In the recent decade, innovative drug development has brought more and more highly efficacious and less toxic drugs to patients. For such highly efficacious drugs, the design of confirmatory clinical trials for next generation treatment often aims to show a non-inferior treatment effect to that of the standard of care (SOC). Such comparative trials, however, may be prohibitive due to the large sample size needed with the consideration of budget and enrollment difficulty. Therefore, methods based on likelihood and least-squares are adapted to compare the overall treatment effect of the experimental drug to SOC using combined publicly or in-house available clinical trial data with adjustment for covariates. Simulation studies are conducted to suggest the minimum number of trials of different types, as well as number of patients per trial for more accurate and precise estimation. The proposed method is also applied to real studies to show non-inferiority of one drug to another after adjusting for several covariates. We also show that a similar approach can be extended to time-to-event data.