UK-Cuboid Randomized Complete Block Design: A Complete Three Way Layout for Large Multi-Center Clinical Trials

In clinical epidemiology research, if principal investigator needs to know whether specific drug which has been invented is advantageous or not in the treatment of particular disease, there are many experimental designs to approach, in which most significant one is Latin Square Design (LSD). The LSD has its own merits and demerits. The major disadvantages of LSD are 1) Incomplete three-way layout.2) Difficult to handle large sample size 3) The analysis of data in LSD is conditional. Our aim is to provide a new design of experiment to overcome all these disadvantages with additional benefits and also useful for large multi-center clinical trials. A new design of an experiment is an effort to extend UK-Cubic complete random design (UKCD) towards UK-Cuboid randomized complete block design (UKCBD) to provide a way to include larger number of treatments and unequal/equal number of levels (characteristics that differentiate factors) of the various independent variables effortlessly. The depending study variable is simulated using the simulated model. Each term in the simulated model, plus the simulated model as a whole, is tested using ANOVA for its ability to account for variation in the dependent variable, to reveal the efficiency of the UKCBD design. Simulation model is used to have values of depending study variable hemoglobin level of anemic patients for the operation of the assumed Phase III multi-center clinical trial for the 192 anemic patients arranged in the original order of the three factors such as, 2 age-groups, 3 centers and 4 districts in such a way that four treatments randomly allocated to twenty four cells which have 2 blocks based on gender (Female, Male) and 4 anemic patients in each block. The analysis tools ANOVA and post hoc tests imply with the independent variables such as treatment, age-groups, centers, districts and blocks that the outcome of treatments of anemic patients is correctly found as per simulation for the randomized data of depending study variable. The mean hemoglobin level of anemic patients under the new drugs is significantly better than that under placebo as in simulated order. It reveals the efficiency of this design and its good randomization. In real life situation, it is not always possible to form equal number of levels with all the associate factors. Using this new design, it is possible to accommodate either unequal or equal number of levels of the associated factors. This design will be useful for large multi-center clinical trials. It includes larger number of participants, more number of treatments, different geographic locations, wider range of population groups and the ability to compare results among centers, all of which increase the generalizablity of the multi-center clinical trial study. It also controls more variation and results in a smaller mean square error.