In jianxiaoyang/RareDiseases:

library(PatientLevelPrediction)
library(dplyr)
library(ggplot2)
library(DatabaseConnector)

Diseases <- c("HSAN","Craniopharyngioma","Brain neoplasm","Kabuki syndrome","GNB-5 mutation")
cohortsId <- c("4445","4450","4452","4453","4454")

studyFolder <- "/Users/yang/Desktop/RareDisease"

connectionDetails <- createConnectionDetails(dbms = "postgresql", 
                                             server = "localhost/ohdsi", 
                                             user = "joe", 
                                             password = "supersecret")

cdmDatabaseSchema <- "my_cdm_data"
cohortsDatabaseSchema <- "my_results"
cdmVersion <- "5"

createAnalysesDetails <- function(connectionDetails,
                                  cdmDatabaseSchema,
                                  studyFolder,
                                  cdmVersion,
                                  disease,
                                  cohortId){
    # Define which types of covariates must be constructed ----
    covariateSettings <- FeatureExtraction::covariateSettings <- createCovariateSettings(useCovariateDemographics = TRUE,
                                                                                         useCovariateDemographicsGender = TRUE,
                                                                                         useCovariateDemographicsRace = TRUE,
                                                                                         useCovariateDemographicsEthnicity = TRUE,
                                                                                         useCovariateDemographicsAge = TRUE,
                                                                                         useCovariateDemographicsYear = TRUE,
                                                                                         useCovariateDemographicsMonth = TRUE,
                                                                                         useCovariateConditionOccurrence = TRUE,
                                                                                         useCovariateConditionOccurrenceLongTerm = TRUE,
                                                                                         useCovariateConditionOccurrenceShortTerm = TRUE,
                                                                                         useCovariateConditionOccurrenceInptMediumTerm = TRUE,
                                                                                         useCovariateConditionEra = TRUE,
                                                                                         useCovariateConditionEraEver = TRUE,
                                                                                         useCovariateConditionEraOverlap = TRUE,
                                                                                         useCovariateConditionGroup = TRUE,
                                                                                         useCovariateConditionGroupMeddra = TRUE,
                                                                                         useCovariateConditionGroupSnomed = TRUE,
                                                                                         useCovariateDrugExposure = TRUE,
                                                                                         useCovariateDrugExposureLongTerm = TRUE,
                                                                                         useCovariateDrugExposureShortTerm = TRUE,
                                                                                         useCovariateDrugEra = TRUE,
                                                                                         useCovariateDrugEraLongTerm = TRUE,
                                                                                         useCovariateDrugEraShortTerm = TRUE,
                                                                                         useCovariateDrugEraOverlap = TRUE,
                                                                                         useCovariateDrugEraEver = TRUE,
                                                                                         useCovariateDrugGroup = TRUE,
                                                                                         useCovariateProcedureOccurrence = TRUE,
                                                                                         useCovariateProcedureOccurrenceLongTerm = TRUE,
                                                                                         useCovariateProcedureOccurrenceShortTerm = TRUE,
                                                                                         useCovariateProcedureGroup = TRUE,
                                                                                         useCovariateObservation = TRUE,
                                                                                         useCovariateObservationLongTerm = TRUE,
                                                                                         useCovariateObservationShortTerm = TRUE,
                                                                                         useCovariateObservationCountLongTerm = TRUE,
                                                                                         useCovariateMeasurement = TRUE,
                                                                                         useCovariateMeasurementLongTerm = TRUE,
                                                                                         useCovariateMeasurementShortTerm = TRUE,
                                                                                         useCovariateMeasurementCountLongTerm = TRUE,
                                                                                         useCovariateMeasurementBelow = TRUE,
                                                                                         useCovariateMeasurementAbove = TRUE,
                                                                                         useCovariateConceptCounts = TRUE,
                                                                                         useCovariateRiskScores = TRUE,
                                                                                         useCovariateRiskScoresCharlson = TRUE,
                                                                                         useCovariateRiskScoresDCSI = TRUE,
                                                                                         useCovariateRiskScoresCHADS2 = TRUE,
                                                                                         useCovariateRiskScoresCHADS2VASc = TRUE,
                                                                                         excludedCovariateConceptIds = c(),
                                                                                         deleteCovariatesSmallCount = 100)

    plpData <- PatientLevelPrediction::getPlpData(connectionDetails = connectionDetails,
                                                  cdmDatabaseSchema = cdmDatabaseSchema,
                                                  cohortDatabaseSchema = cdmDatabaseSchema,
                                                  cohortTable = "cohort",
                                                  cohortId = cohortId,
                                                  washoutPeriod = 183,
                                                  covariateSettings = covariateSettings,
                                                  outcomeDatabaseSchema = cdmDatabaseSchema,
                                                  outcomeTable = "cohort",
                                                  outcomeIds = cohortId,
                                                  cdmVersion = cdmVersion)

    setwd(file.path(paste0(studyFolder,"/",disease,"/plp_data")))
    PatientLevelPrediction::savePlpData(plpData,paste0(disease,"plp_data"))
}

createAnalysesDetails(connectionDetails,
                                  cdmDatabaseSchema,
                                  studyFolder,
                                  cdmVersion,
                                  disease,
                                  cohortId)

    setwd(file.path(paste0(studyFolder,"/",disease)))

    plpData <- PatientLevelPrediction::loadPlpData(file.path(paste0(studyFolder,"/",disease,"/plp_data")))

    covariatesinfo <- as.data.frame(plpData$covariateRef[,1:3]) # get the match infromation of covariateId, covariateName and analysisId
    nameid <- as.data.frame(covariatesinfo[,1:2])  # get the match infromation of covariateId and covariateName
    covariateswhole <- as.data.frame(plpData$covariates[,1:2]) # all covariates with rowId (individual)

    childrenage <- c(10:28)
    childrenrowid <- dplyr::filter(covariateswhole,covariateswhole$covariateId %in% childrenage)
    childrencovariate <- merge(childrenrowid,covariateswhole,by = "rowId",all.x = TRUE, sort = TRUE)
    names(childrencovariate) <- c("rowId","Age group","covariateId")

    mergecovariate <- merge(childrencovariate,covariatesinfo, by = "covariateId",all.x = TRUE, sort = TRUE) # all covariates with the infromation of covariates name (name and id), covariates type (analysisId), age group and rowId

Demographics

summaryDemo <- function(mergecovariate){

        ## Age group
        age_group <- dplyr::filter(mergecovariate,mergecovariate$analysisId==5)

        agegroup <- as.data.frame(table(age_group$covariateId))
        agegroup <- dplyr::arrange(agegroup,desc(Freq))
        agegroup$prevalence <- agegroup$Freq/length(childrenrowid[,1])
        names(agegroup) <- c("covariateId","Person Count","Prevalence")

        agegroup <- merge(agegroup,nameid, by = "covariateId",all.x = TRUE, sort = TRUE)
        agegroup <- dplyr::arrange(agegroup,desc(agegroup$`Person Count`))

        # age_group$covariateName <- factor(age_group$covariateName, levels=c("Age group: 5-9","Age group: 10-14","Age group: 15-19"))

        plotage <- ggplot2::ggplot(age_group,aes(x=age_group$covariateName))+geom_bar(width=0.5,fill=brewer.pal(length(agegroup),"Blues"))+labs(x="Age group",y="Person Count")+geom_text(aes(label=..count..),
                                                                                                                                                                                      stat="count", color = "black",
                                                                                                                                                                                      vjust = 0, size = 4)
        ggplot2::ggsave(filename="Age Group.png",plot=plotage,device="png",path = file.path(paste0(studyFolder,"/",disease)))
        # write.csv(age_group,file="age_group.csv")
        write.csv(agegroup,file="agegroup.csv")


        ## Gender

        Gender<- dplyr::filter(mergecovariate,mergecovariate$analysisId==2)

        gender <- as.data.frame(table(Gender$covariateId))
        g <- c("MALE","FEMALE")
        counts <- c(gender[1,2],length(childrenrowid[,1])-gender[1,2])
        gender <- data.frame(g,counts)
        names(gender) <- c("Gender","Patient Counts")

        # write.csv(Gender,file="Gender.csv")
        write.csv(gender,file="gender.csv")


        ## demo

        age_gender <- merge(age_group, Gender, by = "rowId",all.x = TRUE, sort = TRUE)
        new <- c("rowId","Age group.x","covariateName.x","covariateId.y","covariateName.y")
        age_gender <- age_gender[new]
        names(age_gender) <- c("rowId","age group.x","Age Group","covariateId.y","Gender")
        age_gender$covariateId.y[is.na(age_gender$covariateId.y)] <- 8532
        age_gender$Gender[is.na(age_gender$Gender)] <- "Gender = FEMALE"

        plotgender <- ggplot2::ggplot(age_gender,aes(x=age_gender$Gender))+geom_bar(width=0.4,fill=c("lightcoral","royalblue3"))+labs(x="Gender",y="Person Count")+geom_text(aes(label=..count..),stat="count", color = "black",
                                                                                                                                                                    vjust = 0, size = 4)
        ggplot2::ggsave(filename="Gender.png",plot=plotgender,device="png",path = file.path(paste0(studyFolder,"/",disease)))

        write.csv(age_gender,file="age_gender.csv")
    }

summaryDemo(mergecovariate)

Rank covariates

rankcov <- function(mergecovariate){

        ## Condition Occurrence

        condition_occurrence <- dplyr::filter(mergecovariate,mergecovariate$analysisId>100 & mergecovariate$analysisId<200)

        conditionoccur <- as.data.frame(table(condition_occurrence$covariateId))
        conditionoccur <- dplyr::arrange(conditionoccur,desc(Freq))
        conditionoccur$prevalence <- conditionoccur$Freq/length(childrenrowid[,1])
        names(conditionoccur) <- c("covariateId","Person Count","Prevalence")

        conditionoccur <- merge(conditionoccur,nameid, by = "covariateId",all.x = TRUE, sort = TRUE)
        conditionoccur <- dplyr::arrange(conditionoccur,desc(conditionoccur$`Person Count`))

        # write.csv(condition_occurrence,file="condition_occurence.csv")
        write.csv(conditionoccur,file="conditionoccur.csv")


        ## Condition Era

        condition_era <- dplyr::filter(mergecovariate,mergecovariate$analysisId>200 & mergecovariate$analysisId<300)

        conditionera <- as.data.frame(table(condition_era$covariateId))
        conditionera <- dplyr::arrange(conditionera,desc(Freq))
        conditionera$prevalence <- conditionera$Freq/length(childrenrowid[,1])
        names(conditionera) <- c("covariateId","Person Count","Prevalence")

        conditionera <- merge(conditionera,nameid, by = "covariateId",all.x = TRUE, sort = TRUE)
        conditionera <- dplyr::arrange(conditionera,desc(conditionera$`Person Count`))

        # write.csv(condition_era,file="condition_era.csv")
        write.csv(conditionera,file="conditionera.csv")


        ## Drug Exposure

        drug_exposure <- dplyr::filter(mergecovariate,mergecovariate$analysisId>400 & mergecovariate$analysisId<500)

        drugexpo <- as.data.frame(table(drug_exposure$covariateId))
        drugexpo <- dplyr::arrange(drugexpo,desc(Freq))
        drugexpo$prevalence <- drugexpo$Freq/length(childrenrowid[,1])
        names(drugexpo) <- c("covariateId","Person Count","Prevalence")

        drugexpo <- merge(drugexpo,nameid, by = "covariateId",all.x = TRUE, sort = TRUE)
        drugexpo <- dplyr::arrange(drugexpo,desc(drugexpo$`Person Count`))

        # write.csv(drug_exposure,file="drug_exposure.csv")
        write.csv(drugexpo,file="drugexpo.csv")


        ## Drug Era

        drug_era <- dplyr::filter(mergecovariate,mergecovariate$analysisId>500 & mergecovariate$analysisId<600)

        drugera <- as.data.frame(table(drug_era$covariateId))
        drugera <- dplyr::arrange(drugera,desc(Freq))
        drugera$prevalence <- drugera$Freq/length(childrenrowid[,1])
        names(drugera) <- c("covariateId","Person Count","Prevalence")

        drugera <- merge(drugera,nameid, by = "covariateId",all.x = TRUE, sort = TRUE)
        drugera <- dplyr::arrange(drugera,desc(drugera$`Person Count`))

        # write.csv(drug_era,file="drug_era.csv")
        write.csv(drugera,file="drugera.csv")


        ## number of drugs within each ATC3 groupings all time prior

        drug_num <- dplyr::filter(mergecovariate,mergecovariate$analysisId>600 & mergecovariate$analysisId<700)

        drugnum <- as.data.frame(table(drug_num$covariateId))
        drugnum <- dplyr::arrange(drugnum,desc(Freq))
        drugnum$prevalence <- drugnum$Freq/length(childrenrowid[,1])
        names(drugnum) <- c("covariateId","Person Count","Prevalence")

        drugnum <- merge(drugnum,nameid, by = "covariateId",all.x = TRUE, sort = TRUE)
        drugnum <- dplyr::arrange(drugnum,desc(drugnum$`Person Count`))

        # write.csv(drug_num,file="drug_num.csv")
        write.csv(drugnum,file="drugnum.csv")


        ## Procedure Occurrence

        procedure_occurrence <- dplyr::filter(mergecovariate,mergecovariate$analysisId>700 & mergecovariate$analysisId<800)

        procedureoccur <- as.data.frame(table(procedure_occurrence$covariateId))
        procedureoccur <- dplyr::arrange(procedureoccur,desc(Freq))
        procedureoccur$prevalence <- procedureoccur$Freq/length(childrenrowid[,1])
        names(procedureoccur) <- c("covariateId","Person Count","Prevalence")

        procedureoccur <- merge(procedureoccur,nameid, by = "covariateId",all.x = TRUE, sort = TRUE)
        procedureoccur <- dplyr::arrange(procedureoccur,desc(procedureoccur$`Person Count`))

        # write.csv(procedure_occurrence,file="procedure_occurrence.csv")
        write.csv(procedureoccur,file="procedureoccur.csv")


        ## Observation

        Observation <- dplyr::filter(mergecovariate,mergecovariate$analysisId>900 & mergecovariate$analysisId<1000)

        Ob <- as.data.frame(table(Observation$covariateId))
        Ob <- dplyr::arrange(Ob,desc(Freq))
        Ob$prevalence <- Ob$Freq/length(childrenrowid[,1])
        names(Ob) <- c("covariateId","Person Count","Prevalence")

        Ob <- merge(Ob,nameid, by = "covariateId",all.x = TRUE, sort = TRUE)
        Ob <- dplyr::arrange(Ob,desc(Ob$`Person Count`))

        # write.csv(Observation,file="Observation.csv")
        write.csv(Ob,file="Ob.csv")

    }

rankcov(mergecovariate)

Tree map (unfinished!)

library(readr)
library(treemap)
library(d3treeR)
library(googleVis)
# devtools::install_github("timelyportfolio/d3treeR")

class <- c("SOC","ATC1")

treemap <- function(disease, covariatetype, classification){

    covariatetype <-read_csv(file.path(paste0(studyFolder,"/",disease,"/",covariatetype,".csv")), col_types = cols(`Person Count` = col_double()))
    paste0(covariatetype,"200") <- dataset[1:200,]
    names(conditionoccur200) <- c("X1","covariateId", "Person Count","Prevalence","SOC","covariateName")
    write.csv(conditionoccur200, file = "conditionoccur200.csv")

    treemap::treemap(paste0(covariatetype,"200"), 
                            index=c(classification,"covariateName"),  
                            vSize = "Person Count",  
                            vColor= "Prevalence", 
                            type="value", 
                            palette = "-RdGy", 
                            range=c(0,1),
                            bg.labels = 100,
                            title=paste0(disease,"plp_data"), 
                            fontsize.title = 14,
                            fontface.labels = 1,
                            fontfamily.title = "serif",
                            fontfamily.labels = "serif",
                            fontfamily.legend = "serif",
                            overlap.labels = 1,
                            border.col="white",
                            lowerbound.cex.labels=0,
                            aspRatio = 16/9)

d3tree(conditiontreemap,rootname="HSAN Condition Occurrence")
}