David A. Springate 2014
This package provides an R interface for downloading clinical code lists and research objects from the repository
You can install the package from CRAN:
install.packages("rClinicalCodes")
library(rClinicalCodes)
You can install the development version from github using the devtools package:
install.packages("devtools")
require(devtools)
install_github("rClinicalCodes", "rOpenHealth")
require(rClinicalCodes)
The master branch will always be checked with devtools::check
angina_codes <- get_ClinicalCodes(url = "https://clinicalcodes.rss.mhs.man.ac.uk/medcodes/article/6/codelist/angina/download/")
depression_codes <- get_ClinicalCodes(article_id = 6, codelist_name = "depression")
codelists = get_ClinicalCodes(article_id = 6)
all_articles <- all_ClinicalCodes_articles()
all_articles[1,]
## Type Title Journal Year Authors link ID
## 1 QOF Business Rules QOF Business Rules v24 NA 2012 NA https://clinicalcodes.rss.mhs.man.ac.uk/medcodes/article/1/ 1
The ClinicalCodes repository supplies article and codelist metadata in the form of a JSON research object. Research objects contain metadata describing the article (URI, abstract, ID, title, authors, doi, journal etc.), comments on the article, codelist metadata (associated articles, name, url, number of codes in the list, user field names, comments) and optional full codelists.
rClinicalCodes provides functions to access these and to import them as R objects:
RO <- research_object(article_ids = 5, download_codes = TRUE)
ROs <- research_object(article_ids = all_articles$ID[1:3], download_codes = TRUE)
sapply(ROs, names)
## 1: QOF Business Rules v24 2: QOF Business Rules v5 5: Withdrawing Performance Indicators: Retrospective
## [1,] "URI" "URI" "URI"
## [2,] "abstract" "abstract" "abstract"
## [3,] "article_ID" "article_ID" "article_ID"
## [4,] "article_title" "article_title" "article_title"
## [5,] "article_type" "article_type" "article_type"
## [6,] "authors" "authors" "authors"
## [7,] "codelists" "codelists" "codelists"
## [8,] "comments" "comments" "comments"
## [9,] "correspondence_author" "correspondence_author" "correspondence_author"
## [10,] "correspondence_email" "correspondence_email" "correspondence_email"
## [11,] "date_accessed" "date_accessed" "date_accessed"
## [12,] "doi" "doi" "doi"
## [13,] "fulltext" "fulltext" "fulltext"
## [14,] "journal" "journal" "journal"
## [15,] "link" "link" "link"
## [16,] "publication_year" "publication_year" "publication_year"
## [17,] "uploading_user" "uploading_user" "uploading_user"
Once you have downloaded the codelists, you can extract the main keywords from the descriptions
codelist_keywords(angina_codes, extra_stopwords = c("good", "poor"))
## [1] "angina" "anginal" "anginosa" "anginosus"
## [5] "antianginal" "attack" "atypical" "cardiac"
## [9] "cardiomyopathy" "chest" "control" "crescendo"
## [13] "decubitus" "effort" "forms" "hypertension"
## [17] "improving" "infarct" "inversa" "ischaemic"
## [21] "nocturnal" "nos" "onset" "pain"
## [25] "pectoris" "post" "refractory" "rest"
## [29] "stable" "status" "stenocardia" "syncope"
## [33] "syndrome" "therapy" "unstable" "worsening"