AMR/R/first_isolates.R

605 lines
23 KiB
R
Executable File

# ==================================================================== #
# TITLE #
# Antimicrobial Resistance (AMR) Analysis #
# #
# AUTHORS #
# Berends MS (m.s.berends@umcg.nl), Luz CF (c.f.luz@umcg.nl) #
# #
# LICENCE #
# This program is free software; you can redistribute it and/or modify #
# it under the terms of the GNU General Public License version 2.0, #
# as published by the Free Software Foundation. #
# #
# This program is distributed in the hope that it will be useful, #
# but WITHOUT ANY WARRANTY; without even the implied warranty of #
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the #
# GNU General Public License for more details. #
# ==================================================================== #
#' Determine first (weighted) isolates
#'
#' Determine first (weighted) isolates of all microorganisms of every patient per episode and (if needed) per specimen type.
#' @param tbl a \code{data.frame} containing isolates.
#' @param col_date column name of the result date (or date that is was received on the lab)
#' @param col_patient_id column name of the unique IDs of the patients
#' @param col_bactid column name of the unique IDs of the microorganisms (should occur in the \code{\link{microorganisms}} dataset). Get your bactid's with the function \code{\link{guess_bactid}}, that takes microorganism names as input.
#' @param col_testcode column name of the test codes. Use \code{col_testcode = NA} to \strong{not} exclude certain test codes (like test codes for screening). In that case \code{testcodes_exclude} will be ignored. Supports tidyverse-like quotation.
#' @param col_specimen column name of the specimen type or group
#' @param col_icu column name of the logicals (\code{TRUE}/\code{FALSE}) whether a ward or department is an Intensive Care Unit (ICU)
#' @param col_keyantibiotics column name of the key antibiotics to determine first \emph{weighted} isolates, see \code{\link{key_antibiotics}}. Supports tidyverse-like quotation.
#' @param episode_days episode in days after which a genus/species combination will be determined as 'first isolate' again
#' @param testcodes_exclude character vector with test codes that should be excluded (case-insensitive)
#' @param icu_exclude logical whether ICU isolates should be excluded
#' @param filter_specimen specimen group or type that should be excluded
#' @param output_logical return output as \code{logical} (will else be the values \code{0} or \code{1})
#' @param type type to determine weighed isolates; can be \code{"keyantibiotics"} or \code{"points"}, see Details
#' @param ignore_I logical to determine whether antibiotic interpretations with \code{"I"} will be ignored when \code{type = "keyantibiotics"}, see Details
#' @param points_threshold points until the comparison of key antibiotics will lead to inclusion of an isolate when \code{type = "points"}, see Details
#' @param info print progress
#' @param col_genus (deprecated, use \code{col_bactid} instead) column name of the genus of the microorganisms
#' @param col_species (deprecated, use \code{col_bactid} instead) column name of the species of the microorganisms
#' @details \strong{WHY THIS IS SO IMPORTANT} \cr
#' To conduct an analysis of antimicrobial resistance, you should only include the first isolate of every patient per episode \href{https://www.ncbi.nlm.nih.gov/pubmed/17304462}{[1]}. If you would not do this, you could easily get an overestimate or underestimate of the resistance of an antibiotic. Imagine that a patient was admitted with an MRSA and that it was found in 5 different blood cultures the following week. The resistance percentage of oxacillin of all \emph{S. aureus} isolates would be overestimated, because you included this MRSA more than once. It would be \href{https://en.wikipedia.org/wiki/Selection_bias}{selection bias}.
#'
#' \strong{DETERMINING WEIGHTED ISOLATES} \cr
#' \strong{1. Using} \code{type = "keyantibiotics"} \strong{and parameter} \code{ignore_I} \cr
#' To determine weighted isolates, the difference between key antibiotics will be checked. Any difference from S to R (or vice versa) will (re)select an isolate as a first weighted isolate. With \code{ignore_I = FALSE}, also differences from I to S|R (or vice versa) will lead to this. This is a reliable method and 30-35 times faster than method 2. \cr
#' \strong{2. Using} \code{type = "points"} \strong{and parameter} \code{points_threshold} \cr
#' To determine weighted isolates, difference between antimicrobial interpretations will be measured with points. A difference from I to S|R (or vice versa) means 0.5 points, a difference from S to R (or vice versa) means 1 point. When the sum of points exceeds \code{points_threshold}, an isolate will be (re)selected as a first weighted isolate. This method is being used by the Infection Prevention department (Dr M. Lokate) of the University Medical Center Groningen (UMCG).
#' @keywords isolate isolates first
#' @export
#' @importFrom dplyr arrange_at lag between row_number filter mutate arrange
#' @return A vector to add to table, see Examples.
#' @source Methodology of this function is based on: "M39 Analysis and Presentation of Cumulative Antimicrobial Susceptibility Test Data, 4th Edition", 2014, Clinical and Laboratory Standards Institute. \url{https://clsi.org/standards/products/microbiology/documents/m39/}.
#' @examples
#' # septic_patients is a dataset available in the AMR package
#' ?septic_patients
#' my_patients <- septic_patients
#'
#' library(dplyr)
#' my_patients$first_isolate <- my_patients %>%
#' first_isolate(col_date = "date",
#' col_patient_id = "patient_id",
#' col_bactid = "bactid")
#'
#' \dontrun{
#'
#' # set key antibiotics to a new variable
#' tbl$keyab <- key_antibiotics(tbl)
#'
#' tbl$first_isolate <-
#' first_isolate(tbl)
#'
#' tbl$first_isolate_weighed <-
#' first_isolate(tbl,
#' col_keyantibiotics = 'keyab')
#'
#' tbl$first_blood_isolate <-
#' first_isolate(tbl,
#' filter_specimen = 'Blood')
#'
#' tbl$first_blood_isolate_weighed <-
#' first_isolate(tbl,
#' filter_specimen = 'Blood',
#' col_keyantibiotics = 'keyab')
#'
#' tbl$first_urine_isolate <-
#' first_isolate(tbl,
#' filter_specimen = 'Urine')
#'
#' tbl$first_urine_isolate_weighed <-
#' first_isolate(tbl,
#' filter_specimen = 'Urine',
#' col_keyantibiotics = 'keyab')
#'
#' tbl$first_resp_isolate <-
#' first_isolate(tbl,
#' filter_specimen = 'Respiratory')
#'
#' tbl$first_resp_isolate_weighed <-
#' first_isolate(tbl,
#' filter_specimen = 'Respiratory',
#' col_keyantibiotics = 'keyab')
#' }
first_isolate <- function(tbl,
col_date,
col_patient_id,
col_bactid = NA,
col_testcode = NA,
col_specimen = NA,
col_icu = NA,
col_keyantibiotics = NA,
episode_days = 365,
testcodes_exclude = '',
icu_exclude = FALSE,
filter_specimen = NA,
output_logical = TRUE,
type = "keyantibiotics",
ignore_I = TRUE,
points_threshold = 2,
info = TRUE,
col_genus = NA,
col_species = NA) {
# bactid OR genus+species must be available
if (is.na(col_bactid) & (is.na(col_genus) | is.na(col_species))) {
stop('`col_bactid or both `col_genus` and `col_species` must be available.')
}
# check if columns exist
check_columns_existance <- function(column, tblname = tbl) {
if (NROW(tblname) <= 1 | NCOL(tblname) <= 1) {
stop('Please check tbl for existance.')
}
if (!is.na(column)) {
if (!(column %in% colnames(tblname))) {
stop('Column `', column, '` not found.')
}
}
}
check_columns_existance(col_date)
check_columns_existance(col_patient_id)
check_columns_existance(col_bactid)
check_columns_existance(col_genus)
check_columns_existance(col_species)
check_columns_existance(col_testcode)
check_columns_existance(col_icu)
check_columns_existance(col_keyantibiotics)
if (!is.na(col_bactid)) {
tbl <- tbl %>% left_join_microorganisms(by = col_bactid)
col_genus <- "genus"
col_species <- "species"
}
if (is.na(col_testcode)) {
testcodes_exclude <- NA
}
# remove testcodes
if (!is.na(testcodes_exclude[1]) & testcodes_exclude[1] != '' & info == TRUE) {
cat('Isolates from these test codes will be ignored:\n', toString(testcodes_exclude), '\n')
}
if (is.na(col_icu)) {
icu_exclude <- FALSE
} else {
tbl <- tbl %>%
mutate(col_icu = tbl %>% pull(col_icu) %>% as.logical())
}
if (is.na(col_specimen)) {
filter_specimen <- ''
}
weighted.notice <- ''
# filter on specimen group and keyantibiotics when they are filled in
if (!is.na(filter_specimen) & filter_specimen != '') {
check_columns_existance(col_specimen, tbl)
if (info == TRUE) {
cat('Isolates other than of specimen group \'', filter_specimen, '\' will be ignored. ', sep = '')
}
} else {
filter_specimen <- ''
}
if (col_keyantibiotics %in% c(NA, '')) {
col_keyantibiotics <- ''
} else {
tbl <- tbl %>% mutate(key_ab = tbl %>% pull(col_keyantibiotics))
}
if (is.na(testcodes_exclude[1])) {
testcodes_exclude <- ''
}
# create new dataframe with original row index and right sorting
tbl <- tbl %>%
mutate(first_isolate_row_index = 1:nrow(tbl),
date_lab = tbl %>% pull(col_date),
patient_id = tbl %>% pull(col_patient_id),
species = tbl %>% pull(col_species),
genus = tbl %>% pull(col_genus)) %>%
mutate(species = if_else(is.na(species) | species == "(no MO)", "", species),
genus = if_else(is.na(genus) | genus == "(no MO)", "", genus))
if (filter_specimen == '') {
if (icu_exclude == FALSE) {
if (info == TRUE) {
cat('Isolates from ICU will *NOT* be ignored.\n')
}
tbl <- tbl %>%
arrange_at(c(col_patient_id,
col_genus,
col_species,
col_date))
row.start <- 1
row.end <- nrow(tbl)
} else {
if (info == TRUE) {
cat('Isolates from ICU will be ignored.\n')
}
tbl <- tbl %>%
arrange_at(c(col_icu,
col_patient_id,
col_genus,
col_species,
col_date))
suppressWarnings(
row.start <- which(tbl %>% pull(col_icu) == FALSE) %>% min(na.rm = TRUE)
)
suppressWarnings(
row.end <- which(tbl %>% pull(col_icu) == FALSE) %>% max(na.rm = TRUE)
)
}
} else {
# sort on specimen and only analyse these row to save time
if (icu_exclude == FALSE) {
if (info == TRUE) {
cat('Isolates from ICU will *NOT* be ignored.\n')
}
tbl <- tbl %>%
arrange_at(c(col_specimen,
col_patient_id,
col_genus,
col_species,
col_date))
suppressWarnings(
row.start <- which(tbl %>% pull(col_specimen) == filter_specimen) %>% min(na.rm = TRUE)
)
suppressWarnings(
row.end <- which(tbl %>% pull(col_specimen) == filter_specimen) %>% max(na.rm = TRUE)
)
} else {
if (info == TRUE) {
cat('Isolates from ICU will be ignored.\n')
}
tbl <- tbl %>%
arrange_at(c(col_icu,
col_specimen,
col_patient_id,
col_genus,
col_species,
col_date))
suppressWarnings(
row.start <- which(tbl %>% pull(col_specimen) == filter_specimen
& tbl %>% pull(col_icu) == FALSE) %>% min(na.rm = TRUE)
)
suppressWarnings(
row.end <- which(tbl %>% pull(col_specimen) == filter_specimen
& tbl %>% pull(col_icu) == FALSE) %>% max(na.rm = TRUE)
)
}
}
if (abs(row.start) == Inf | abs(row.end) == Inf) {
if (info == TRUE) {
cat('No isolates found.\n')
}
# NA's where genus is unavailable
tbl <- tbl %>%
mutate(real_first_isolate = if_else(genus == '', NA, FALSE))
if (output_logical == FALSE) {
tbl$real_first_isolate <- tbl %>% pull(real_first_isolate) %>% as.integer()
}
return(tbl %>% pull(real_first_isolate))
}
# suppress warnings because dplyr want us to use library(dplyr) when using filter(row_number())
suppressWarnings(
scope.size <- tbl %>%
filter(
row_number() %>% between(row.start,
row.end),
genus != '') %>%
nrow()
)
# Analysis of first isolate ----
all_first <- tbl %>%
mutate(other_pat_or_mo = if_else(patient_id == lag(patient_id)
& genus == lag(genus)
& species == lag(species),
FALSE,
TRUE),
days_diff = 0) %>%
mutate(days_diff = if_else(other_pat_or_mo == FALSE,
(date_lab - lag(date_lab)) + lag(days_diff),
0))
if (col_keyantibiotics != '') {
if (info == TRUE) {
if (type == 'keyantibiotics') {
cat('Comparing key antibiotics for first weighted isolates (')
if (ignore_I == FALSE) {
cat('NOT ')
}
cat('ignoring I)...\n')
}
if (type == 'points') {
cat(paste0('Comparing antibiotics for first weighted isolates (using points threshold of '
, points_threshold, ')...\n'))
}
}
type_param <- type
# suppress warnings because dplyr want us to use library(dplyr) when using filter(row_number())
suppressWarnings(
all_first <- all_first %>%
mutate(key_ab_lag = lag(key_ab)) %>%
mutate(key_ab_other = !key_antibiotics_equal(x = key_ab,
y = key_ab_lag,
type = type_param,
ignore_I = ignore_I,
points_threshold = points_threshold,
info = info)) %>%
mutate(
real_first_isolate =
if_else(
between(row_number(), row.start, row.end)
& genus != ''
& (other_pat_or_mo
| days_diff >= episode_days
| key_ab_other),
TRUE,
FALSE))
)
if (info == TRUE) {
cat('\n')
}
} else {
# suppress warnings because dplyr want us to use library(dplyr) when using filter(row_number())
suppressWarnings(
all_first <- all_first %>%
mutate(
real_first_isolate =
if_else(
between(row_number(), row.start, row.end)
& genus != ''
& (other_pat_or_mo
| days_diff >= episode_days),
TRUE,
FALSE))
)
}
# first one as TRUE
all_first[row.start, 'real_first_isolate'] <- TRUE
# no tests that should be included, or ICU
if (!is.na(col_testcode)) {
all_first[which(all_first[, col_testcode] %in% tolower(testcodes_exclude)), 'real_first_isolate'] <- FALSE
}
if (icu_exclude == TRUE) {
all_first[which(all_first[, col_icu] == TRUE), 'real_first_isolate'] <- FALSE
}
# NA's where genus is unavailable
all_first <- all_first %>%
mutate(real_first_isolate = if_else(genus %in% c('', '(no MO)', NA), NA, real_first_isolate))
all_first <- all_first %>%
arrange(first_isolate_row_index) %>%
pull(real_first_isolate)
if (info == TRUE) {
cat(paste0('\nFound ',
all_first %>% sum(na.rm = TRUE),
' first ', weighted.notice, 'isolates (',
(all_first %>% sum(na.rm = TRUE) / scope.size) %>% percent(),
' of isolates in scope [where genus was not empty] and ',
(all_first %>% sum(na.rm = TRUE) / tbl %>% nrow()) %>% percent(),
' of total)\n'))
}
if (output_logical == FALSE) {
all_first <- all_first %>% as.integer()
}
all_first
}
#' Key antibiotics based on bacteria ID
#'
#' @param tbl table with antibiotics coloms, like \code{amox} and \code{amcl}.
#' @inheritParams first_isolate
#' @param amcl,amox,cfot,cfta,cftr,cfur,cipr,clar,clin,clox,doxy,gent,line,mero,peni,pita,rifa,teic,trsu,vanc column names of antibiotics, case-insensitive
#' @export
#' @importFrom dplyr %>% mutate if_else
#' @return Character of length 1.
#' @seealso \code{\link{mo_property}} \code{\link{antibiotics}}
#' @examples
#' \donttest{
#' #' # set key antibiotics to a new variable
#' tbl$keyab <- key_antibiotics(tbl)
#' }
key_antibiotics <- function(tbl,
col_bactid = 'bactid',
info = TRUE,
amcl = 'amcl',
amox = 'amox',
cfot = 'cfot',
cfta = 'cfta',
cftr = 'cftr',
cfur = 'cfur',
cipr = 'cipr',
clar = 'clar',
clin = 'clin',
clox = 'clox',
doxy = 'doxy',
gent = 'gent',
line = 'line',
mero = 'mero',
peni = 'peni',
pita = 'pita',
rifa = 'rifa',
teic = 'teic',
trsu = 'trsu',
vanc = 'vanc') {
keylist <- character(length = nrow(tbl))
if (!col_bactid %in% colnames(tbl)) {
stop('Column ', col_bactid, ' not found.', call. = FALSE)
}
# check columns
col.list <- c(amox, cfot, cfta, cftr, cfur, cipr, clar,
clin, clox, doxy, gent, line, mero, peni,
pita, rifa, teic, trsu, vanc)
col.list <- check_available_columns(tbl = tbl, col.list = col.list, info = info)
amox <- col.list[amox]
cfot <- col.list[cfot]
cfta <- col.list[cfta]
cftr <- col.list[cftr]
cfur <- col.list[cfur]
cipr <- col.list[cipr]
clar <- col.list[clar]
clin <- col.list[clin]
clox <- col.list[clox]
doxy <- col.list[doxy]
gent <- col.list[gent]
line <- col.list[line]
mero <- col.list[mero]
peni <- col.list[peni]
pita <- col.list[pita]
rifa <- col.list[rifa]
teic <- col.list[teic]
trsu <- col.list[trsu]
vanc <- col.list[vanc]
# join microorganisms
tbl <- tbl %>% left_join_microorganisms(col_bactid)
tbl$key_ab <- NA_character_
# Staphylococcus
list_ab <- c(clox, trsu, teic, vanc, doxy, line, clar, rifa)
list_ab <- list_ab[list_ab %in% colnames(tbl)]
tbl <- tbl %>% mutate(key_ab =
if_else(genus == 'Staphylococcus',
apply(X = tbl[, list_ab],
MARGIN = 1,
FUN = function(x) paste(x, collapse = "")),
key_ab))
# Rest of Gram +
list_ab <- c(peni, amox, teic, vanc, clin, line, clar, trsu)
list_ab <- list_ab[list_ab %in% colnames(tbl)]
tbl <- tbl %>% mutate(key_ab =
if_else(gramstain %like% '^Positive ',
apply(X = tbl[, list_ab],
MARGIN = 1,
FUN = function(x) paste(x, collapse = "")),
key_ab))
# Gram -
list_ab <- c(amox, amcl, pita, cfur, cfot, cfta, cftr, mero, cipr, trsu, gent)
list_ab <- list_ab[list_ab %in% colnames(tbl)]
tbl <- tbl %>% mutate(key_ab =
if_else(gramstain %like% '^Negative ',
apply(X = tbl[, list_ab],
MARGIN = 1,
FUN = function(x) paste(x, collapse = "")),
key_ab))
# format
tbl <- tbl %>%
mutate(key_ab = gsub('(NA|NULL)', '-', key_ab) %>% toupper())
tbl$key_ab
}
#' @importFrom dplyr progress_estimated %>%
#' @noRd
key_antibiotics_equal <- function(x,
y,
type = c("keyantibiotics", "points"),
ignore_I = TRUE,
points_threshold = 2,
info = FALSE) {
# x is active row, y is lag
type <- type[1]
if (length(x) != length(y)) {
stop('Length of `x` and `y` must be equal.')
}
result <- logical(length(x))
if (info == TRUE) {
p <- dplyr::progress_estimated(length(x))
}
for (i in 1:length(x)) {
if (info == TRUE) {
p$tick()$print()
}
if (is.na(x[i])) {
x[i] <- ''
}
if (is.na(y[i])) {
y[i] <- ''
}
if (nchar(x[i]) != nchar(y[i])) {
result[i] <- FALSE
} else if (x[i] == '' & y[i] == '') {
result[i] <- TRUE
} else {
x2 <- strsplit(x[i], "")[[1]]
y2 <- strsplit(y[i], "")[[1]]
if (type == 'points') {
# count points for every single character:
# - no change is 0 points
# - I <-> S|R is 0.5 point
# - S|R <-> R|S is 1 point
# use the levels of as.rsi (S = 1, I = 2, R = 3)
suppressWarnings(x2 <- x2 %>% as.rsi() %>% as.double())
suppressWarnings(y2 <- y2 %>% as.rsi() %>% as.double())
points <- (x2 - y2) %>% abs() %>% sum(na.rm = TRUE)
result[i] <- ((points / 2) >= points_threshold)
} else if (type == 'keyantibiotics') {
# check if key antibiotics are exactly the same
# also possible to ignore I, so only S <-> R and S <-> R are counted
if (ignore_I == TRUE) {
valid_chars <- c('S', 's', 'R', 'r')
} else {
valid_chars <- c('S', 's', 'I', 'i', 'R', 'r')
}
# remove invalid values (like "-", NA) on both locations
x2[which(!x2 %in% valid_chars)] <- '?'
x2[which(!y2 %in% valid_chars)] <- '?'
y2[which(!x2 %in% valid_chars)] <- '?'
y2[which(!y2 %in% valid_chars)] <- '?'
result[i] <- all(x2 == y2)
} else {
stop('`', type, '` is not a valid value for type, must be "points" or "keyantibiotics". See ?first_isolate.')
}
}
}
if (info == TRUE) {
cat('\n')
}
result
}