tcc: Transform meta-analysis data from long or wide arm-based format into paired-preference format

Description

This function transforms data that are given in wide or long arm-based format (e.g. input format for WinBUGS or JAGS) to a paired-preference format needed as input to mtrank. The function can transform data with binary and continuous arm-based to preference-based format.

Usage

tcc(
  treat,
  event,
  n,
  mean,
  sd,
  data = NULL,
  studlab,
  mcid = NULL,
  lower.equi = NULL,
  upper.equi = NULL,
  small.values = gs("small.values"),
  relax = FALSE,
  level = 0.95,
  sm,
  keepdata = gs("keepdata"),
  ...
)
# S3 method for tcc
print(x, ...)

Value

The initial data in a paired-preference format.
The correspondence between the initial study names (passed in the argument studlab) and the index name of the paired-preference format data.

Arguments

treat: Either a pairwise object, or a list or vector with treatment information for individual treatment arms (see Details).
event: A list or vector with information on number of events for individual treatment arms (see Details).
n: A list or vector with information on number of observations for individual treatment arms (see Details).
mean: A list vector with estimated means for individual treatment arms (see Details).
sd: A list or vector with information on the standard deviation for individual treatment arms (see Details).
data: A data frame containing the study information.
studlab: A vector with study labels.
mcid: A numeric vector specifying the minimal clinically important value (see Details).
lower.equi: A numeric value specifying the lower limit of the range of equivalence (see Details).
upper.equi: A numeric value specifying the upper limit of the range of equivalence (see Details).
small.values: A character string specifying whether small treatment effects indicate a beneficial ("desirable") or harmful ("undesirable") effect.
relax: A logical optional argument. If TRUE it 'relaxes' the tcc to only consider the bounds of ROE when specifying 'wins' and ties. The default FALSE uses the criterion described by Evrenoglou et al. (2024) and considers also the statistical significance on top of the ROE bounds (see Details).
level: The level used to calculate confidence intervals for log-abilities.
sm: The effect measure of interest (see Details).
keepdata: A logical indicating whether original data should be kept in tcc object.
...: Additional arguments (passed on to pairwise).
x: An object of class tcc.

Details

R function mtrank expects data in a paired-preference format, where for each study-specific pairwise comparison in the network a treatment preference or tie is indicated. For example, for the study-specific comparison between treatments A and B the potential outcomes are:

A > B
A < B
A = B

The data transformation takes place based on the study-specific treatment effects and the treatment choice criterion. R function pairwise is called internally to calculate the study-specific treatment effect estimates and standard errors. This ensures that data given in either 'long' or 'wide' arm-based format will be suitably used to calculate the study-specific treatment effect estimates and standard errors while ensuring that a network of multi-arm studies gets an equivalent representation as a network of two-arm studies. It is also possible to provide a pairwise as the main input. In this case, inputs for the arguments event, n, mean, sd, data, studlab, or keepdata are ignored.

This function implements treatment choice criteria based on the method by Evrenoglou et al. (2024). Namely, a range of equivalence (ROE) can be specified by

argument mcid. Then the limits of the ROE will be defined based on the values (i) mcid, 1/mcid for ratio measures and (ii) mcid and -mcid for difference measures.
arguments lower.equi and upper.equi. These arguments allow the users to define their own limits of the ROE, given the restriction that the lower limit will always be smaller than the upper limit.

Note that when the argument mcid is specified, the arguments lower.equi and upper.equi are ignored. Either only the mcid or both of the lower.equi and upper.equi must be specified for the proper definition of the ROE.

After setting the ROE, each study-specific treatment effect will be categorised as a treatment preference or a tie. The argument relax controls the amount of conservatism of the treatment choice criterion. If set to FALSE (default), the treatment choice criterion is equivalent to the one described by Evrenoglou et al. (2024). In this case, study-specific treatment effects need to be both statistically and clinically significant to indicate a treatment preference. If set to TRUE, the criterion is relaxed and the study-specific treatment effects need to be only clinically significant to indicate a treatment preference.

This function can transform data with binary and continuous outcomes. Depending on the outcome, the following arguments are mandatory:

treat, event, n (for binary outcomes);
treat, n, mean, sd (for continuous outcomes).

Finally, the argument sm is used to define the effect measure of interest for transforming the data into paired-preference format; see metabin and metacont for a list of available effect measures.

References

Evrenoglou T, Nikolakopoulou A, Schwarzer G, Rücker G, Chaimani A (2024): Producing treatment hierarchies in network meta-analysis using probabilistic models and treatment-choice criteria. https://arxiv.org/abs/2406.10612

Examples

Run this code

data(diabetes)
#
ranks <- tcc(treat = t, studlab = study, event = r, n = n, data = diabetes,
  mcid = 1.20, sm = "OR", small.values = "desirable")
#
forest(ranks, treat = "ARB")

Run the code above in your browser using DataLab