Closes #11 in stata, plus warning added about length of group names

R error created if group names are too short. V0.9.3 released.

SAS bug with merging soc_codes resolved.

guidance provided in FAQ on systematic & non-systematci

R/eudract/DESCRIPTION

@@ -1,8 +1,8 @@
 Package: eudract
 Type: Package
 Title: Creates Safety Results Summary in XML to Upload to EudraCT
-Date: 2020-04-03
-Version: 0.9.2.9000
+Date: 2021-02-24
+Version: 0.9.3
 Authors@R: c(
     person("Simon", "Bond", email="simon.bond@addenbrookes.nhs.uk", role="cre"),
     person("Beatrice","Pantaleo", email="bp403@medschl.cam.ac.uk", role="aut"))
@@ -15,7 +15,7 @@ Encoding: UTF-8
 LazyData: true
 Imports: tidyr, xslt, dplyr, xml2, utils, magrittr
 Depends: R (>= 3.5.0)
-RoxygenNote: 7.0.1
+RoxygenNote: 7.1.0
 Suggests: 
     testthat,
     knitr,

R/eudract/NAMESPACE

@@ -1,4 +1,4 @@
-# Generated by roxygen2: do not edit by hand
+# Generated by roxygen2: do not edit by hand
 
 S3method(print,safety_summary)
 export(create.safety_summary)

R/eudract/NEWS.md

@@ -1,6 +1,6 @@
 # eudract v0.9.3
 
-Placeholder for the next version
+Error is caused if any of the group names are less than 4 characters in length, as EudraCT will reject this.
 
 
 # eudract v 0.9.2

R/eudract/R/safety_summary.R

@@ -57,6 +57,8 @@ safety_summary <- function(data, exposed, excess_deaths=0, freq_threshold=0, soc
 
   data$group <- factor(as.character(data$group), levels=group_names)
 
+  #Check length of group name text
+  if( min(nchar(group_names))<4){stop("Group names must be at least 4 characters in length.")}
 
   # check length, value and names of excess_deaths
   if( length(excess_deaths)==1 && length(group_names)>1 && excess_deaths==0){

R/eudract/build_dashboard.R

@@ -13,11 +13,11 @@ devtools::document(roclets = c('rd', 'collate', 'namespace'))
 devtools::build()
 
 #Check you've got the right version number
-install.packages("../eudract_0.9.2.tar.gz",
+install.packages("../eudract_0.9.3.tar.gz",
                  repos = NULL, type = "source")
-devtools::build("../eudract_0.9.2.tar.gz", binary=TRUE)
+devtools::build("../eudract_0.9.3.tar.gz", binary=TRUE)
 
-install.packages("V:/STATISTICS/NON STUDY FOLDER/Software/R/R Code Library/cctu_0.9.2.zip",
+install.packages("V:/STATISTICS/NON STUDY FOLDER/Software/R/R Code Library/eudract_0.9.3.zip",
                  repos = NULL, type="win.binary")
 
 #devtools::build( binary=TRUE)

R/eudract/cran-comments.md

@@ -4,6 +4,11 @@
 * Windows Server 2012 R2 x64 (build 9600) on (appveyor), R3.6.1
 * check_win_release
 
+## Change 24FEB2021
+
+an error is created if input fails the check of group names being at least 4 characters in length.
+
+
 ## Change 03APR2020
 
 Minor corrections to test code, setting a resetting options(StringsAsFactors).

R/eudract/tests/testthat/test_errors.R

@@ -62,6 +62,17 @@ test_that("missing variable",{
   expect_error(safety_summary(aes[,-2], exposed=c(700,750,730), soc_index = "soc_term"),"your input data are missing the following variables:")
 })
 
+
+test_that("Too short group names",{
+
+  aes <- read.csv(file.path(path,"data/events.csv"), stringsAsFactors = FALSE)
+  aes$group <- ifelse(aes$group=="Group A", "A", aes$group)
+  expect_error(safety_summary(aes, exposed=c(700,750,730), soc_index = "soc_term"),
+               "Group names must be at least 4 characters in length.")
+
+})
+
+
 test_that("exposed has too few elements",{
   aes <- read.csv(file.path(path,"data/events.csv"), stringsAsFactors = FALSE)
   expect_is(safety_summary(aes, exposed=c(700,750,730), soc_index = "soc_term"),"safety_summary")

R/eudract/vignettes/eudract.Rmd

@@ -4,7 +4,7 @@ author: "Simon Bond"
 date: "`r Sys.Date()`"
 output: rmarkdown::html_vignette
 vignette: >
-  %\VignetteIndexEntry{eudract}
+  %\VignetteIndexEntry{Eudract Safety XML}
   %\VignetteEngine{knitr::rmarkdown}
   %\VignetteEncoding{UTF-8}
 ---

Stata/safety_scriptv0.2.do

@@ -20,7 +20,7 @@ save soc_code, replace
 
 import delimited using ${datafile} , varnames(1) clear
 save raw, replace
-
+* The length of the group names must be 4 characters or longer
 
 
 gen nonserious=1-serious

docs/_posts/2021-02-24-stata-bug-fix.md

@@ -0,0 +1,18 @@
+---
+layout: post
+title: Stata bug fix and minor improvements.
+date: 2021-02-24
+---
+
+A bug was identified in the stata code where the serious events were being used to
+create the output for the non-serious events.
+
+SAS code was in error if the SOC code linked to the soc_code data set using soc_term
+rather than meddra code.
+
+Comments added to SAS and Stata to warn that the group names need to be 4 characters
+in length. An error is created in the R package for this scenario; new version
+released on CRAN.
+
+Advice provide in the FAQ on how to default to systematic, rather than non-systematic,
+as the method of data collection.

docs/faq.md

@@ -63,6 +63,24 @@ to clean the data.
 Yes, please do review and ammend by hand using the Eudract portal manually. You may want to confirm if study-wide default values
 (AE dictionary, version number, incidence threshold,...) are accurate.
 
+## Systematic vs Non-Systematic reporting
+
+The majority of studies use non-systematic reporting, whereby a patient is asked what AEs happened to them.
+The alternative is where a specific list of AEs is provided and the incidence of each AE is explicitly determined.
+
+The default in the tool is to assume that non-systematic reporting was used.
+To change this manually inside the EudraCT portal would require it to be done for each AE in the list,
+which may be too many changes to be practical.  
+
+An automated alternative, _at the user's risk_, is to edit the final xslt file
+[simpleToEudraCT.xlst]( {{ site.ghpath }}/xml%20tools/simpleToEudraCT.xslt)
+in lines 15 and 71.
+
+In the R package, this new xslt file would need to be saved locally and the filepath
+used as an argument to the `xslt=` argument within `eudract_convert()` .
+In SAS, line 297 in the script would need to be edited to use the revised xslt file.
+In Stata, line 226 would need to be edited to use the revised xslt file.
+
 ## I want to use my own code to calculate the summary statistics
 
 As described in the [Specification]( {{ site.ghpath }}/Specification/Eudract%20Tool%20Specification%20V0.3.docx?raw=true ) this is