The hardware and bandwidth for this mirror is donated by dogado GmbH, the Webhosting and Full Service-Cloud Provider. Check out our Wordpress Tutorial.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]dogado.de.
Simulate adverse event reporting in clinical trials with the goal of detecting under-reporting sites.
Monitoring of Adverse Event (AE) reporting in clinical trials is important for patient safety. We use bootstrap-based simulation to assign an AE under-reporting probability to each site in a clinical trial. The method is inspired by the ‘infer’ R package and Allen Downey’s blog article: “There is only one test!”.
install.packages("simaerep")
You can install the development version from GitHub with:
# install.packages("devtools")
::install_github("openpharma/simaerep") devtools
simaerep
has been published as workproduct of the
Inter-Company Quality Analytics (IMPALA) consortium. IMPALA
aims to engage with Health Authorities inspectors on defining guiding
principles for the use of advanced analytics to complement, enhance and
accelerate current QA practices. simaerep
has initially
been developed at Roche but is currently evaluated by other companies
across the industry to complement their quality assurance activities (see
testimonials).
Koneswarakantha, B., Adyanthaya, R., Emerson, J. et al. An Open-Source R Package for Detection of Adverse Events Under-Reporting in Clinical Trials: Implementation and Validation by the IMPALA (Inter coMPany quALity Analytics) Consortium. Ther Innov Regul Sci 58, 591–599 (2024). https://doi.org/10.1007/s43441-024-00631-8
Koneswarakantha, B., Barmaz, Y., Ménard, T. et al. Follow-up on the Use of Advanced Analytics for Clinical Quality Assurance: Bootstrap Resampling to Enhance Detection of Adverse Event Under-Reporting. Drug Saf (2020). https://doi.org/10.1007/s40264-020-01011-5
Download as pdf in the release section generated using thevalidatoR.
Recommended Threshold:
aerep$dfeval$prob_low_prob_ur: 0.95
suppressPackageStartupMessages(library(simaerep))
suppressPackageStartupMessages(library(tidyverse))
suppressPackageStartupMessages(library(knitr))
set.seed(1)
<- sim_test_data_study(
df_visit n_pat = 1000, # number of patients in study
n_sites = 100, # number of sites in study
frac_site_with_ur = 0.05, # fraction of sites under-reporting
ur_rate = 0.4, # rate of under-reporting
ae_per_visit_mean = 0.5 # mean AE per patient visit
)
$study_id <- "A"
df_visit
%>%
df_visit select(study_id, site_number, patnum, visit, n_ae) %>%
head(25) %>%
::kable() knitr
study_id | site_number | patnum | visit | n_ae |
---|---|---|---|---|
A | S0001 | P000001 | 1 | 0 |
A | S0001 | P000001 | 2 | 1 |
A | S0001 | P000001 | 3 | 1 |
A | S0001 | P000001 | 4 | 2 |
A | S0001 | P000001 | 5 | 3 |
A | S0001 | P000001 | 6 | 3 |
A | S0001 | P000001 | 7 | 3 |
A | S0001 | P000001 | 8 | 3 |
A | S0001 | P000001 | 9 | 3 |
A | S0001 | P000001 | 10 | 3 |
A | S0001 | P000001 | 11 | 3 |
A | S0001 | P000001 | 12 | 3 |
A | S0001 | P000001 | 13 | 4 |
A | S0001 | P000001 | 14 | 4 |
A | S0001 | P000001 | 15 | 4 |
A | S0001 | P000001 | 16 | 6 |
A | S0001 | P000001 | 17 | 6 |
A | S0001 | P000002 | 1 | 0 |
A | S0001 | P000002 | 2 | 0 |
A | S0001 | P000002 | 3 | 0 |
A | S0001 | P000002 | 4 | 0 |
A | S0001 | P000002 | 5 | 0 |
A | S0001 | P000002 | 6 | 0 |
A | S0001 | P000002 | 7 | 0 |
A | S0001 | P000002 | 8 | 1 |
<- simaerep(df_visit)
aerep
plot(aerep, study = "A")
Left panel shows mean AE reporting per site (lightblue and darkblue lines) against mean AE reporting of the entire study (golden line). Single sites are plotted in descending order by AE under-reporting probability on the right panel in which grey lines denote cumulative AE count of single patients. Grey dots in the left panel plot indicate sites that were picked for single plotting. AE under-reporting probability of dark blue lines crossed threshold of 95%. Numbers in the upper left corner indicate the ratio of patients that have been used for the analysis against the total number of patients. Patients that have not been on the study long enough to reach the evaluation point (visit_med75, see introduction) will be ignored.
Following the recommendation of our latest performance benchmark statistical performance can be increased by using the inframe algorithm without multiplicity correction.
Note that the plot is more noisy because no patients are excluded and only a few patients contribute to the event count at higher visits
Recommended Threshold:
aerep$dfeval$prob_low_prob_ur: 0.99
<- simaerep(
aerep
df_visit,inframe = TRUE,
visit_med75 = FALSE,
mult_corr = FALSE
)
plot(aerep, study = "A")
The inframe
algorithm uses only dbplyr
compatible table operations and
can be executed within a database backend as we demonstrate here using
duckdb
.
However, we need to provide a in database table that has as many rows
as the desired replications in our simulation, instead of providing an
integer for the r
parameter.
<- DBI::dbConnect(duckdb::duckdb(), dbdir = ":memory:")
con <- tibble(rep = seq(1, 1000))
df_r
::copy_to(con, df_visit, "visit")
dplyr::copy_to(con, df_r, "r")
dplyr
<- tbl(con, "visit")
tbl_visit <- tbl(con, "r")
tbl_r
<- simaerep(
aerep
tbl_visit,r = tbl_r,
inframe = TRUE,
visit_med75 = FALSE,
mult_corr = FALSE
)
plot(aerep, df_visit = tbl_visit)
#> study = NULL, defaulting to study:A
::dbDisconnect(con) DBI
These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.
Health stats visible at Monitor.