In clinical trials, best result will be obtained with controlled designs using randomised enrollment. By randomised we mean the process of subject allocation determines only by pure chance, and there is no selection bias in play. Furthermore by randomisation we aim to have balance over background variables such as demographic and other important prognostic factors, among branches of clinical trial. The rational behind having balanced groups is that if there is a significant deffernces among groups with respect to one or more baseline factors, then the results of clinical trial in unreliable which needs complication data analysis with overal loss of study power and validity. For this reason published articles present at the begininng of result sectiob the overal distribution of baseline factors among trial groups to demonstrate the state of balance among them and valifity of results and conclusions. Every study chooses different subsets of baseline factors to for the purpose of balancing thme among groups, though general factors such as gender, age and weight are included in most trials as the baseline factors. It must be emphasized absence of significant difference among groups with respect to baseline factors does not nessecerily equvalent to balanced groups.
Unfortunately is spite of random allocation of subject among groups, sometimes significant differences occure among trial groups with respect to baseline factors which bring considerable diturbances in analysis and interpretation of the results. The probability of this phenomenon is not low, though depends on the number of baseline factors, sample and effect sizes. The probability is high in small samples with large effect sizes and many baseline factors.
A method for preventing this phenomenon is minimisation which largely unknow among clinical researchers. In this method, allocation is not based solely and purely on chance but also the allocation is in the favor of minimising the difference among groups with respect of baseline factors. Therefore the subject in enrolled into a favored group with a more probability that non-favored groups. This process bring increasing balance with continuing enrollment of subjects into arms of trial.
Unfortunately the whole process of minimisation is calculation intensive and requires a lot of data keeping and maintenance which make manual application of minimisation practically impossible. There are several desktop or online application for implementation of minimisation. Some of these include Minim, MinimPy and QMinim.