Create a codelist specifying the following (enter terms separated by a new line) :
You will see three tables:
Unmatched descendants will show any not included entries that, within the selected column, are the same or start with any of the included entries.
In hierarchical coding systems, such as ICD or Read, unmatched descendants can be found. For example, if code “E10” is included, all codes will be shown that start with “E10”, such as “E10.1”, “E10.2”, etc., that aren’t already included in the final codelist.
If multiple coding systems are present, e.g. SNOMED codes mapped to BNF chapters, entries with the same code or within the same chapter can be found. For example, if “Carvedilol” with the BNF chapter 20400080 is included, then “Eucardic” (a brand name for Carvedilol), which has the same BNF chapter, will be shown.
Cross-tabulation shows a contingency table on a specified column.
By default, terms are matched exactly and case-insensitively.
The underlying code takes the DATA and keeps only rows where in the specified COLUMN at least one of the SEARCHTERMS is found and none of the EXCLUSIONTERMS are present. This is implemented using R and dplyr, and can be used in a standalone R script if needed:
termsearch <- function(lookup, terms) {
stringr::str_detect(lookup, stringr::regex(paste(terms, collapse = '|'), ignore_case = TRUE))
}
initial <- dplyr::filter(DATA, termsearch(COLUMN, SEARCHTERMS))
excluded <- dplyr::filter(initial, termsearch(COLUMN, EXCLUSIONTERMS))
final <- dplyr::setdiff(initial, excluded)
If “Termset search method” is selected, search rules as devised by Williams, R. et al. (2019) are used (this is not optimised for speed here and therefore may be slow when searching large lookup datasets, and the implementation of the search strategy is more complex):
Williams, R. et al. (2019) ‘Term sets: A transparent and reproducible representation of clinical code sets’, PLOS ONE. Edited by I. Olier, 14(2), p. e0212291. Available at: https://doi.org/10.1371/journal.pone.0212291.
Press the download button to download a zip archive containing:
Compare two codelists and look for any matches. You can get a codelist directly from the codelist maker or upload an existing codelist (a variety of file formats are supported). Select which column should be matched on (usually it will be the column containing the clinical code, such as ICD, SNOMED, Readcode, etc…) and if an entry has a match in the other column it will be highlighted in green, else in red.
Check which are the most commonly used combinations of words in a codelist. You can get a codelist directly from the codelist maker or upload an existing codelist (a variety of file formats are supported). Select the column containing the terms (descriptions). A monogram is a single word, a bigram is two consecutive words, a trigram is a group of three consecutive words.