Keywords: Real World Inferences, Selection of Control Group, Missing Data
It is helpful when clinical trials are designed, conducted and analyzed to provide information relevant to a patient making a treatment decision. The design should consider the treatment options from which a patient would choose, and the control arm should receive an appropriate representation of standard of care. The conduct of the study should be consistent with the conduct a patient may take in the real world. All relevant outcomes should be captured from each subject. The results should be analyzed in a manner to inform treatment decisions. We will discuss what is needed to achieve these goals and why some designs and analyses fall short of achieving these goals.