Implementation of the Generalized Pairwise Comparisons.
BuyseTest is the main function of the package. See the vignette of an overview of the functionalities of the package.
Run citation("BuyseTest") in R for how to cite this package in scientific publications.
See the section reference below for examples of application in clinical studies.
The Generalized Pairwise Comparisons form all possible pairs of observations, one observation being taken from the intervention group and the other is taken from the control group, and compare the difference in endpoints (\(Y-X\)) to the threshold of clinical relevance (\(\tau\)).
For a single endpoint, if the difference is greater or equal than the threshold of clinical relevance (\(Y \ge X + \tau\)), the pair is classified as favorable (i.e. win). If the difference is lower or equal than minus the threshold of clinical relevance (\(X \ge Y + \tau\)), the pair is classified as unfavorable (i.e. loss). Otherwise the pair is classified as neutral. In presence of censoring, it might not be possible to compare the difference to the threshold. In such cases the pair is classified as uninformative.
Simultaneously analysis of several endpoints is performed by prioritizing the endpoints, assigning the highest priority to the endpoint considered the most clinically relevant. The endpoint with highest priority is analyzed first, and neutral and uninformative pair are analyzed regarding endpoint of lower priority.
Keywords: documented methods/functions are classified according to the following keywords
models: function fitting a statistical model/method based on a dataset (e.g. auc, brier, BuyseTest, BuyseTTEM, CasinoTest, performance)
htest: methods performing statistical inference based on an existing model (e.g. BuyseMultComp, performanceResample, powerBuyseTest, sensitivity)
methods: extractors (e.g. getCount, getPairScore, getPseudovalue, getSurvival, getIid)
print: concise display of an object in the console (e.g. print, summary)
utilities: function used to facilitate user interactions (e.g. BuyseTest.options, constStrata)
hplot: graphical display (e.g. autoplot.S3sensitivity)
internal: function used internally but that need to be exported for parallel calculations (e.g. GPC_cpp)
datagen: function for generating data sets (e.g. simBuyseTest, simCompetingRisks)
classes: definition of S4 classes
Maintainer: Brice Ozenne brice.mh.ozenne@gmail.com (ORCID)
Authors:
Eva Cantagallo
William Anderson
Other contributors:
Julien Peron [contributor]
Johan Verbeeck [contributor]
Method papers on the GPC procedure and its extensions:
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 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
On the stratified win ratio: G. Dong et al. (2018). The stratified win ratio. Journal of biopharmaceutical statistics. 28(4):778-796
On the Peron's scoring rule: J. Peron, M. Buyse, B. Ozenne, L. Roche and P. Roy (2018). An extension of generalized pairwise comparisons for prioritized outcomes in the presence of censoring. Statistical Methods in Medical Research 27: 1230-1239.
On the Gehan's scoring rule: Gehan EA (1965). A generalized two-sample Wilcoxon test for doubly censored data. Biometrika 52(3):650-653
On inference in GPC using the U-statistic theory: Ozenne B, Budtz-Jorgensen E, Peron J (2021). The asymptotic distribution of the Net Benefit estimator in presence of right-censoring. Statistical Methods in Medical Research 2021 doi:10.1177/09622802211037067
On how to handle right-censoring: J. Peron, M. Idlhaj, D. Maucort-Boulch, et al. (2021) Correcting the bias of the net benefit estimator due to right-censored observations. Biometrical Journal 63: 893–906.
On how using a restriction time: Piffoux M, Ozenne B, De Backer M, Buyse M, Chiem JC, Péron J (2024). Restricted Net Treatment Benefit in oncology. Journal of Clinical Epidemiology. Jun;170:111340.
Examples of application in clinical studies:
J. Peron, P. Roy, K. Ding, W. R. Parulekar, L. Roche, M. Buyse (2015). Assessing the benefit-risk of new treatments using generalized pairwise comparisons: the case of erlotinib in pancreatic cancer. British journal of cancer 112:(6)971-976.
J. Peron, P. Roy, T. Conroy, F. Desseigne, M. Ychou, S. Gourgou-Bourgade, T. Stanbury, L. Roche, B. Ozenne, M. Buyse (2016). An assessment of the benefit-risk balance of FOLFORINOX in metastatic pancreatic adenocarcinoma. Oncotarget 7:82953-60, 2016.
Discussion about the relevance of GPC based measures of treatment effect:
J. Peron, P. Roy, B. Ozenne, L. Roche, M. Buyse (2016). The net chance of a longer survival as a patient-oriented measure of benefit in randomized clinical trials. JAMA Oncology 2:901-5.
E. D. Saad , J. R. Zalcberg, J. Peron, E. Coart, T. Burzykowski, M. Buyse (2018). Understanding and communicating measures of treatment effect on survival: can we do better?. J Natl Cancer Inst.
Ezimamaka Ajufo, Aditi Nayak, Mandeep R. Mehra (2023). Fallacies of Using the Win Ratio in Cardiovascular Trials: Challenges and Solutions. JACC: Basic to Translational Science. 8(6):720-727.
Useful links: