vigicaen (development version)
New features
tb_vigibase()
now accepts a new argumentrm_suspdup
(default TRUE) to automatically remove cases identified as suspected duplicates (fromSUSPECTEDDUPLICATES.txt
) when creating the main tables. You can disable this behavior by settingrm_suspdup = FALSE
.tb_vigibase()
now supports a new argumentoverwrite_existing_tables
(default FALSE). If set to FALSE, the function will skip the creation of any .parquet tables that already exist, allowing you to resume the process after a crash or interruption without rebuilding completed tables. Set to TRUE to force rebuilding all tables. This makes the table-building process more robust and efficient for large datasets or limited hardware.-
vigi_routine()
now supports two new arguments:-
suspect_only
(default FALSE): controls whether only suspected drug cases are included in IC analysis. -
d_code_2
(optional): if provided, a single analysis is performed on cases exposed to both drugs ind_code
andd_code_2
. A CLI message is displayed when dual drug analysis is performed.
-
Bug fixes
-
vigi_routine()
now has left/right justification of the patient label centered on 3 months (90 days) instead of the observed median. This resolves an issue where the label could be misplaced on the time to onset plot when the median did not match the visual center of the graph.(#145)
Documenting
The ‘smq’ argument in
get_llt_smq()
no longer requires length 1.(#126)The example for
get_atc_code()
now uses explicit names for atc_sel.(#126)The Value section of
desc_tto()
now clarifies the default descriptive format and how to change it.(#126)Typos in routine pharmacovigilance vignette.(#126)
vigicaen 0.15.6
CRAN release: 2025-03-13
New features
vigi_routine()
now plots details on drug liability and rechallenge when available (#119).desc_rch()
,desc_dch()
anddesc_outcome()
now handle out of memory arrow Table (#127).compute_dispro()
now handles factors asx
ory
arguments, if levels are strictly equal to 0 and 1 (#134).
Minor
Error and warnings all turned into
cli
syntax. Gathered redundant checkers internally (#133).vigi_routine()
now works with out of memory arrow Tables. Additional checkers tovigi_routine()
for input data type.compute_or_mod()
now has an informative error message if either argsestimate
orstd_er
were not supplied (#123).get_drecno()
now handles non-WHO names matching a unique DrecNo that has several WHO names, and throws an error if matching multiple DrecNos (#139).Adding dependency to
glue
.
vigicaen 0.15.3
Minor
Addressing CRAN queries on package DESCRIPTION.
Condensing examples and
tb_vigibase()
default output.
vigicaen 0.15.1
New features
tb_*
family now has acli
style progress barget_llt_soc()
output is reworked withcli
features. The function also now correctly supports non-data.table data.frame.vigi_routine()
fails if there are no drug or adr cases found indemo_data
.
vigicaen 0.15.0
Breaking changes
-
mp
replacesmp_short
throughout the package (#110). Users will have to change existing code to replace calls tomp_short
withmp
.# the old way mp_short <- dt_parquet(path_who, "mp_short") d_names <- list(drug1 = "paracetamol") d_drecno <- get_drecno(d_names, mp_short = mp_short) # the new way mp <- dt_parquet(path_who, "mp") d_names <- list(drug1 = "paracetamol") d_drecno <- get_drecno(d_names, mp = mp)
-
In
get_llt_smq()
,smq_list
andsmq_content
replacesmq_list_content
(#74). Users will have to change existing code to replace calls tosmq_list_content
withsmq_list
andsmq_content
. Example tablessmq_list_
andsmq_content_
are added to the package. This change was made to streamline the process of querying sub-SMQs.# the old way smq_list_content <- dt_parquet(path_meddra, "smq_list_content") smq_llt <- get_llt_smq( list(ihd = "Ischaemic heart disease (SMQ)"), smq_list_content) # the new way smq_list <- dt_parquet(path_meddra, "smq_list") smq_content <- dt_parquet(path_meddra, "smq_content") smq_llt <- get_llt_smq( list(ihd = "Ischaemic heart disease (SMQ)"), smq_list = smq_list, smq_content = smq_content)
compute_dispro()
is the new name ofcompute_or_abcd()
, to reflect that other disproportionality measures are available.
New features
Args
inspect
andshow_all
ofget_drecno()
are replaced byverbose
(#102).get_drecno()
andget_llt_smq()
messages are cleaner and written with thecli
package.New
top_n
argument added toscreen_adr()
(#86).screen_drug()
let you screen most reported drugs indrug
(#103).get_llt_smq()
now queries sub-SMQs and return all relevant codes (#74).Order of output columns in
compute_dispro()
has changed, to allow easier access to formatted values. Also new argumentsmin_n_obs
andexport_raw_values
(#101).tb_subset()
handles ID code lists as input, instead of numeric vectors, to make workflow more consistent with other processes likeadd_*
. (#112)
Minor and bug Fixes
screen_adr()
now counts effects at the case level (#86).arg
data_type
ofadd_*
functions is now deprecated. Data_type is internally detected.add_drug()
andadd_adr()
have informative error if id code list is incorrect (#88).
vigicaen 0.14.0
New features
#90 Introducing vigi_routine()
to streamline the process for pharmacovigilance practitioners. This function is intended for non-expert users, to produce an Information Component and time to onset graph.
#94 The luda
table no longer exists. It is replaced by an augmented version of link
- built directly from tb_vigibase()
, which is used downstream in place of luda.
Minor
#92 tb_main()
and tb_sub()
are fusionned into tb_vigibase()
.
#93 It is now possible to export raw data composing the value
column of desc_facvar()
and desc_cont()
, with argument export_raw_values
.
#83 vignette("descriptive")
was translated in English.
Bug fixes
#89 and #87 the get_*
family of functions now correctly handle out of memory tables (by pulling them into memory). The same is true for extract_tto()
.
#85 Useless default values to some arguments of extract_tto()
were removed.
vigicaen 0.13.5
Patch
#91 tb_* family correctly handles different ways of providing paths, and check for the existence of the files before working on tables.
vigicaen 0.13.3
Patch
#75 clearer example for get_llt_soc()
#76 SizeUnit_Lx file in sub, which is encoded in an old format, is now appropriately treated as such.