R: Simultaneous Confidence Intervals for Ratios of Means of...
SimCiRat
R Documentation
Simultaneous Confidence Intervals for Ratios of Means of Multiple Endpoints
Description
Simultaneous confidence intervals for ratios of contrasts (linear functions) of
normal means (e.g., "Dunnett", "Tukey", "Williams" ect.) when there is more than
one primary response variable (endpoint). The procedure of Hasler and Hothorn
(2012) is applied for ratios of means of normally distributed data. The covariance
matrices (containing the covariances between the endpoints) may be assumed to be
equal or possibly unequal for the different groups (Hasler, 2014). For the case of
only a single endpoint and unequal covariance matrices (variances), the procedure
coincides with the PI procedure of Hasler and Hothorn (2008).
Usage
SimCiRat(data, grp, resp = NULL, type = "Dunnett", base = 1, Num.Contrast = NULL,
Den.Contrast = NULL, alternative = "two.sided", covar.equal = FALSE,
conf.level = 0.95)
Arguments
data
a data frame containing a grouping variable and the endpoints as
columns
grp
a character string with the name of the grouping variable
resp
a vector of character strings with the names of the endpoints; if
resp=NULL (default), all column names of the data frame
without the grouping variable are chosen automatically
type
a character string, defining the type of contrast, with the following
options:
"Dunnett": many-to-one comparisons, with control in the
denominator
"Tukey": all-pair comparisons
"Sequen": comparisons of consecutive groups, where the group
with lower order is the denomniator
"AVE": comparison of each group with average of all others,
where the average is taken as denominator
"GrandMean": comparison of each group with grand mean of all
groups, where the grand mean is taken as
denominator
"Changepoint": ratios of averages of groups of higher order
divided by averages of groups of lower order
"Marcus": Marcus contrasts as ratios
"McDermott": McDermott contrasts as ratios
"Williams": Williams contrasts as ratios
"UmbrellaWilliams": Umbrella-protected Williams contrasts as
ratios
note that type is ignored if Num.Contrast and
Den.Contrast are specified by the user (see below)
base
a single integer specifying the control (i.e. denominator) group for
Dunnett contrasts, ignored otherwise
Num.Contrast
a numerator contrast matrix, where columns correspond to
groups and rows correspond to contrasts
Den.Contrast
a denominator contrast matrix, where columns correspond to
groups and rows correspond to contrasts
alternative
a character string specifying the alternative hypothesis,
must be one of "two.sided" (default), "greater"
or "less"
covar.equal
a logical variable indicating whether to treat the covariance
matrices (containing the covariances between the endpoints)
for the different groups as being equal;
if TRUE then the pooled covariance matrix is used,
otherwise the Satterthwaite approximation to the degrees of
freedom is used according to Hasler and Hothorn (2008)
conf.level
a numeric value defining the simultaneous confidence level
Details
The interest is in simultaneous confidence intervals for several ratios of linear
combinations (contrasts) of treatment means in a one-way ANOVA model, and
simultaneously for multiple endpoints. For example, corresponding intervals for the
all-pair comparison of Tukey (1953) and the many-to-one comparison of Dunnett (1955)
for ratios of means are implemented, but allowing for multiple endpoints. Also, the
user is free to create other interesting problem-specific contrasts. An approximate
multivariate t-distribution is used to calculate lower and upper limits (see
Hasler and Hothorn, 2012). Simultaneous tests based on these intervals control
the familywise error rate in an admissible range and in the strong sense. The
covariance matrices of the treatment groups (containing the covariances between the
endpoints) can be assumed to be equal (covar.equal=TRUE) or unequal
(covar.equal=FALSE). If being equal, the pooled covariance matrix is used,
otherwise approximations to the degrees of freedom (Satterthwaite, 1946) are used
(see Hasler, 2014). Unequal covariance matrices occure if variances or correlations
of some endpoints differ depending on the treatment groups.
Value
An object of class SimCi containing:
estimate
a matrix of estimated differences
lower.raw
a matrix of raw (unadjusted) lower limits
upper.raw
a matrix of raw (unadjusted) upper limits
lower
a matrix of lower limits adjusted for multiplicity
upper
a matrix of upper limits adjusted for multiplicity
CorrMatDat
either the estimated common correlation matrix of the data
(covar.equal=TRUE) or the list of the different (one for
each treatment) estimated correlation matrices of the data
(covar.equal=FALSE)
CorrMatComp
the estimated correlation matrix to be used for the multivariate
t-distribution
degr.fr
either a single degree of freedom (covar.equal=TRUE) or a
vector of degrees of freedom (covar.equal=FALSE) related
to the comparisons
Note
All measurement objects of each treatment group must have values for each endpoint.
If there are missing values then the procedure stops. If covar.equal=TRUE,
then the number of endpoints must not be greater than the total sample size minus
the number of treatment groups. If covar.equal=FALSE, the number of endpoints
must not be greater than the minimal sample size minus 1. Otherwise the procedure
stops.
All intervals have the same direction for all comparisons and endpoints
(alternative="..."). In case of doubt, use "two.sided".
In contrast to simultaneous confidence intervals for differences, the correlation
matrix for the multivariate t-distribution depends on the unknown ratios. The
same problem also arises for the degrees of freedom if the covariance matrices for
the different groups are assumed to be unequal (covar.equal=FALSE). Both
problems are handled by a plug-in approach, see the references therefore.
Author(s)
Mario Hasler
References
Hasler, M. (2014): Multiple contrast tests for multiple endpoints in the presence of
heteroscedasticity. The International Journal of Biostatistics 10, 17–28.
Hasler, M. and Hothorn, L.A. (2012): A multivariate Williams-type trend procedure.
Statistics in Biopharmaceutical Research 4, 57–65.
Hasler, M. and Hothorn, L.A. (2008): Multiple contrast tests in the presence of
heteroscedasticity. Biometrical Journal 50, 793–800.
Dilba, G. et al. (2006): Simultaneous confidence sets and confidence intervals for
multiple ratios. Journal of Statistical Planning and Inference 136, 2640–2658.
Satterthwaite, F.E. (1946): An approximate distribution of estimates of variance
components. Biometrics 2, 110–114.
See Also
SimCiDiff, SimTestRat,
SimTestDiff
Examples
# Example 1:
# Simultaneous confidence intervals for ratios of means, related to a
# comparison of the groups B and H against the standard S, on endpoint
# Thromb.count, assuming unequal variances for the groups. This is an extension
# of the well-known Dunnett-intervals to the case of heteroscedasticity and in
# terms of ratios of means instead of differences.
data(coagulation)
interv1 <- SimCiRat(data=coagulation, grp="Group", resp="Thromb.count",
type="Dunnett", base=3, alternative="greater", covar.equal=FALSE)
interv1
# Example 2:
# Simultaneous confidence intervals for ratios of means, related to a
# comparison of the groups B and H against the standard S, simultaneously on
# all endpoints, assuming unequal covariance matrices for the groups. This is an
# extension of the well-known Dunnett-intervals to the case of heteroscedasticity
# and multiple endpoints, and in terms of ratios of means instead of differences.
data(coagulation)
interv2 <- SimCiRat(data=coagulation, grp="Group", resp=c("Thromb.count","ADP","TRAP"),
type="Dunnett", base=3, alternative="greater", covar.equal=FALSE)
summary(interv2)
Results
R version 3.3.1 (2016-06-21) -- "Bug in Your Hair"
Copyright (C) 2016 The R Foundation for Statistical Computing
Platform: x86_64-pc-linux-gnu (64-bit)
R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under certain conditions.
Type 'license()' or 'licence()' for distribution details.
R is a collaborative project with many contributors.
Type 'contributors()' for more information and
'citation()' on how to cite R or R packages in publications.
Type 'demo()' for some demos, 'help()' for on-line help, or
'help.start()' for an HTML browser interface to help.
Type 'q()' to quit R.
> library(SimComp)
> png(filename="/home/ddbj/snapshot/RGM3/R_CC/result/SimComp/SimCiRat.Rd_%03d_medium.png", width=480, height=480)
> ### Name: SimCiRat
> ### Title: Simultaneous Confidence Intervals for Ratios of Means of
> ### Multiple Endpoints
> ### Aliases: SimCiRat
> ### Keywords: htest
>
> ### ** Examples
>
> # Example 1:
> # Simultaneous confidence intervals for ratios of means, related to a
> # comparison of the groups B and H against the standard S, on endpoint
> # Thromb.count, assuming unequal variances for the groups. This is an extension
> # of the well-known Dunnett-intervals to the case of heteroscedasticity and in
> # terms of ratios of means instead of differences.
>
> data(coagulation)
>
> interv1 <- SimCiRat(data=coagulation, grp="Group", resp="Thromb.count",
+ type="Dunnett", base=3, alternative="greater", covar.equal=FALSE)
> interv1
Simultaneous 95% confidence intervals for ratios of means of multiple endpoints
Assumption: Heterogeneous covariance matrices for the groups
comparison endpoint estimate lower.raw upper.raw lower upper
1 B/S Thromb.count 1.139 0.9609 Inf 0.9287 Inf
2 H/S Thromb.count 1.050 0.8533 Inf 0.8175 Inf
>
> # Example 2:
> # Simultaneous confidence intervals for ratios of means, related to a
> # comparison of the groups B and H against the standard S, simultaneously on
> # all endpoints, assuming unequal covariance matrices for the groups. This is an
> # extension of the well-known Dunnett-intervals to the case of heteroscedasticity
> # and multiple endpoints, and in terms of ratios of means instead of differences.
>
> data(coagulation)
>
> interv2 <- SimCiRat(data=coagulation, grp="Group", resp=c("Thromb.count","ADP","TRAP"),
+ type="Dunnett", base=3, alternative="greater", covar.equal=FALSE)
> summary(interv2)
Numerator contrast matrix:
B H S
B/S 1 0 0
H/S 0 1 0
Denominator contrast matrix:
B H S
B/S 0 0 1
H/S 0 0 1
Estimated covariance matrices of the data:
$B
Thromb.count ADP TRAP
Thromb.count 0.0626 0.0565 -0.0102
ADP 0.0565 0.0638 0.0054
TRAP -0.0102 0.0054 0.0963
$H
Thromb.count ADP TRAP
Thromb.count 0.0943 0.0637 0.0663
ADP 0.0637 0.0518 0.0446
TRAP 0.0663 0.0446 0.1157
$S
Thromb.count ADP TRAP
Thromb.count 0.0318 0.0132 0.0598
ADP 0.0132 0.0079 0.0269
TRAP 0.0598 0.0269 0.1376
Estimated correlation matrices of the data:
$B
Thromb.count ADP TRAP
Thromb.count 1.0000 0.8937 -0.1314
ADP 0.8937 1.0000 0.0687
TRAP -0.1314 0.0687 1.0000
$H
Thromb.count ADP TRAP
Thromb.count 1.0000 0.9121 0.6348
ADP 0.9121 1.0000 0.5770
TRAP 0.6348 0.5770 1.0000
$S
Thromb.count ADP TRAP
Thromb.count 1.0000 0.8338 0.9033
ADP 0.8338 1.0000 0.8161
TRAP 0.9033 0.8161 1.0000
Estimated correlation matrix of the comparisons:
[,1] [,2] [,3] [,4] [,5] [,6]
[1,] 1.0000 0.8494 0.3781 0.3198 0.2025 0.4258
[2,] 0.8494 1.0000 0.2918 0.1697 0.1545 0.2448
[3,] 0.3781 0.2918 1.0000 0.3740 0.2566 0.6102
[4,] 0.3198 0.1697 0.3740 1.0000 0.8869 0.7084
[5,] 0.2025 0.1545 0.2566 0.8869 1.0000 0.5874
[6,] 0.4258 0.2448 0.6102 0.7084 0.5874 1.0000
comparison endpoint estimate lower.raw upper.raw lower upper
1 B/S Thromb.count 1.139 0.9609 Inf 0.8848 Inf
2 B/S ADP 1.262 1.0853 Inf 1.0116 Inf
3 B/S TRAP 1.145 0.8370 Inf 0.7231 Inf
4 H/S Thromb.count 1.050 0.8533 Inf 0.7712 Inf
5 H/S ADP 1.104 0.9522 Inf 0.8893 Inf
6 H/S TRAP 1.098 0.7907 Inf 0.6771 Inf
>
>
>
>
>
> dev.off()
null device
1
>