A Novel Statistical Method For Comparing Effectiveness Of Two Treatments-simulated Randomized Controlled Trials

Objective: To develop a new statistical method for comparing effectiveness of twotreatments for the same disease indication. A statistical method needs to be developed tocompare effectiveness of different treatments by using existing clinical data when lack ofevidence from randomized controlled trials. Methods: Datasets including an outcomevariable and a confounder variable for two treatments were simulated，with simulatedknown group differences and statistical powers. Subjects in each dataset were randomizedinto two arms for100times，and were analyzed in the way of RCTs for100times.Per-protocol analysis strategy was applied. The means of the two arms were compared byusing student-t tests after excluding the observations with simulated allocation that werecontradictory to their actual received treatments. The ratio of the frequency of hypothesistests that rejected H₀to the frequency of hypothesis tests that did not reject H₀，calledodds，was used as the statistic of the method to indicate the probability of significant testsfor group difference. To document the consistency and reliability of the method， thedistribution of odds and its95%CI were plotted in simulated datasets with variousbetween-group mean differences and statistical power （ranging from0.5to0.85） withvaried sample sizes （n=50，100，500，and1000）. Stata11.0was used to program andperform the analysis. A set of real clinical data about two kinds of interferons as therapiesto hepatitis C were used to test the validity of the method. Results: The odds and its95%CI of simulated RCTs were perfectly and linearly correlated with the change ofbetween group mean differences and statistical power，by difference sample size. Aconclusion can be made based on the hypothesis of the simulated randomized controlledtrials （sRCT）. The probability of loss of balance of confounding was over95%for equaland unequal sample size of two arms after excluding misclassified subjects.Conclusions:The proposed novel analytical method，simulated RCTs based on real clinical treatmentdata，to synthesize a new kind of evidence helpful for clinical decision making.