Request for Help: Analyzing Antibiotic Coverage in Clinical Data

[ghost]

Hello everyone,

I am working on a project that involves analyzing the therapeutic coverage of empirically prescribed antibiotics based on susceptibility results from an antibiogram. I have a dataset containing patient information, prescribed antibiotics, and the corresponding antibiogram results.

Objective
The goal is to determine if the empirically prescribed antibiotics are effective against the organisms identified in the antibiogram. Specifically, I want to perform a coverage analysis that returns a binary output:

"therapeutic coverage" if the empirically prescribed antibiotic is found effective (the organism is susceptible).
"therapeutic failure" if the empirically prescribed antibiotic is not effective (the organism is resistant).

Dataset
The dataset contains the following columns:

participant_id: Unique identifier for each participant.
gender: Gender of the participant.
specimen: Type of specimen collected.
origin: Origin of the specimen.
emp_thp: Indicates if empiric therapy was given.
t1 to t5: Empirically prescribed antibiotics.
infection: Type of infection.
organisms: Organism identified.
Antibiotic sensitivity columns (pen, fox, amp, etc.): Susceptibility results for various antibiotics.
status_update: Status update for the participant.

Here is a bit of what i did, it would be very helpful if i could get help

Load necessary libraries

library(dplyr)
library(tidyr)
library(AMR)

Create the dataset

data <- tibble::tribble(
~participant_id, ~gender, ~specimen, ~origin, ~emp_thp, ~t1, ~t2, ~t3, ~t4, ~t5, ~infection, ~organisms, ~pen, ~fox, ~amp, ~gen, ~cip, ~cli, ~ery, ~tcy, ~sxt, ~fac, ~lzd, ~status_update,
"1811STPH3522", "Female", "tissue samples", "Surgery", "Yes", "Amo_Clav", "Metronidazole", NA, NA, NA, "SSTI", "B_STPHY_AURS", "R", "S", "R", NA, NA, NA, "S", "R", "S", NA, NA, "MSSA",
"1812STPH3538", "Female", "throat", "Community", "Yes", "Amo_Clav", NA, NA, NA, NA, "Others", "B_STPHY_AURS", "R", "S", "R", "S", "S", "S", "S", "S", "S", "S", "S", "MSSA",
"1812STPH3539", "Female", "tissue samples", "Surgery", "Yes", "Amo_Clav", "Ciprofloxacin", "Metronidazole", NA, NA, "SSTI", "B_STPHY_AURS", "R", "R", "R", "S", "R", "S", "S", "S", "S", "R", "S", "MRSA",
"1812STPH3541", "Female", "blood", "Paediatric", "Yes", "Ampicillin", "Gentamicin", NA, NA, NA, "Bacteremia", "B_STPHY_AURS", "R", "S", "R", "S", "S", "S", "S", "R", "R", "S", "S", "MSSA",
"1901STPH3578", "Female", "tissue samples", "Community", "Yes", "Flucloxacillin", NA, NA, NA, NA, "SSTI", "B_STPHY_AURS", "R", NA, "R", NA, NA, NA, "R", NA, "S", NA, NA, "MSSA",
"1902STPH3582", "Male", "swabs", "Community", "Yes", "Amo_Clav", NA, NA, NA, NA, "Others", "B_STPHY_AURS", NA, "S", "S", "S", "R", NA, NA, "R", "S", "S", "S", "MSSA",
"1902STPH3586", "Male", "tissue samples", "Community", "Yes", "Oxacillin", "Metronidazole", NA, NA, NA, "SSTI", "B_STPHY_AURS", "R", "S", "R", NA, NA, NA, "S", "R", "S", NA, NA, "MSSA",
"1902STPH3587", "Male", "tissue samples", "Paediatric", "Yes", "Cloxacillin", NA, NA, NA, NA, "SSTI", "B_STPHY_AURS", "R", "S", "R", NA, NA, NA, "R", "S", "S", NA, NA, "MSSA",
"1902STPH3616", "Female", "tissue samples", "Community", "Yes", "Ampicillin", "Gentamicin", NA, NA, NA, "SSTI", "B_STPHY_AURS", "R", "S", "R", "S", "R", "S", "S", "S", "S", "S", "S", "MSSA",
"1903STPH3637", "Male", "tissue samples", "Community", "Yes", "Ceftriaxone", NA, NA, NA, NA, "SSTI", "B_STPHY_AURS", "R", "S", "R", "S", "R", "S", "S", "R", "S", "S", "S", "MSSA",
"1903STPH3648", "Female", "swabs", "Community", "Yes", "Ampicillin", "Ceftriaxone", "Gentamycin", NA, NA, "Others", "B_STPHY_AURS", "R", "S", "R", "S", "R", "S", "S", "R", "S", "S", "S", "MSSA",
"1904STPH3661", "Male", "tissue samples", "Community", "Yes", "Amo_Clav", NA, NA, NA, NA, "SSTI", "B_STPHY_AURS", "R", "S", NA, NA, NA, NA, "S", "R", "S", NA, NA, "MSSA",
"1904STPH3663", "Male", "tissue samples", "Community", "Yes", "Ciprofloxacin", "Azithomycin", "Gentamycin", "Amo_Clav", NA, "SSTI", "B_STPHY_AURS", "R", "S", "R", "S", "R", NA, "S", "R", "S", "S", "S", "MSSA",
"1904STPH3677", "Male", "blood", "Paediatric", "Yes", "Ampicillin", "Gentamicin", NA, NA, NA, "Bacteremia", "B_STPHY_AURS", "R", "S", "R", "S", "R", "S", "S", "S", "S", "S", "S", "MSSA",
"1904STPH3681", "Male", "throat", "Community", "Yes", "Amo_Clav", "Ceftriaxone", "Cefixime", NA, NA, "Others", "B_STPHY_AURS", "R", "S", "R", "S", "R", "S", "S", "R", "S", "S", "S", "MSSA",
"1905STPH3704", "Male", "tissue samples", "Community", "Yes", "Metronidazole", NA, NA, NA, NA, "SSTI", "B_STPHY_AURS", "R", "S", "S", "S", "R", "S", "S", "R", "S", "S", "S", "MSSA",
"2002STPH3887", "Male", "fluid aspirates", "Internal", "Yes", "Amo_Clav", NA, NA, NA, NA, "SSTI", "B_STPHY_AURS", "R", "S", "R", NA, NA, NA, "S", "I", "S", NA, NA, "MSSA",
"2003STPH3901", "Female", "fluid aspirates", "Surgery", "Yes", "Amo_Clav", NA, NA, NA, NA, "SSTI", "B_STPHY_AURS", "R", "S", "R", "S", "R", "S", "S", "S", "S", "S", "S", "MSSA",
"2008STPH3964", "Male", "tissue samples", "Community", "Yes", "Flucloxacillin", NA, NA, NA, NA, "SSTI", "B_STPHY_AURS", "S", "S", "S", "S", "R", "S", "S", "S", "S", "R", "S", "MSSA",
"2012STPH4022", "Male", "tissue samples", "Community", "Yes", "Bacitracin", NA, NA, NA, NA, "SSTI", "B_STPHY_AURS", "R", "S", "R", "S", "R", "S", "S", "R", "S", "R", "S", "MSSA"
)

Define function to check therapeutic coverage

check_coverage <- function(emp_thp, antibiogram) {
if (emp_thp %in% antibiogram && antibiogram[emp_thp] == "S") {
return("therapeutic coverage")
} else {
return("therapeutic failure")
}
}

Reshape data to long format for antibiotic sensitivity columns

data_long <- data %>%
pivot_longer(cols = starts_with("pen"):starts_with("lzd"), names_to = "antibiotic", values_to = "sensitivity") %>%
filter(!is.na(sensitivity))

Perform coverage analysis

coverage_results <- data %>%
rowwise() %>%
mutate(coverage_result = check_coverage(emp_thp, c(pen = pen, fox = fox, amp = amp, gen = gen, cip = cip, cli = cli, ery = ery, tcy = tcy, sxt = sxt, fac = fac, lzd = lzd)))

View the results

print(coverage_results)

[msberends]

This is a lot of code to just get the empiric coverage of prescribed antibiotics.

You might want to read about analysing data in our manual: https://msberends.github.io/AMR/reference/index.html#analysing-data

More specifically, I think antibiogram() might help you to get a fast answer to your question (see here).

[ghost]

Thanks for your reply.

I also wanted to know if it was possible to get the 95%CI when getting the WISCA results just as in the article from

Barbieri E et al. (2021). Development of a Weighted-Incidence Syndromic Combination Antibiogram (WISCA) to guide the choice of the empiric antibiotic treatment for urinary tract infection in paediatric patients: a Bayesian approach Antimicrobial Resistance & Infection Control May 1;10(1):74; doi:10.1186/s13756-021-00939-2

[msberends]

Hi, great that you ask this. Not yet, but we are working on the implementation of it, see #133. We hope to have it implemented soon to test it!

[ghost]

Thanks for your reply