ComputePredictiveCheck {MIfuns} | R Documentation |
Computes individual patient parameters for observed data and for each simulated study and computes statistics of individual parameters for each study (beta)
ComputePredictiveCheck(NonmemData, SimulatedData, SimpleFun = c("min", "max", "mean", "median"), CombFun = c("Tmin", "Tmax", "AUC"), StatFun = c("min", "max", "mean", "median", "Q25", "Q75"), IDname = "ID", TimeName = "TIME", DVname = "DV", CName = "C", Ignore = "C")
NonmemData |
data frame that contains the data set used for simulations. This data frame should contain MDV column. Only records with MDV=0 will be used. |
SimulatedData |
simulated data (possibly, output of the simulations with uncertainty function) |
SimpleFun |
list of functions (in quotes) that accept vector of DV values as an input and return a scalar. These functions will be used to compute individual parameters. |
CombFun |
list of functions (in quotes) that accept data frame with three columns,
ID, TIME and DV, and return the data frame with two columns, ID and the columns with the name
that coincide with the name of the function. These functions will be used to compute
individual parameters in the case when computation requires not only the vector of observations
but also the times of observations. Examples of this type of function are AUC , Tmin
and Tmax . The user can also create custom functions and use those with ComputePredictiveCheck . |
StatFun |
list of functions (in quotes) that accept vector of values as an input and return
a scalar. These functions will be used to compute study statistics. Examples of this type of function
are Q25 and Q75 . The user can also create custom functions and use those with
ComputePredictiveCheck . |
IDname |
Name of the column that assigns unique patient ID. Default value is “ID”. Each of the data
frames NonmemData and code{SimulatedData} should have IDname column. |
TimeName |
Name of the column that assigns time of the observation. Default value is “TIME”. Each of
the data frames NonmemData and code{SimulatedData} should have TimeName column. |
DVname |
Name of the column that assigns dependent variable value. Default value is “DV”. Each of the
data frames NonmemData and SimulatedData should have DVname column. DVname
column of NonmemData data frame should contain observed values. DVname column of
SimulatedData data frame should contain simulated values. |
CName |
Name of the column that assigns the comment field. Default value is “C”. The data frame
NonmemData should have Cname column. Records with the character assigned by Ignore
parameter in the Cname column will be removed from the analysis. |
Ignore |
The character that denotes the record of the NonmemData data frame that should be
removed from the analysis. Default value is “C”. |
~~ If necessary, more details than the description above ~~
The output of the program is the list with two elements, Combined and Stats. Combined is the data frame that
contains individual parameters. It has the columns ID, STUDY and one column for each element of the
SimpleFun
and CombFun
vectors. Stats is the list, a collection of data frames, one per statistics
requested. The names of the data frame are given by StatFun
vector. Individual element of the list can be
called as Stats[[k]]
where k
is the order of the desired data frame in the StatFun
list. The
same element can be called by name as Stats[ StutFun[k] ]
. Each data frame (element of Stats list) contains
summaries of individual parameters by study. It has STUDY column and one column for each element of the SimpleFun
and code{CombFun} vectors. STUDY column give the order number of the simulated study. STUDY=0 record describes
observed data.
Developed by Metrum Institute (Bill Knebel, Tim Bergsma, and others).
MIfuns package available from Metrum Institute website (http://metruminstitute.org/
).