Impact of subject attrition on sample size determinations for longitudinal cluster randomized clinical trials

Moonseong Heo

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

Subject attrition is a ubiquitous problem in any type of clinical trial and, thus, needs to be taken into consideration at the design stage particularly to secure adequate statistical power. Here, we focus on longitudinal cluster randomized clinical trials (cluster-RCT) that aim to test the hypothesis that an intervention has an effect on the rate of change in the outcome over time. In this setting, the cluster-RCT assumes a three-level hierarchical data structure in which subjects are nested within a higher level unit such as clinics and are evaluated for outcome repeatedly over the study period. Furthermore, the subject-specific slopes can be modeled in terms of fixed or random coefficients in a mixed-effects linear model. Closed-form sample size formulas for testing the preceding hypothesis have been developed under an assumption of no attrition. In this article, we propose closed-form approximate samples size determinations with anticipated attrition rates by modifying those existing sample size formulas. With extensive simulations, we examine performances of the modified formulas under three attrition mechanisms: attrition completely at random, attrition at random, and attrition not at random. In conclusion, the proposed modification is very effective under fixed-slope models but yields biased, perhaps substantially so, statistical power under random slope models.

Original languageEnglish (US)
Pages (from-to)507-522
Number of pages16
JournalJournal of Biopharmaceutical Statistics
Volume24
Issue number3
DOIs
StatePublished - May 4 2014

Fingerprint

Sample Size Determination
Attrition
Randomized Clinical Trial
Sample Size
Randomized Controlled Trials
Slope
Statistical Power
Linear Models
Clinical Trials
Closed-form
Linear Mixed Effects Model
Hierarchical Data
Random Coefficients
Rate of change
Hierarchical Structure
Biased
Data Structures
Testing
Unit
Model

Keywords

  • Attrition
  • Effect size
  • Longitudinal cluster RCT
  • Power
  • Sample size
  • Three-level data.

ASJC Scopus subject areas

  • Pharmacology (medical)
  • Pharmacology
  • Statistics and Probability
  • Medicine(all)

Cite this

Impact of subject attrition on sample size determinations for longitudinal cluster randomized clinical trials. / Heo, Moonseong.

In: Journal of Biopharmaceutical Statistics, Vol. 24, No. 3, 04.05.2014, p. 507-522.

Research output: Contribution to journalArticle

@article{28c163f840f448bbb539eb27f46b3a09,
title = "Impact of subject attrition on sample size determinations for longitudinal cluster randomized clinical trials",
abstract = "Subject attrition is a ubiquitous problem in any type of clinical trial and, thus, needs to be taken into consideration at the design stage particularly to secure adequate statistical power. Here, we focus on longitudinal cluster randomized clinical trials (cluster-RCT) that aim to test the hypothesis that an intervention has an effect on the rate of change in the outcome over time. In this setting, the cluster-RCT assumes a three-level hierarchical data structure in which subjects are nested within a higher level unit such as clinics and are evaluated for outcome repeatedly over the study period. Furthermore, the subject-specific slopes can be modeled in terms of fixed or random coefficients in a mixed-effects linear model. Closed-form sample size formulas for testing the preceding hypothesis have been developed under an assumption of no attrition. In this article, we propose closed-form approximate samples size determinations with anticipated attrition rates by modifying those existing sample size formulas. With extensive simulations, we examine performances of the modified formulas under three attrition mechanisms: attrition completely at random, attrition at random, and attrition not at random. In conclusion, the proposed modification is very effective under fixed-slope models but yields biased, perhaps substantially so, statistical power under random slope models.",
keywords = "Attrition, Effect size, Longitudinal cluster RCT, Power, Sample size, Three-level data.",
author = "Moonseong Heo",
year = "2014",
month = "5",
day = "4",
doi = "10.1080/10543406.2014.888442",
language = "English (US)",
volume = "24",
pages = "507--522",
journal = "Journal of Biopharmaceutical Statistics",
issn = "1054-3406",
publisher = "Taylor and Francis Ltd.",
number = "3",

}

TY - JOUR

T1 - Impact of subject attrition on sample size determinations for longitudinal cluster randomized clinical trials

AU - Heo, Moonseong

PY - 2014/5/4

Y1 - 2014/5/4

N2 - Subject attrition is a ubiquitous problem in any type of clinical trial and, thus, needs to be taken into consideration at the design stage particularly to secure adequate statistical power. Here, we focus on longitudinal cluster randomized clinical trials (cluster-RCT) that aim to test the hypothesis that an intervention has an effect on the rate of change in the outcome over time. In this setting, the cluster-RCT assumes a three-level hierarchical data structure in which subjects are nested within a higher level unit such as clinics and are evaluated for outcome repeatedly over the study period. Furthermore, the subject-specific slopes can be modeled in terms of fixed or random coefficients in a mixed-effects linear model. Closed-form sample size formulas for testing the preceding hypothesis have been developed under an assumption of no attrition. In this article, we propose closed-form approximate samples size determinations with anticipated attrition rates by modifying those existing sample size formulas. With extensive simulations, we examine performances of the modified formulas under three attrition mechanisms: attrition completely at random, attrition at random, and attrition not at random. In conclusion, the proposed modification is very effective under fixed-slope models but yields biased, perhaps substantially so, statistical power under random slope models.

AB - Subject attrition is a ubiquitous problem in any type of clinical trial and, thus, needs to be taken into consideration at the design stage particularly to secure adequate statistical power. Here, we focus on longitudinal cluster randomized clinical trials (cluster-RCT) that aim to test the hypothesis that an intervention has an effect on the rate of change in the outcome over time. In this setting, the cluster-RCT assumes a three-level hierarchical data structure in which subjects are nested within a higher level unit such as clinics and are evaluated for outcome repeatedly over the study period. Furthermore, the subject-specific slopes can be modeled in terms of fixed or random coefficients in a mixed-effects linear model. Closed-form sample size formulas for testing the preceding hypothesis have been developed under an assumption of no attrition. In this article, we propose closed-form approximate samples size determinations with anticipated attrition rates by modifying those existing sample size formulas. With extensive simulations, we examine performances of the modified formulas under three attrition mechanisms: attrition completely at random, attrition at random, and attrition not at random. In conclusion, the proposed modification is very effective under fixed-slope models but yields biased, perhaps substantially so, statistical power under random slope models.

KW - Attrition

KW - Effect size

KW - Longitudinal cluster RCT

KW - Power

KW - Sample size

KW - Three-level data.

UR - http://www.scopus.com/inward/record.url?scp=84898952276&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84898952276&partnerID=8YFLogxK

U2 - 10.1080/10543406.2014.888442

DO - 10.1080/10543406.2014.888442

M3 - Article

C2 - 24697555

AN - SCOPUS:84898952276

VL - 24

SP - 507

EP - 522

JO - Journal of Biopharmaceutical Statistics

JF - Journal of Biopharmaceutical Statistics

SN - 1054-3406

IS - 3

ER -