Type: Package
Title: KONP Tests: Powerful K-Sample Tests for Right-Censored Data
Version: 1.0.4
Date: 2021-12-27
Author: Matan Schlesinger [cre, aut], Malka Gorfine [aut]
Maintainer: Matan Schlesinger <matan.schles@gmail.com>
Description: The K-sample omnibus non-proportional hazards (KONP) tests are powerful non-parametric tests for comparing K (>=2) hazard functions based on right-censored data (Gorfine, Schlesinger and Hsu, 2020, <doi:10.1177/0962280220907355>). These tests are consistent against any differences between the hazard functions of the groups. The KONP tests are often more powerful than other existing tests, especially under non-proportional hazard functions.
License: GPL-2 | GPL-3 [expanded from: GPL (≥ 2)]
Imports: survival,Rcpp (≥ 0.12.16)
LinkingTo: Rcpp
RoxygenNote: 7.0.0
Encoding: UTF-8
Suggests: testthat
NeedsCompilation: yes
Packaged: 2022-01-03 16:28:18 UTC; matan.schlesinger
Repository: CRAN
Date/Publication: 2022-01-03 17:10:12 UTC

KONP Tests for Testing the Equality of K Distributions for Right-Censored Data

Description

An implementation of the K-sample omnibus non-proportional hazrds (KONP) tests.

The KONP tests are powerful non-parametric tests for comparing K (>=2) hazard functions based on right-censored data. These tests are consistent against any differences between the hazard functions of the groups. The KONP tests are often more powerful than other existing tests, especially under non-proportional hazard functions.

Details

The package contains one function:

konp_test: non-parametric tests for equality of K distributions using right-censored data.

Author(s)

Author and Maintainer: Matan Schlesinger matan.schles@gmail.com

Author: Malka Gorfine gorfinem@tauex.tau.ac.il

References

Gorfine, M., Schlesinger, M., & Hsu, L. (2020). K-sample omnibus non-proportional hazards tests based on right-censored data. Statistical Methods in Medical Research, 29(10), 2830–2850. doi: 10.1177/0962280220907355

Examples

  
# gastric cancer data  
data(gastric)

konp_test(gastric$time, gastric$status, gastric$group, n_perm=10^3)
  

Urothelial carcinoma.

Description

Survival data from a trial comparing chemotherapy versus atezolizumab in the treatment of Urothelial carcinoma.

Usage

data(carcinoma)

Format

A data frame with 625 observations (316 in the atezolizumab group and 309 chemotherapy group) with the following 3 columns:

time

the observed follow-up times in days.

status

the event indicators, 0=right censored, 1= event.

group

the group labels, 1 = atezolizumab, 2 = chemotherapy.

References

Powles T, Dura?n I, van der Heijden MS, et al. Atezolizumab versus chemotherapy in patients with platinum-treated locally advanced or metastatic urothelial carcinoma (IMvigor211): a multicentre, open-label, phase 3 randomised controlled trial. Lancet 2018; 391: 748-757.


Gastric Cancer Data.

Description

Survival data from a trial comparing chemotherapy versus combined chemotherapy plus radiotherapy in the treatment of gastric cancer.

Usage

data(gastric)

Format

A data frame with 90 observations (45 in each treatment group) with the following 3 columns:

time

the observed follow-up times in days.

status

the event indicators, 0=right censored, 1= event.

group

the group labels, 1 = chemotherapy, 2 = chemotherapy plus radiotherapy.

Source

Stablein, D. M. and Koutrouvelis, I. A. (1985) A two-sample test sensitive to crossing hazards in uncensored and singly censored data. Biometrics 41, 643–652. (Page 649).

References

Gastrointestinal Tumor Study Group: Schein, P. D., Stablein, D. M., Bruckner, H. W., Douglass, H. O., Mayer, R., et al. (1982). A comparison of combination chemotherapy and combined modality therapy for locally advanced gastricarcinoma. Cancer 49, 1771-1777.


KONP tests are K-sample Omnibus Non-Proportional hazards tests for right-censored data.

Description

KONP tests are K-sample Omnibus Non-Proportional hazards tests for right-censored data.

Usage

konp_test(time, status, group, n_perm, n_impu = 1)

Arguments

time

A vector of the observed follow-up times.

status

A vector of event indicators, 0=right censored, 1= event at time.

group

A vector denoting the group labels, must contain at least two different values.

n_perm

The number of permutations.

n_impu

The number of imputations, for each imputation n_perm permutations will be executed.

Details

The KONP tests are powerful non-parametric tests for comparing K (>=2) hazard functions based on right-censored data. These tests are consistent against any differences between the hazard functions of the groups. The KONP tests are often more powerful than other existing tests, especially under non-proportional hazard functions.

Value

Three test statistics and their respective p-values are returned:

pv_chisq - returns the p-value based on the KONP test chi-square statistic.
pv_lr - returns the p-value based on the KONP test likelihood ratio statistic.
pv_cauchy - returns the p-value based on the KONP-based Cauchy-combination test statistic.
chisq_test_stat - returns the KONP test chi-squared test statistic.
lr_test_stat - returns the KONP test likelihood-ratio test statistic.
cauchy_test_stat - returns the KONP-based Cauchy-combination test statistic.

Examples

# gastric cancer data
data(gastric)

konp_test(gastric$time, gastric$status, gastric$group, n_perm=10^3) 

mirror server hosted at Truenetwork, Russian Federation.