| This thesis examines some alternative methods to the conventional designs and analyses of clinical trials. In the first manuscript, we review various randomized control trials designs for accommodating participants' and professionals' preferences. Specifically, we discuss the advantages, limitations, applicability, ethical issues and statistical issues of each design. We also discuss the estimation of treatment effect, selection effect, and preference effect.;In the third manuscript, we examine four patterns of departures from assigned treatments and present an alternative randomized-based method of estimating treatment benefits. Specifically, we develop new estimators for continuous outcomes for each pattern and address the statistical properties of preference-based efficacy estimators for continuous outcomes. In addition, a comparison of the preference-based analysis, the intention-to-treat (ITT) analysis, per protocol analysis (PP) and as-treated analysis (AT) in the estimation of treatment efficacy in compliant patients is carried out using illustrative trials.;In the fourth manuscript, we compare the ITT, PP, AT and PB analyses of the effect of reamed or non-reamed intramedullary nailing of tibial shaft fractures on reoperation and/or autodynamization before one year III a multicenter, blinded randomized trial of 1339 adults.;In the fifth manuscript, we introduce the expertise-based randomized controlled trial, an alternative to conventional RCT, in which health care providers administer only one of the treatment interventions they are most familiar and comfortable with and then they use it for all their patients. We examined the relative efficiency of estimating the treatment effect in the expertise-based and conventional designs. These concepts were illustrated with data taken from a randomized trial of two alternative surgical techniques for tibia fractures.;In the second manuscript, we focus on three types of designs: the two-stage design, the preference-conventional clinical trial, and the partially randomized patient-preference clinical trial. We briefly review each of the designs and discuss the methods for estimating treatment, selection and preference effects from each design using illustrative trial data from the literature.;The sixth manuscript of the thesis focuses on crossover issues in conventional versus expertise-based randomized controlled trial. We examined the association between surgeon expertise and compliance with assigned intervention. We also looked at the relationship between surgeon expertise and patients outcomes using the SPRINT trial. |
