Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Doctoral Award - Frederick Banting and Charles Best Canada Graduate Scholarships

McMaster University

$105,000.00

Statistical design and analysis methodology for clinical research

FY 2009-10

Summary
The proposed research intends to bridge the gap between biostatistics and clinical research, by means of developing, identifying and promoting valid statistical approaches to answer clinical questions arising from multicentre randomized controlled trials (RCTs), systematic reviews of RCTs and clinical guidelines. It addresses the following two specific issues that often act together and complicate the analysis of multicentre randomized controlled trials: (1)Prognostic heterogeneity/imbalance: Randomized Controlled Trials (RCTs) have become the "gold standard" for assessing health care interventions. Nevertheless, randomization does not guarantee complete balance of baseline prognostic factors between treatment groups, especially when the sample size is small. Neglecting imbalance of known or unknown prognostics between groups can result in biased estimate of treatment effect. Frequently used statistical methods are not capable of adjusting for the imbalance of unknown prognostic factors. We propose a simulations study to investigate the likelihood of occurrence of prognostic imbalance and its impact on statistical modeling. This work will inform researchers about the design and analysis of individual RCTs, and can also serve as a statistical foundation for rating quality of evidence in systematic reviews of RCTs and clinical guidelines. (2)Intraclass correlation in multicentre RCTs: In multicentre RCTs, similarity (or intraclass correlation) among outcomes from patients managed in the same centre compared with those in different centres is often ignored. The independence assumption of many statistical models is therefore violated, and the validity of study results is jeopardized. We propose to examine the impact of intraclass correlation on six models that are most frequently used in multicentre RCTs using simulation techniques. Study findings will be used to make accurate and efficient statistical inference in multicentre studies.
Program
Doctoral Award - Frederick Banting and Charles Best Canada Graduate Scholarships
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
CIHR Funding Decisions Database

As published

Id
206096
Name
Chu Rong
Region
Ontario
Country
Canada
Orgtype
Institution Paid; Research Institution
Instname2
Circulatory and Respiratory Health
Theme2
Biomedical
Prcname2
Doctoral Research Awards - A
Projecttitle
Statistical design and analysis methodology for clinical research
Programtype2
Studentship Programs
Orgnameinp2
McMaster University
Cihramount2
$105,000
Cihrequipment2
$0
Programname2
Doctoral Award - Frederick Banting and Charles Best Canada Graduate Scholarships
Keyworddelim
Adaptive Methods In Clinical Research; Biostatiatics; Design And Analysis Of Clinical Trials; Health Research Methdology; Intraclass Correlation; Meta-Analysis; Monte Carlo Simulation; Multicentre Clinical Trials; Prognostic Heterogeneity/Imbalance
Abstract
The proposed research intends to bridge the gap between biostatistics and clinical research, by means of developing, identifying and promoting valid statistical approaches to answer clinical questions arising from multicentre randomized controlled trials (RCTs), systematic reviews of RCTs and clinical guidelines. It addresses the following two specific issues that often act together and complicate the analysis of multicentre randomized controlled trials: (1)Prognostic heterogeneity/imbalance: Randomized Controlled Trials (RCTs) have become the "gold standard" for assessing health care interventions. Nevertheless, randomization does not guarantee complete balance of baseline prognostic factors between treatment groups, especially when the sample size is small. Neglecting imbalance of known or unknown prognostics between groups can result in biased estimate of treatment effect. Frequently used statistical methods are not capable of adjusting for the imbalance of unknown prognostic factors. We propose a simulations study to investigate the likelihood of occurrence of prognostic imbalance and its impact on statistical modeling. This work will inform researchers about the design and analysis of individual RCTs, and can also serve as a statistical foundation for rating quality of evidence in systematic reviews of RCTs and clinical guidelines. (2)Intraclass correlation in multicentre RCTs: In multicentre RCTs, similarity (or intraclass correlation) among outcomes from patients managed in the same centre compared with those in different centres is often ignored. The independence assumption of many statistical models is therefore violated, and the validity of study results is jeopardized. We propose to examine the impact of intraclass correlation on six models that are most frequently used in multicentre RCTs using simulation techniques. Study findings will be used to make accurate and efficient statistical inference in multicentre studies.
Competitiondate
200910
Approvedterm2
3 yrs 0 mth
Pinamesdelim
Chu, Rong
Approvedterm
3 ans 0 mois
Cihrcontribution2
~ $105,000~ $0~
McMaster University - $105,000 Grant (2009-10) | Canada Spends