Display Balance Statistics in a Table for match_time objects

This function extends the bal.tab function from the cobalt package to allow the input of a match_time object. This allows users to directly call the bal.tab function on output of the match_time function, mimicking the usage of the same function when using the matchit function of the MatchIt package. The existence of this function also allows users to directly call the love.plot function on match_time objects.

Usage

# S3 method for class 'match_time'
bal.tab(x, s.d.denom, remove_unmatched=TRUE,
        n_required=x$info$ratio, ...)

Arguments

x: A match_time object created using the match_time function.
s.d.denom: Argument of the same name in bal.tab.
remove_unmatched: Whether to remove unmatched individuals before calculating the balance statistics. Internally, the get_match_data function is called for this, please see the documentation of that function for more information. This has no influence on the computed sample sizes.
n_required: Same as the argument of the same name in get_match_data.
...: Further arguments passed to bal.tab.

Author

Robin Denz

Details

All balance statistics calculated using this method are always considered "unadjusted". The reason is that it is not possible to make a distinction between "unadjusted" and "adjusted" balance statistics directly when using time-dependent matching, because balance is assessed at "baseline" (e.g. the time at which individuals were first included into the matching process). This is easy for the matched data. Here we just use the covariate value at time of entry (.treat_time), which is the whole point of the matching process. Since there is no such baseline period for the unmatched data, however, there are no such balance statistics to be calculated.

The only conceivable strategy to get "unadjusted" and "adjusted" estimates would be to only match on time (using match_method="none" in match_time) and to calculate the balance statistics immediatly afterwards. Then one could adjust for unbalanced covariates using any regular method such as weighting, assessing the balance statistics again afterwards. Once matching on covariates is included in match_time, this no longer works.

When calling love.plot, estimates are thus also always displayed as "unadjusted". Users can change or entirely remove this labelling through options in the love.plot function.

Value

Returns a bal.tab object.

Examples

library(data.table)
library(MatchTime)

# only execute if packages are available
if (requireNamespace("cobalt") & requireNamespace("survival") &
    requireNamespace("MatchIt")) {

library(cobalt)
library(survival)
library(MatchIt)

# set random seed to make the output replicably
set.seed(1234)

# load "heart" data from survival package
data("heart")

# perform nearest neighbor time-dependent matching on "age" and "surgery"
# (plus exact matching on time)
out <- match_time(transplant ~ age + surgery, data=heart, id="id",
                  match_method="nearest")
bal.tab(out)
}
#>  cobalt (Version 4.6.0, Build Date: 2025-04-15)
#> 
#> Attaching package: ‘cobalt’
#> The following object is masked from ‘package:MatchIt’:
#> 
#>     lalonde
#> Warning: glm.fit: algorithm did not converge
#> Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
#> Warning: glm.fit: algorithm did not converge
#> Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
#> Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
#> Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
#> Balance Measures
#>            Type Diff.Un
#> event    Binary  0.4118
#> age     Contin.  0.1677
#> year    Contin.  0.0048
#> surgery  Binary  0.0000
#> 
#> Sample sizes
#>     FALSE TRUE
#> All    51   51

Display Balance Statistics in a Table for `match_time` objects