Extract summary statistics (net benefit, win ratio, ...) from GPC.
# S4 method for S4BuyseTest
coef(
object,
endpoint = NULL,
statistic = NULL,
strata = FALSE,
cumulative = NULL,
resampling = FALSE,
simplify = TRUE,
...
)When resampling=FALSE and simplify=FALSE, a matrix (strata, endpoint).
When resampling=FALSE and simplify=FALSE, an array (sample, strata, endpoint).
a S4BuyseTest object, output of BuyseTest.
[character] for which endpoint(s) the summary statistic should be output?
If NULL returns the summary statistic for all endpoints.
[character] the statistic summarizing the pairwise comparison:
"netBenefit": displays the net benefit, as described in Buyse (2010) and Peron et al. (2016)),
"winRatio": win ratio or win odds.
"favorable": proportion strictly in favor of the treatment or Mann-Whitney parameter.
"unfavorable": proportion in favor of the control.
"neutral": proportion of neutral pairs.
"uninf": proportion of uninformative pairs.
"count.favorable": number of pairs in favor of the treatment.
"count.unfavorable": number of pairs in favor of the control.
"count.neutral": number of neutral pairs.
"count.uninf": number of uninformative pairs.
Default value read from BuyseTest.options().
[character vector] the strata relative to which the statistic should be output.
Can also be "global" or FALSE to output the statistic pooled over all strata,
or TRUE to output each strata-specific statistic.
[logical] should the summary statistic be cumulated over endpoints? Otherwise display the contribution of each endpoint.
[logical] should the summary statistic obtained by resampling be output?
[logical] should the result be coerced to the lowest possible dimension?
ignored.
Brice Ozenne
statistic: with a single endpoint denoted \(Y\) and \(X\) in the treatment and control group and a threshold of clinical relevance \(\tau\):
"netBenefit": \(P[Y \ge X + \tau] - P[X \ge Y + \tau]\). See Buyse (2010).
"winRatio": the win ratio \(\frac{P[Y \ge X + \tau]}{P[X \ge Y + \tau]}\) or the win odds \(\frac{P[Y \ge X + \tau]+0.5P[|Y - X|<\tau]}{P[X \ge Y + \tau]+0.5P[|Y - X|<\tau]}\). see Wang (2016) and Dong (2019).
"favorable": \(P[Y \ge X + \tau]\) or the Mann-Whitney parameter \(P[Y \ge X + \tau]+0.5P[|Y - X|<\tau]\). See Fay (2018).
"unfavorable": \(P[Y \le X + \tau]\) or \(P[Y \le X + \tau]+0.5P[|Y - X|<\tau]\).
The value of the argument add.halfNeutral used when running BuyseTest decides whether \(0.5P[|Y - X|<\tau]\) is considered, e.g. whether the win ratio or win odds is output.
On the GPC procedure: Marc Buyse (2010). Generalized pairwise comparisons of prioritized endpoints in the two-sample problem. Statistics in Medicine 29:3245-3257
On the Mann-Whitney parameter: Fay, Michael P. et al (2018). Causal estimands and confidence intervals asscoaited with Wilcoxon-Mann-Whitney tests in randomized experiments. Statistics in Medicine 37:2923-2937
On the win odds: Dong, G., Hoaglin, D. C., Qiu, J., Matsouaka, R. A., Chang, Y. W., Wang, J., & Vandemeulebroecke, M. (2019). The Win Ratio: On Interpretation and Handling of Ties. Statistics in Biopharmaceutical Research, 12(1), 99–106. https://doi.org/10.1080/19466315.2019.1575279
On the win ratio: D. Wang, S. Pocock (2016). A win ratio approach to comparing continuous non-normal outcomes in clinical trials. Pharmaceutical Statistics 15:238-245