Power and sample size calculation for paired recurrent events data based on robust nonparametric tests

Pei-Fang Su, Chia Hua Chung, Yu Wen Wang, Yun-Chan Chi, Ying-Ju Chang

Research output: Contribution to journalArticle

Abstract

The purpose of this paper is to develop a formula for calculating the required sample size for paired recurrent events data. The developed formula is based on robust non-parametric tests for comparing the marginal mean function of events between paired samples. This calculation can accommodate the associations among a sequence of paired recurrent event times with a specification of correlated gamma frailty variables for a proportional intensity model. We evaluate the performance of the proposed method with comprehensive simulations including the impacts of paired correlations, homogeneous or nonhomogeneous processes, marginal hazard rates, censoring rate, accrual and follow-up times, as well as the sensitivity analysis for the assumption of the frailty distribution. The use of the formula is also demonstrated using a premature infant study from the neonatal intensive care unit of a tertiary center in southern Taiwan.

Original languageEnglish
Pages (from-to)1823-1838
Number of pages16
JournalStatistics in Medicine
Volume36
Issue number11
DOIs
Publication statusPublished - 2017 May 20

Fingerprint

Sample Size Calculation
Robust Tests
Recurrent Events
Non-parametric test
Sample Size
Frailty
Neonatal Intensive Care Units
Taiwan
Premature Infants
Hazard Rate
Censoring
Sensitivity Analysis
Directly proportional
Specification
Unit
Evaluate
Simulation
Model

All Science Journal Classification (ASJC) codes

  • Epidemiology
  • Statistics and Probability

Cite this

@article{1d6a89264b6f42d19b147d35e606433c,
title = "Power and sample size calculation for paired recurrent events data based on robust nonparametric tests",
abstract = "The purpose of this paper is to develop a formula for calculating the required sample size for paired recurrent events data. The developed formula is based on robust non-parametric tests for comparing the marginal mean function of events between paired samples. This calculation can accommodate the associations among a sequence of paired recurrent event times with a specification of correlated gamma frailty variables for a proportional intensity model. We evaluate the performance of the proposed method with comprehensive simulations including the impacts of paired correlations, homogeneous or nonhomogeneous processes, marginal hazard rates, censoring rate, accrual and follow-up times, as well as the sensitivity analysis for the assumption of the frailty distribution. The use of the formula is also demonstrated using a premature infant study from the neonatal intensive care unit of a tertiary center in southern Taiwan.",
author = "Pei-Fang Su and Chung, {Chia Hua} and Wang, {Yu Wen} and Yun-Chan Chi and Ying-Ju Chang",
year = "2017",
month = "5",
day = "20",
doi = "10.1002/sim.7241",
language = "English",
volume = "36",
pages = "1823--1838",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "John Wiley and Sons Ltd",
number = "11",

}

Power and sample size calculation for paired recurrent events data based on robust nonparametric tests. / Su, Pei-Fang; Chung, Chia Hua; Wang, Yu Wen; Chi, Yun-Chan; Chang, Ying-Ju.

In: Statistics in Medicine, Vol. 36, No. 11, 20.05.2017, p. 1823-1838.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Power and sample size calculation for paired recurrent events data based on robust nonparametric tests

AU - Su, Pei-Fang

AU - Chung, Chia Hua

AU - Wang, Yu Wen

AU - Chi, Yun-Chan

AU - Chang, Ying-Ju

PY - 2017/5/20

Y1 - 2017/5/20

N2 - The purpose of this paper is to develop a formula for calculating the required sample size for paired recurrent events data. The developed formula is based on robust non-parametric tests for comparing the marginal mean function of events between paired samples. This calculation can accommodate the associations among a sequence of paired recurrent event times with a specification of correlated gamma frailty variables for a proportional intensity model. We evaluate the performance of the proposed method with comprehensive simulations including the impacts of paired correlations, homogeneous or nonhomogeneous processes, marginal hazard rates, censoring rate, accrual and follow-up times, as well as the sensitivity analysis for the assumption of the frailty distribution. The use of the formula is also demonstrated using a premature infant study from the neonatal intensive care unit of a tertiary center in southern Taiwan.

AB - The purpose of this paper is to develop a formula for calculating the required sample size for paired recurrent events data. The developed formula is based on robust non-parametric tests for comparing the marginal mean function of events between paired samples. This calculation can accommodate the associations among a sequence of paired recurrent event times with a specification of correlated gamma frailty variables for a proportional intensity model. We evaluate the performance of the proposed method with comprehensive simulations including the impacts of paired correlations, homogeneous or nonhomogeneous processes, marginal hazard rates, censoring rate, accrual and follow-up times, as well as the sensitivity analysis for the assumption of the frailty distribution. The use of the formula is also demonstrated using a premature infant study from the neonatal intensive care unit of a tertiary center in southern Taiwan.

UR - http://www.scopus.com/inward/record.url?scp=85012988686&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85012988686&partnerID=8YFLogxK

U2 - 10.1002/sim.7241

DO - 10.1002/sim.7241

M3 - Article

C2 - 28183151

AN - SCOPUS:85012988686

VL - 36

SP - 1823

EP - 1838

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 11

ER -