In drug development, the control of type-I error is important as the consequence of approving a drug that is not efficacious can potentially cause public harm. Generally, control of type-I error is translated to being able to estimate the effects of treatments accurately in a randomized controlled trial (RCT). The concept of an informative data-based prior then are sometimes at odds with the rigor instituted by RCTs. To mitigate this problem apart from just using objective priors, the concept of exchangeability needs to be considered, i.e., the trial can be assumed to be exchangeable with other previous trials when the previous trials are considered to be good prior information. What this concept also implies is that the choice of the prior necessitates a sound exploration and synthesis of information from multiple sources at various stages. In this talk, we use the propensity scoring technique and explore strategies how historical data can be used effectively to form a basis for drug approval.