Real world evidence (RWE) offers an opportunity to conduct timely, large, low-cost studies to answer important clinical questions that may not always be feasible to investigate using randomized controlled trials for practical, financial or ethical reasons. However, RWE research also faces several important challenges.
One of these is unmeasured confounding that can invalidate comparisons between different treatment groups. While unmeasured confounders cannot, by definition, be measured, in this webinar we will discuss several indicators of unmeasured confounding that can help determine whether outcomes of a particular treatment group can be accurately compared to others. These will include:
- positive and negative control outcomes;
- distribution of known confounders; and
- clinically unexpected temporal relationships.
Discussion of these methods will be illustrated on the example of a multi-center PCORI-funded BESTMED study that investigated optimal 2nd line therapy of type 2 diabetes.