Electronic Thesis and Dissertation Repository

Thesis Format



Master of Science


Epidemiology and Biostatistics


Zou, Guangyong

2nd Supervisor

Jairath, Vipul

Joint Supervisor


This thesis aimed to review the literature on multiple-arm randomized control trials in inflammatory bowel disease (IBD) and to illustrate how to analyze these trials, focusing on appropriately controlling the type 1 error rates. The literature review found 247 trials published from the inception of each database to April 2014, of which 122 (49%) trials were multiple-arm trials and of those, 59 (48%) trials were on ulcerative colitis and 63 (52%) on Crohn’s disease. A published assessment tool was adopted to assess whether controlling of Type I error rates was needed. Despite the common use of this trial design and the need for multiple comparison procedures (MCPs) based on the assessment tool, only 20% of trials applied any MCPs. Failing to apply appropriate procedures may have inflated the Type I error rates, yielding false positive results. Data from a multiple-arm trial evaluating the efficacy of MLN02 (vedolizumab) was used as a case study to present how to implement common MCPs. Assessment tools regarding MCPs are helpful to have readily available to improve the scientific rigour of such research. Future research into the utility of online resources that guide multiple-arm trial design and analysis within IBD could promote such tools.

Summary for Lay Audience

Inflammatory bowel disease (IBD) is a term that describes disorders involving long-term (chronic) inflammation of tissues in the digestive tract. Two common forms of IBD are ulcerative colitis (UC) and Crohn’s disease (CD). Randomized controlled trials (RCTs) are the standard tool for evaluating the effect of treatments for IBD. This thesis reviewed the frequency of RCTs that used a design of more than two comparison groups. Such trials are termed multiple-arm trials. Since a critical statistical complication with such trials is controlling the false positive rate (Type I error) within the analysis, this thesis also enumerated and explored the procedures used to control the Type I error.

The literature review identified 217 randomized control trials about IBD, and among those, 59 UC and 63 CD disease trials had a multiple-arm design. An assessment tool evaluated these trials to determine if they possessed specific study characteristics, suggesting using a statistical procedure for multiple study groups to maintain statistical validity. Trials were also assessed on whether statistical procedures were initially implemented in their methods to determine the types of procedures already being used within IBD multiple-arm trials.

Data from an example study was analyzed and presented using some commonly used statistical procedures for trials with multiple study groups. This thesis established the importance of such statistical methods as they can help multiple group studies with specific characteristics control for statistical validity and maintain credible results and conclusions. Having tools and guidelines that assist with the decision-making around using such statistical methods could benefit the quality and efficacy of research work, especially within the field of IBD. Future research efforts could analyze online resources, which help set up the required statistical logistics for conducting multiple group IBD trials.

Creative Commons License

Creative Commons Attribution-Noncommercial 4.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 4.0 License.