whichPatientDrug: Match patients against a look-up table of treatments

In n8thangreen/HESmanip: Hospital length of stay data preprocessing and analysis

Description

Determine which patients are susceptible or resistant to the guideline drugs, accounting for age category and other user defined covariates.

Usage

whichPatientDrug(inf.data, drugGuideline.data, patientdrugs)

Arguments

inf.data	Linked HES-LabBase2 data set of exposed patients.
drugGuideline.data	A data.frame of which antibiotics are first-line treatments for which pathogen.
patientdrugs	List of drugs per patient.

Value

returns: A list of susceptible or resistant status per patient:

Per drug	TRUE entry if guideline antibiotic and resistant/susceptible
Aggregated across all drugs	TRUE entry if guideline antibiotic and resistant/susceptible

Note

This is an alternative to the whichPatientDrug.join function which uses joining operations.

Author(s)

N. Green

See Also

AgeCatModify, whichPatientDrug.joins

Examples

data(infdata_sample)
data(antibioticLookup)

drug.names <<- factor(c("AMPAMOX","CTX","CTZ","CTRI","GENT","IMI","MERO",
                        "MET","PIPTAZ","CoNSmet","AMINOAMOX1","AMINOMET1","AMINOVAN1","CTXI1",
                        "pen","met","van2","AMINOAMOXAMP1","AMINOPIPTAZ1","AMINOCTZ1","AMINOIMI1","AMINOMER1","CTXI2"))

##################
# 0: susceptible #
# 1: resistant   #
# NA: not tested #
##################

## extract drug columns from original inf.data dataset
inf.drug.data <- infdata_sample[, paste("lab_",drug.names,sep="")]
inf.drug.data <- data.frame(lapply(inf.drug.data, factor))

## make a list of the drugs administered to each patient
## for each response type

patientdrugs.res <- patientdrugs.sus <- list()
for (i in 1:nrow(inf.drug.data)){ # each patient
  patientdrugs.res[[i]] <- drug.names[!is.na(inf.drug.data[i,]) & inf.drug.data[i,]==1]  # resistant 
  patientdrugs.sus[[i]] <- drug.names[!is.na(inf.drug.data[i,]) & inf.drug.data[i,]==0]  # susceptible
}

whichPatientDrug.res <- whichPatientDrug(infdata_sample, antibioticLookup.dat, patientdrugs.res)
whichPatientDrug.sus <- whichPatientDrug(infdata_sample, antibioticLookup.dat, patientdrugs.sus)





## The function is currently defined as
function (inf.data, drugGuideline.data, patientdrugs) 
{
    npatient <- length(patientdrugs)
    ndrugs <- length(drug.names)
    nguidelines <- nrow(drugGuideline.data)
    FollowGuideline.tot <- rep(FALSE, npatient)
    FollowGuideline.indiv <- matrix(NA, nrow = npatient, ncol = ndrugs)
    colnames(FollowGuideline.indiv) <- drug.names
    patient.age_cat <- AgeCatModify(inf.data)
    for (i in 1:nguidelines) {
        if (drugGuideline.data$drug_code[i] != "" & drugGuideline.data$drug_code[i] %in% 
            drug.names) {
            drugMatch <- sapply(patientdrugs, function(x) drugGuideline.data$drug_code[i] %in% 
                x) & (drugGuideline.data$ageCat[i] == patient.age_cat | 
                drugGuideline.data$ageCat[i] == "") & (toupper(drugGuideline.data$suspected_organism[i]) == 
                inf.data$lab_organismname | toupper(drugGuideline.data$suspected_organism[i]) == 
                inf.data$lab_group | toupper(drugGuideline.data$suspected_organism[i]) == 
                "")
            drugCol <- as.character(drugGuideline.data$drug_code[i])
            FollowGuideline.indiv[, drugCol] <- FollowGuideline.indiv[, 
                drugCol] | drugMatch
            FollowGuideline.tot <- FollowGuideline.tot | drugMatch
        }
    }
    list(indiv = FollowGuideline.indiv, all = FollowGuideline.tot)
  }

n8thangreen/HESmanip documentation built on March 21, 2020, 12:20 a.m.