View source: R/samplesize_scABEL.R
sampleN.scABEL  R Documentation 
This function performs the sample size estimation via power calculations of the BE decision via scaled (expanded) BE acceptance limits, based on simulations.
sampleN.scABEL(alpha = 0.05, targetpower = 0.8, theta0, theta1, theta2, CV, design = c("2x3x3", "2x2x4", "2x2x3"), regulator, nsims = 1e+05, nstart, imax = 100, print = TRUE, details = TRUE, setseed = TRUE)
alpha 
Type I error probability. Per convention mostly set to 0.05. 
targetpower 
Power to achieve at least. Must be >0 and <1. 
theta0 
‘True’ or assumed T/R ratio. 
theta1 
Conventional lower ABE limit to be applied in the mixed procedure if

theta2 
Conventional upper ABE limit to be applied in the mixed procedure if

CV 
Intrasubject coefficient(s) of variation as ratio (not percent).

design 
Design of the study to be planned. 
regulator 
Regulatory settings for the widening of the BE acceptance limits. 
nsims 
Number of simulations to be performed to obtain the (empirical) power.
The default value 100,000 = 1e+5 is usually sufficient. Consider to rise
this value if 
nstart 
Set this to a start for the sample size search if a previous run failed. 
imax 
Maximum number of steps in sample size search. Defaults to 100. 
print 
If 
details 
If set to 
setseed 
Simulations are dependent on the starting point of the (pseudo) random number
generator. To avoid differences in power for different runs a 
The simulations are done via the distributional properties of the statistical
quantities necessary for deciding BE based on ABEL (‘Average Bioequivalence with Expanded Limits’). For more details see a description in the /doc
subdirectory of the package.
Function sampleN.scABEL()
is based on power calculations via simulations
using the distributional characteristics of the ‘key’ statistics obtained from
the EMA recommended evaluation via ANOVA if regulator="EMA"
or if the
regulator component est_method
is set to "ANOVA"
if regulator is an object
of class 'regSet'.
Otherwise, the simulations are based on the distributional characteristis of the
‘key’ statistics obtained from evaluation via intrasubject contrasts (ISC),
as recommended by the FDA.
The estimated sample size gives always the total number of subjects
(not subject/sequence – like in some other software packages).
Function sampleN.scABEL2()
is solely based on power calculations via
simulation using the distributional characteristics of the ‘key’ statistics
obtained from evaluation via intrasubject contrasts (ISC). This function is deprecated.
Returns a data.frame with the input settings and sample size results.
The Sample size
column contains the total sample size.
The nlast
column contains the last n
value. May be useful for restarting.
Although some designs are more ‘popular’ than others, sample size estimations are valid for all of the following designs:
"2x2x4"  TRTR  RTRT 
TRRT  RTTR  
TTRR  RRTT  
"2x2x3"  TRT  RTR 
TRR  RTT  
"2x3x3"  TRR  RTR  RRT 
The sample size estimation for extreme theta0
(<0.83 or >1.21) may be very
time consuming and will eventually also fail since the start values chosen are
not really reasonable in that ranges. This is especially true in the range around
CV = 0.3 and regulatory constant according to FDA.
If you really need sample sizes in that range be prepared to restart the sample
size estimation via the argument nstart
.
Since the dependence of power from n is very flat in the mentioned region you may
also consider to adapt the number of simulations not to get caught in the simulation
error trap.
If results of power.scABEL
are expected to be inaccurate (partial
replicate design with unbalanced sequences and/or heteroscedasticity in the case of CV_{wT} > CV_{wR}, subject data via sampleN.scABEL.sdsims
should be simulated instead. Very time consuming (easily 100times slower)! Subject data simulations are only supported for regulator="EMA"
and regulator="GCC"
.
We are doing the sample size estimation only for balanced designs since the
break down of the total subject number in case of unbalanced sequence groups
is not unique. Moreover the formulas used are only for balanced designs.
In case of regulator="FDA"
the sample size is only approximate since
the BE decision method is not exactly what is expected by the FDA. But the two Lászlós state that the scABEL method should be ‘operationally’ equivalent to the
FDA method. Thus the sample size should be comparable.
Consider in case of regulator="FDA"
to use the function
sampleN.RSABE()
instead.
In case of regulator="HC"
the underlying power is only approximative
since the Health Canada recommends evaluation by a mixed model approach.
But this could only implemented via subject data simulations which are very
time consuming.
The minimum sample size is 6, even if the power is higher than the intended
targetpower.
D. Labes
Tóthfalusi L, Endrényi L. Sample Sizes for Designing Bioequivalence Studies for Highly Variable Drugs. J Pharm Pharmaceut Sci. 2011;15(1):73–84. open access
power.scABEL
, sampleN.scABEL.sdsims
, sampleN.RSABE
,
reg_const
# using all the defaults: # partial replicate design, targetpower=80%, # true assumed ratio = 0.90, 1E+5 simulated studies # ABE limits, PE constraint 0.8  1.25 # EMA regulatory settings sampleN.scABEL(CV = 0.3) # should result in a sample size n=54, power=0.8159 # Now with former (inofficial) ANVISA settings, CVswitch=40% # (since 2016 ANVISA uses the same settings as EMA) reg < reg_const("USER", r_const = 0.76, CVswitch = 0.4, CVcap = 0.5) reg$name < "Old ANVISA" sampleN.scABEL(CV = 0.3, regulator = reg) # should result in a sample size n=60, power=0.8101 # For the full replicate design, target power = 90% # true assumed ratio = 0.9, FDA regulatory settings # sims based on evalaution via ISC sampleN.scABEL(CV = 0.4, targetpower = 0.9, design = "2x2x4", regulator = "FDA") # should result in a sample size n=32, power=0.9125 # Fixed wider limits (0.7500  1.3333) for the GCC sampleN.scABEL(CV = 0.4, targetpower = 0.9, design = "2x2x4", regulator = "GCC") # should result in a sample size n=40, power=0.9039
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.