Clinical trials are researches with predifined protocols, performed on human subjects, to evaluate the effects of new medical treatments. Clinical trials invloves controlled intervetions on the treatment of subjects, to make the final results valid. The key step in a clinical trial is assignment of subjects to different treatments groups followed by measuring some outcome measures. Randomized controlled trials represent the gold standard in evaluating healthcare interventions, provided the used mathodology is valid.
Problem with random allocation
Subjects should be assigned to trial groups by chance, that is, in a way that can be predictable. Researchers may use different methods to generate the random sequence for assignment of subjects to trial groups. The commonest method of allocation is randomization, that is random allocation. Advantages of random allocation includes elimination of selection bias, validity of differences amoung treatment groups and under in some cases allocation blinding. The main drawback of randomization is the developement of unmatched groups by chance particularly in samples. If this occur the validity of trial results will be questionable. An alternative to randomization is minimization which is targeted at minimizing the differences amoung trial groups with respect to certain preselected prognotic factors. The first patient assign to one of trial groups randomly. Subsequent subject will be allocated to the group which minimize the overall imbalance amoung trial groups, that is the assignment to prefered treatment. To make the assingment unpredictable usually subject will be assign to prefered treatment with a higher probability that the one for non-prefered treatment. The whole process of minimization is computing-intensive particularly with unequal allocation ratios, therefore requires using a computer software.
What is MinimPy
MinimPy is an open source desktop application for mamaging the whole process of minimization. With MinimPy all aspects of a minimization protocole can be configured and controlled. It has optional network synchronization feathur which uses a central SVN repository. This is particularly usefull for multi-center trials. MinimPy program is written using Python programming language which is a powerful platform-independent scripting language suitable for scientific and mathematical calculations. In addition is has an easy to read code which is necessary by design. This makes the program logic transparent and provide an ideal media for future enhancement of program codes by various developers around the world.