## package replicateBE on CRAN [R for BE/BA]

Dear all,

package replicateBE is on CRAN. Background information in its Vignette.

Data can be imported from CSV- or XLS(X)-files. Following ten designs are covered:
• Four period full replicates
• TRTR | RTRT
• TRRT | RTTR
• TTRR | RRTT
• TRTR | RTRT | TRRT | RTTR 1
• TRRT | RTTR | TTRR | RRTT 1
• Two period full replicate
• TR | RT | TT | RR 2
• Three period full replicates
• TRT | RTR
• TRR | RTT
• Three period partial replicates
• TRR | RTR | RRT
• TRR | RTR 3
Implemented:
• Estimation of CVwR (and CVwT in full replicates)
A linear model of log-transformed PK responses and effects
sequence, subject(sequence), period
of the reference (and test, if possible) treatment – where all effects are fixed (i.e., ANOVA).
Estimated via function lm() of library stats.
In full replicate designs assessment of $s_{wT}/s_{wR}$ and the upper confidence limit of $\sigma_{wT}/\sigma_{wR}$ for the WHO’s reference-scaling of AUC.

• Method A
A linear model of log-transformed PK responses and effects
sequence, subject(sequence), period, treatment
– where all effects are fixed (i.e., ANOVA).
Estimated via function lm() of library stats.

• Method B
A linear mixed model of log-transformed PK responses and effects
sequence, subject(sequence), period, treatment
– where subject(sequence) is a random effect and all others are fixed.
Three options
1. Estimated via function lme() of library nlme. Uses degrees of freedom equivalent to SAS’ DDFM=CONTAIN and Phoenix/WinNonlin’s DF Residual. Implicitly preferred according to the EMA’s Q&A document and hence, the default.
2. Estimated via function lmer() of library lmerTest. Uses Satterthwaite’s degrees of freedom equivalent to SAS’ DDFM=SATTERTHWAITE and Phoenix/WinNonlin’s DF Satterthwaite. This is the only option available in SPSS.
Potentially better for highly incomplete data.
3. Estimated via function lmer() of library lmerTest. Uses the Kenward-Roger approximation equivalent to SAS’ DDFM=KENWARDROGER. This is the only option available in JMP.
Potentially better for highly incomplete data.
• ABE
Conventional Average Bioequivalence. Optionally with tighter (EMA: NTIDs) or wider limits (GCC: Cmax).
Assessment of potential outliers in the ABEL-methods is done by box plots of studentized model residuals.

Installation/Update of the latest release from CRAN with

install.packages("replicateBE", repos = "https://cloud.r-project.org/")

Installation of the development version from GitHub:

devtools::install_github("Helmut01/replicateBE")

Enjoy!

Bug reports to GitHub or the maintainer.

1. Confounded effects (design not recommended).
2. Balaam’s design (not recommended due to poor power characteristics).
3. Extra-reference design; biased in the presence of period effects (design not recommended).

Cheers,
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. ☼
Science Quotes