Executive Summary : | The Covid-19 pandemic has highlighted the need for increased emphasis on health sciences research, with billions of dollars spent annually on clinical trials. These trials are used in medical sciences, psychology, and public policy to develop new drugs or treatments. This research project aims to explore the selection of participants and the ideal number of individuals for a Stepped Wedge Cluster Randomized Trial (SW-CRT) study. Clusterrandomized trials (CRTs) are popular in epidemiological and community-based clinical studies, but there is limited literature on crossover cluster randomized trials, particularly Stepped Wedge Cluster Randomized Trials (SW-CRTs). The objective is to establish the most effective sample sizes for detecting intervention effects in a SW-CRT while considering budget limitations. Dropouts from large population-based epidemiological studies can impact the study's analysis, so the project will identify necessary sample sizes at the design level while accounting for random subject attrition. The project will also consider a four-level hierarchical data structure, where repeated observations are nested within patients, doctors, and hospitals. Ignoring hierarchical structure in data can lead to low statistical power and increased sample size, increasing the cost of the study. This project will be the first attempt to consider four-level data in SW-CRTs, as no theoretical research work has been done. The proposed model will be assessed through real-life datasets and comparison studies with existing models. User-friendly software packages will be developed to help experimentalists apply complex models to real-life scenarios. |