precautionary
implements new layers of patient-centered safety analysis
for phase 1 dose-escalation trials, adding diagnostics to examine
the safety characteristics of these designs in light of expected
inter-individual variation in pharmacokinetics and pharmacodynamics.
See Norris (2020b), "Retrospective analysis of a fatal dose-finding trial"
arXiv:2004.12755 and (2020c)
"What Were They Thinking? Pharmacologic priors implicit in a choice of 3+3
dose-escalation design" arXiv:2012.05301.
Releases starting with 0.2.3 incorporate fast numerics implemented in Rust, a modern programming language that emphasizes performance and reliability---attributes crucial to applications such as the analysis of clinical trials.
These innovations have delayed review and acceptance by CRAN, pending which
the newest features of precautionary
will be available only here on GitHub.
# Install release version from GitHub
remotes::install_github("dcnorris/precautionary")
# Install obsolete version from CRAN (where review of new Rust library remains pending)
install.package("precautionary")
To date, those features of precautionary
which depend on the Prolog code
in exec/prolog/
have been prebuilt into the package, for example as the
arrays T[,,,]
written into R/sysdata.rda
by exec/make_sysdata_TUb.R
.
Methodologists who wish to examine, recompute and verify these arrays are
advised to install Scryer Prolog.
It is a near-term goal for precautionary
to reveal more transparently
Prolog's special contributions to its analysis of dose-escalation designs.
Please see the vignettes under the Articles tab above.
The precautionary
package is the pointy end of the spear in a larger
DTAT research programme, of which the
following are key outputs. Several of these citations have accompanying
online resources such as web applications. For the key references,
lay explanations are available.
Any scripts or data that you put into this service are public.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.