value/maxValue . \n')
}

if(preprocessSettings$minFraction > 0){
  cat('\n - Candidate predictors occuring in less than ', preprocessSettings$minFraction*100,'\\% of patients in the target population were removed. \n')
}

if(preprocessSettings$removeRedundancy){
  cat('\n - Candidate predictors that were redundant (completely correlated with another predictor) were removed.  When two predictors are completely correlated, the most common predictor is removed.  For example, if 60\\% of the target population were male and 40\\% were female, then the male gender predictor would be removed since it is redundant and more common. \n')
}

OhdsiReportGenerator documentation built on April 12, 2025, 2:09 a.m.